OV - exec - Loop 1256

	Loop Id: 1256	Module: exec	Source: par_lr_interp.c:1624-1628	Coverage: 0.04%

0x27cd40 KORB	%K5,%K4,%K4

0x27cd44 KORB	%K0,%K2,%K2

0x27cd48 KORB	%K3,%K1,%K1

0x27cd4c ADD	$0x20,%R8

0x27cd50 VMOVAPD	%ZMM6,%ZMM23{%K4}

0x27cd56 KMOVW	0x50(%RSP),%K4    [4]

0x27cd5c VMOVAPD	%ZMM18,%ZMM25{%K2}

0x27cd62 VMOVAPD	%ZMM17,%ZMM26{%K1}

0x27cd68 VMOVAPD	%ZMM23,%ZMM6

0x27cd6e VMOVAPD	%ZMM25,%ZMM18

0x27cd74 VMOVAPD	%ZMM26,%ZMM17

0x27cd7a KORB	%K4,%K6,%K4

0x27cd7e VMOVAPD	%ZMM16,%ZMM24{%K4}

0x27cd84 VMOVAPD	%ZMM24,%ZMM16

0x27cd8a CMP	%R8,%RSI

0x27cd8d JE	27d1a0

0x27cd93 MOV	-0x88(%RCX,%R8,8),%RDX    [2]

0x27cd9b VMOVQ	(%R15,%RDX,8),%XMM25    [3]

0x27cda2 MOV	-0x90(%RCX,%R8,8),%RDX    [2]

0x27cdaa VMOVQ	(%R15,%RDX,8),%XMM30    [3]

0x27cdb1 MOV	-0x98(%RCX,%R8,8),%RDX    [2]

0x27cdb9 VMOVSD	(%R15,%RDX,8),%XMM0    [3]

0x27cdbf MOV	-0xa0(%RCX,%R8,8),%RDX    [2]

0x27cdc7 VPUNPCKLQDQ	%XMM25,%XMM30,%XMM30

0x27cdcd VMOVAPS	%XMM0,0x60(%RSP)    [4]

0x27cdd3 VMOVSD	(%R15,%RDX,8),%XMM0    [3]

0x27cdd9 MOV	-0xa8(%RCX,%R8,8),%RDX    [2]

0x27cde1 VMOVAPS	%XMM0,0x50(%RSP)    [4]

0x27cde7 VMOVSD	(%R15,%RDX,8),%XMM0    [3]

0x27cded MOV	-0xb0(%RCX,%R8,8),%RDX    [2]

0x27cdf5 VMOVAPS	%XMM0,0xc0(%RSP)    [4]

0x27cdfe VMOVSD	(%R15,%RDX,8),%XMM0    [3]

0x27ce04 MOV	-0xb8(%RCX,%R8,8),%RDX    [2]

0x27ce0c VMOVAPS	%XMM0,0x100(%RSP)    [4]

0x27ce15 VMOVSD	(%R15,%RDX,8),%XMM0    [3]

0x27ce1b MOV	-0xc0(%RCX,%R8,8),%RDX    [2]

0x27ce23 VMOVQ	(%R15,%RDX,8),%XMM26    [3]

0x27ce2a MOV	-0x48(%RCX,%R8,8),%RDX    [2]

0x27ce2f VMOVAPS	%XMM0,0x130(%RSP)    [4]

0x27ce38 VMOVSD	(%R15,%RDX,8),%XMM0    [3]

0x27ce3e MOV	-0x50(%RCX,%R8,8),%RDX    [2]

0x27ce43 VMOVQ	(%R15,%RDX,8),%XMM23    [3]

0x27ce4a MOV	-0x58(%RCX,%R8,8),%RDX    [2]

0x27ce4f VMOVAPS	%XMM0,0x240(%RSP)    [4]

0x27ce58 VMOVDQA64	0x50(%RSP),%XMM25    [4]

0x27ce60 VPUNPCKLQDQ	0x130(%RSP),%XMM26,%XMM26    [4]

0x27ce68 VPUNPCKLQDQ	0x240(%RSP),%XMM23,%XMM23    [4]

0x27ce70 VMOVQ	(%R15,%RDX,8),%XMM24    [3]

0x27ce77 MOV	-0x60(%RCX,%R8,8),%RDX    [2]

0x27ce7c VPUNPCKLQDQ	0x60(%RSP),%XMM25,%XMM25    [4]

0x27ce84 VMOVQ	(%R15,%RDX,8),%XMM21    [3]

0x27ce8b MOV	-0x68(%RCX,%R8,8),%RDX    [2]

0x27ce90 VINSERTI32X4	$0x1,%XMM30,%YMM25,%YMM25

0x27ce97 VMOVDQA64	0x100(%RSP),%XMM30    [4]

0x27ce9f VPUNPCKLQDQ	0xc0(%RSP),%XMM30,%XMM30    [4]

0x27cea7 VMOVQ	(%R15,%RDX,8),%XMM22    [3]

0x27ceae MOV	-0x70(%RCX,%R8,8),%RDX    [2]

0x27ceb3 VPUNPCKLQDQ	%XMM24,%XMM21,%XMM21

0x27ceb9 VINSERTI32X4	$0x1,%XMM23,%YMM21,%YMM21

0x27cec0 VMOVQ	(%R15,%RDX,8),%XMM19    [3]

0x27cec7 MOV	-0x78(%RCX,%R8,8),%RDX    [2]

0x27cecc VINSERTI32X4	$0x1,%XMM30,%YMM26,%YMM26

0x27ced3 VINSERTI64X4	$0x1,%YMM25,%ZMM26,%ZMM25

0x27ceda VPCMPNLTQ	%ZMM9,%ZMM25,%K4

0x27cee1 VMOVQ	(%R15,%RDX,8),%XMM20    [3]

0x27cee8 MOV	-0x80(%RCX,%R8,8),%RDX    [2]

0x27ceed VPUNPCKLQDQ	%XMM22,%XMM19,%XMM19

0x27cef3 VMOVQ	(%R15,%RDX,8),%XMM2    [3]

0x27cef9 MOV	-0x8(%RCX,%R8,8),%RDX    [2]

0x27cefe VMOVQ	(%R15,%RDX,8),%XMM4    [3]

0x27cf04 MOV	-0x10(%RCX,%R8,8),%RDX    [2]

0x27cf09 VPUNPCKLQDQ	%XMM20,%XMM2,%XMM2

0x27cf0f VINSERTI32X4	$0x1,%XMM19,%YMM2,%YMM2

0x27cf16 VMOVQ	(%R15,%RDX,8),%XMM11    [3]

0x27cf1c MOV	-0x18(%RCX,%R8,8),%RDX    [2]

0x27cf21 VINSERTI64X4	$0x1,%YMM21,%ZMM2,%ZMM2

0x27cf28 VPCMPNLTQ	%ZMM9,%ZMM2,%K1

0x27cf2f VMOVQ	(%R15,%RDX,8),%XMM7    [3]

0x27cf35 MOV	-0x20(%RCX,%R8,8),%RDX    [2]

0x27cf3a VPUNPCKLQDQ	%XMM4,%XMM11,%XMM4

0x27cf3e VMOVQ	(%R15,%RDX,8),%XMM1    [3]

0x27cf44 MOV	-0x28(%RCX,%R8,8),%RDX    [2]

0x27cf49 VMOVQ	(%R15,%RDX,8),%XMM3    [3]

0x27cf4f MOV	-0x30(%RCX,%R8,8),%RDX    [2]

0x27cf54 VPUNPCKLQDQ	%XMM7,%XMM1,%XMM1

0x27cf58 VINSERTI128	$0x1,%XMM4,%YMM1,%YMM1

0x27cf5e VMOVQ	(%R15,%RDX,8),%XMM10    [3]

0x27cf64 MOV	-0x38(%RCX,%R8,8),%RDX    [2]

0x27cf69 VMOVQ	(%R15,%RDX,8),%XMM5    [3]

0x27cf6f MOV	-0x40(%RCX,%R8,8),%RDX    [2]

0x27cf74 VPUNPCKLQDQ	%XMM3,%XMM10,%XMM3

0x27cf78 VMOVQ	(%R15,%RDX,8),%XMM31    [3]

0x27cf7f MOV	0x38(%RCX,%R8,8),%RDX    [2]

0x27cf84 VMOVQ	(%R15,%RDX,8),%XMM8    [3]

0x27cf8a MOV	0x30(%RCX,%R8,8),%RDX    [2]

0x27cf8f VPUNPCKLQDQ	%XMM5,%XMM31,%XMM4

0x27cf95 VINSERTI128	$0x1,%XMM3,%YMM4,%YMM3

0x27cf9b VMOVQ	(%R15,%RDX,8),%XMM14    [3]

0x27cfa1 MOV	0x28(%RCX,%R8,8),%RDX    [2]

0x27cfa6 VINSERTI64X4	$0x1,%YMM1,%ZMM3,%ZMM1

0x27cfad VPCMPNLTQ	%ZMM9,%ZMM1,%K2

0x27cfb4 VMOVDQA64	0x380(%RSP),%ZMM1    [4]

0x27cfbc VPCMPEQQ	-0xc0(%RCX,%R8,8),%ZMM1,%K3    [2]

0x27cfc4 VPCMPEQQ	-0x80(%RCX,%R8,8),%ZMM1,%K6    [2]

0x27cfcc VPCMPEQQ	-0x40(%RCX,%R8,8),%ZMM1,%K7    [2]

0x27cfd4 VPCMPEQQ	(%RCX,%R8,8),%ZMM1,%K0    [2]

0x27cfdb VMOVQ	(%R15,%RDX,8),%XMM12    [3]

0x27cfe1 MOV	0x20(%RCX,%R8,8),%RDX    [2]

0x27cfe6 VPUNPCKLQDQ	%XMM8,%XMM14,%XMM3

0x27cfeb VMOVQ	(%R15,%RDX,8),%XMM0    [3]

0x27cff1 MOV	0x18(%RCX,%R8,8),%RDX    [2]

0x27cff6 KORB	%K4,%K3,%K4

0x27cffa KORB	%K1,%K6,%K3

0x27cffe KORB	%K2,%K7,%K2

0x27d002 VMOVQ	(%R15,%RDX,8),%XMM15    [3]

0x27d008 MOV	0x10(%RCX,%R8,8),%RDX    [2]

0x27d00d VPUNPCKLQDQ	%XMM12,%XMM0,%XMM0

0x27d012 VINSERTI128	$0x1,%XMM3,%YMM0,%YMM0

0x27d018 VMOVQ	(%R15,%RDX,8),%XMM27    [3]

0x27d01f MOV	0x8(%RCX,%R8,8),%RDX    [2]

0x27d024 VMOVQ	(%R15,%RDX,8),%XMM28    [3]

0x27d02b MOV	(%RCX,%R8,8),%RDX    [2]

0x27d02f VPUNPCKLQDQ	%XMM15,%XMM27,%XMM3

0x27d035 VMOVQ	(%R15,%RDX,8),%XMM29    [3]

0x27d03c VPUNPCKLQDQ	%XMM28,%XMM29,%XMM4

0x27d042 VINSERTI128	$0x1,%XMM3,%YMM4,%YMM3

0x27d048 VINSERTI64X4	$0x1,%YMM0,%ZMM3,%ZMM0

0x27d04f VPCMPNLTQ	%ZMM9,%ZMM0,%K5

0x27d056 KORB	%K5,%K0,%K1

0x27d05a KUNPCKBW	%K4,%K3,%K0

0x27d05e KUNPCKWD	%K0,%K2,%K0

0x27d062 VPMOVM2W	%K1,%ZMM1

0x27d068 VPMOVM2W	%K0,%ZMM0

0x27d06e VINSERTI32X4	$0x3,%XMM1,%ZMM0,%ZMM0

0x27d075 VPMOVW2M	%ZMM0,%K0

0x27d07b KORTESTD	%K0,%K0

0x27d080 KMOVD	%K0,0x50(%RSP)    [4]

0x27d087 JE	27d0f8

0x27d089 VMOVUPD	-0x40(%R10,%R8,8),%ZMM21{%K2}{z}    [1]

0x27d091 VMOVAPD	0x340(%RSP),%ZMM3    [4]

0x27d099 VMOVUPD	-0xc0(%R10,%R8,8),%ZMM19{%K4}{z}    [1]

0x27d0a1 VMOVUPD	-0x80(%R10,%R8,8),%ZMM20{%K3}{z}    [1]

0x27d0a9 VMOVUPD	(%R10,%R8,8),%ZMM22{%K1}{z}    [1]

0x27d0b0 VPXOR	%XMM4,%XMM4,%XMM4

0x27d0b4 VMULPD	%ZMM3,%ZMM21,%ZMM2

0x27d0ba VMULPD	%ZMM3,%ZMM19,%ZMM0

0x27d0c0 VMULPD	%ZMM3,%ZMM20,%ZMM1

0x27d0c6 VMULPD	%ZMM3,%ZMM22,%ZMM3

0x27d0cc VCMPPD	$0x1,%ZMM4,%ZMM2,%K0

0x27d0d3 VCMPPD	$0x1,%ZMM4,%ZMM0,%K5

0x27d0da VCMPPD	$0x1,%ZMM4,%ZMM1,%K6

0x27d0e1 KMOVW	%K0,0x60(%RSP)    [4]

0x27d0e7 VCMPPD	$0x1,%ZMM4,%ZMM3,%K0

0x27d0ee KMOVQ	%K0,0x100(%RSP)    [4]

0x27d0f8 KMOVW	%K6,0xc0(%RSP)    [4]

0x27d101 KMOVQ	%K5,%K0

0x27d106 KUNPCKBW	%K5,%K6,%K5

0x27d10a KMOVW	0x60(%RSP),%K7    [4]

0x27d110 KMOVD	0x50(%RSP),%K6    [4]

0x27d117 KUNPCKWD	%K5,%K7,%K5

0x27d11b KMOVQ	%K1,%K7

0x27d120 KMOVQ	0x100(%RSP),%K1    [4]

0x27d12a VPMOVM2W	%K5,%ZMM0

0x27d130 VPMOVM2W	%K1,%ZMM1

0x27d136 VINSERTI32X4	$0x3,%XMM1,%ZMM0,%ZMM0

0x27d13d VPMOVW2M	%ZMM0,%K5

0x27d143 KTESTD	%K6,%K5

0x27d148 KANDNB	%K4,%K0,%K5

0x27d14c KMOVW	0xc0(%RSP),%K0    [4]

0x27d155 KNOTB	%K4,%K4

0x27d159 KNOTB	%K3,%K6

0x27d15d KANDNB	%K3,%K0,%K0

0x27d161 KANDNB	%K7,%K1,%K3

0x27d165 KNOTB	%K7,%K1

0x27d169 KMOVW	%K0,0x50(%RSP)    [4]

0x27d16f KMOVW	0x60(%RSP),%K0    [4]

0x27d175 KANDNB	%K2,%K0,%K0

0x27d179 KNOTB	%K2,%K2

0x27d17d JE	27cd40

0x27d183 VADDPD	%ZMM6,%ZMM19,%ZMM23

0x27d189 VADDPD	%ZMM16,%ZMM20,%ZMM24

0x27d18f VADDPD	%ZMM18,%ZMM21,%ZMM25

0x27d195 VADDPD	%ZMM17,%ZMM22,%ZMM26

0x27d19b JMP	27cd40

/home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c: 1624 - 1628

--------------------------------------------------------------------------------

1624:                for(jj1 = A_diag_i[i1]+1; jj1 < A_diag_i[i1+1]; jj1++)

1625:                {

1626:                  i2 = A_diag_j[jj1];

1627:                  if((P_marker[i2] >= jj_begin_row || i2 == i) && (sgn*A_diag_data[jj1]) < 0)

1628:                    sum += A_diag_data[jj1];

Coverage (%)	Name	Module
►95.79+	__kmp_invoke_microtask	libomp.so
○	__kmp_invoke_task_func	libomp.so
○	__kmp_launch_thread	libomp.so
○	__kmp_launch_worker(void*)	libomp.so
○	start_thread	libc.so.6
○	__GI___clone3	libc.so.6
►2.45+	__kmp_invoke_microtask	libomp.so
○	__kmp_invoke_task_func	libomp.so
○	__kmp_launch_thread	libomp.so
○	__kmp_launch_worker(void*)	libomp.so

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

Path /

Metric	Value
CQA speedup if no scalar integer	1.75
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.54
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.03
Bottlenecks
Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source	par_lr_interp.c:1624-1628
Source loop unroll info	unrolled by 32
Source loop unroll confidence level	medium
Unroll/vectorization loop type	main
Unroll factor	32
CQA cycles	23.38
CQA cycles if no scalar integer	13.38
CQA cycles if FP arith vectorized	23.38
CQA cycles if fully vectorized	15.17
Front-end cycles	22.69
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	1.17
P4 cycles	1.17
P5 cycles	1.17
P6 cycles	23.38
P7 cycles	23.38
P8 cycles	23.38
P9 cycles	23.38
P10 cycles	19.00
P11 cycles	19.00
P12 cycles	19.00
P13 cycles	19.00
P14 cycles	6.00
P15 cycles	6.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	173.50
Nb uops	181.50
Nb loads	83.50
Nb stores	10.00
Nb stack references	7.50
FLOP/cycle	1.37
Nb FLOP add-sub	16.00
Nb FLOP mul	16.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	51.87
Bytes prefetched	0.00
Bytes loaded	1108.00
Bytes stored	109.00
Stride 0	1.00
Stride 1	1.50
Stride n	0.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	52.70
Vectorization ratio load	20.64
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	70.58
Vector-efficiency ratio all	37.45
Vector-efficiency ratio load	27.25
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	45.79

Metric	Value
CQA speedup if no scalar integer	1.67
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.47
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.01
Bottlenecks	P6, P7, P8, P9,
Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source	par_lr_interp.c:1624-1628
Source loop unroll info	unrolled by 32
Source loop unroll confidence level	medium
Unroll/vectorization loop type	main
Unroll factor	32
CQA cycles	24.25
CQA cycles if no scalar integer	14.50
CQA cycles if FP arith vectorized	24.25
CQA cycles if fully vectorized	16.50
Front-end cycles	24.00
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	1.33
P4 cycles	1.33
P5 cycles	1.33
P6 cycles	24.25
P7 cycles	24.25
P8 cycles	24.25
P9 cycles	24.25
P10 cycles	20.50
P11 cycles	20.50
P12 cycles	20.50
P13 cycles	20.50
P14 cycles	6.50
P15 cycles	6.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	184.00
Nb uops	192.00
Nb loads	86.00
Nb stores	11.00
Nb stack references	8.00
FLOP/cycle	2.64
Nb FLOP add-sub	32.00
Nb FLOP mul	32.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	56.99
Bytes prefetched	0.00
Bytes loaded	1268.00
Bytes stored	114.00
Stride 0	1.00
Stride 1	2.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	56.76
Vectorization ratio load	25.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	71.93
Vector-efficiency ratio all	42.45
Vector-efficiency ratio load	31.25
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	47.59

Metric	Value
CQA speedup if no scalar integer	1.73
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.53
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P6, P7, P8, P9,
Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source	par_lr_interp.c:1624-1628
Source loop unroll info	unrolled by 32
Source loop unroll confidence level	medium
Unroll/vectorization loop type	main
Unroll factor	32
CQA cycles	24.25
CQA cycles if no scalar integer	14.00
CQA cycles if FP arith vectorized	24.25
CQA cycles if fully vectorized	15.87
Front-end cycles	23.38
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	1.00
P4 cycles	1.00
P5 cycles	1.00
P6 cycles	24.25
P7 cycles	24.25
P8 cycles	24.25
P9 cycles	24.25
P10 cycles	19.50
P11 cycles	19.50
P12 cycles	19.50
P13 cycles	19.50
P14 cycles	6.50
P15 cycles	6.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	179.00
Nb uops	187.00
Nb loads	86.00
Nb stores	11.00
Nb stack references	8.00
FLOP/cycle	1.32
Nb FLOP add-sub	0.00
Nb FLOP mul	32.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	56.99
Bytes prefetched	0.00
Bytes loaded	1268.00
Bytes stored	114.00
Stride 0	1.00
Stride 1	2.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	55.14
Vectorization ratio load	25.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	71.93
Vector-efficiency ratio all	40.30
Vector-efficiency ratio load	31.25
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	47.59

Metric	Value
CQA speedup if no scalar integer	1.76
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.55
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.02
Bottlenecks	P6, P7, P8, P9,
Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source	par_lr_interp.c:1624-1628
Source loop unroll info	unrolled by 32
Source loop unroll confidence level	medium
Unroll/vectorization loop type	main
Unroll factor	32
CQA cycles	22.50
CQA cycles if no scalar integer	12.75
CQA cycles if FP arith vectorized	22.50
CQA cycles if fully vectorized	14.47
Front-end cycles	22.00
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	1.33
P4 cycles	1.33
P5 cycles	1.33
P6 cycles	22.50
P7 cycles	22.50
P8 cycles	22.50
P9 cycles	22.50
P10 cycles	18.50
P11 cycles	18.50
P12 cycles	18.50
P13 cycles	18.50
P14 cycles	5.50
P15 cycles	5.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	168.00
Nb uops	176.00
Nb loads	81.00
Nb stores	9.00
Nb stack references	7.00
FLOP/cycle	1.42
Nb FLOP add-sub	32.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	46.76
Bytes prefetched	0.00
Bytes loaded	948.00
Bytes stored	104.00
Stride 0	1.00
Stride 1	1.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	50.52
Vectorization ratio load	16.28
Vectorization ratio store	100.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	69.23
Vector-efficiency ratio all	34.92
Vector-efficiency ratio load	23.26
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	43.99

Metric	Value
CQA speedup if no scalar integer	1.84
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.63
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.05
Bottlenecks	P6, P7, P8, P9,
Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source	par_lr_interp.c:1624-1628
Source loop unroll info	unrolled by 32
Source loop unroll confidence level	medium
Unroll/vectorization loop type	main
Unroll factor	32
CQA cycles	22.50
CQA cycles if no scalar integer	12.25
CQA cycles if FP arith vectorized	22.50
CQA cycles if fully vectorized	13.83
Front-end cycles	21.38
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	1.00
P4 cycles	1.00
P5 cycles	1.00
P6 cycles	22.50
P7 cycles	22.50
P8 cycles	22.50
P9 cycles	22.50
P10 cycles	17.50
P11 cycles	17.50
P12 cycles	17.50
P13 cycles	17.50
P14 cycles	5.50
P15 cycles	5.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	163.00
Nb uops	171.00
Nb loads	81.00
Nb stores	9.00
Nb stack references	7.00
FLOP/cycle	0.00
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	46.76
Bytes prefetched	0.00
Bytes loaded	948.00
Bytes stored	104.00
Stride 0	1.00
Stride 1	1.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	48.39
Vectorization ratio load	16.28
Vectorization ratio store	100.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	69.23
Vector-efficiency ratio all	32.12
Vector-efficiency ratio load	23.26
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	43.99

Path /

Average path: Display a virtual path defined by average values of all real paths

Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source file and lines	par_lr_interp.c:1624-1628
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1624-1628.

It is main loop of related source loop which is unrolled by 32 (including vectorization).
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 37% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 23.38 to 15.17 cycles (1.54x speedup).

Details

52% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

20% of SSE/AVX loads are used in vector version.
70% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

FMA

Presence of both ADD/SUB and MUL operations.

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible. For instance a + b*c is a valid FMA (MUL then ADD). However (a+b)* c cannot be translated into an FMA (ADD then MUL).

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

KMOVD: 0 occurrences
KMOVQ: 0 occurrences
KMOVW: 0 occurrences
KORTESTD: 0 occurrences
KTESTD: 0 occurrences
VPCMPEQQ: 0 occurrences

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Irregular (variable stride) or indirect: 1 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Type of elements and instruction set

6 AVX-512 instructions are processing arithmetic or math operations on double precision FP elements in vector mode (eight at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 32 FP arithmetical operations:

16: addition or subtraction
16: multiply

The binary loop is loading 1108 bytes (138 double precision FP elements). The binary loop is storing 109 bytes (13 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.03 FP operations per loaded or stored byte.

General properties

nb instructions	173.50
nb uops	181.50
loop length	1050
used x86 registers	6.50
used mmx registers	0
used xmm registers	26
used ymm registers	8
used zmm registers	16.50
nb stack references	7.50
ADD-SUB / MUL ratio	1.00

Front-end

MACRO FUSION NOT POSSIBLE

micro-operation queue	22.69 cycles
front end	22.69 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.67	0.67	0.67	1.17	1.17	1.17	23.38	23.38	23.38	23.38	19.00	19.00	19.00	19.00	6.00	6.00
cycles	0.67	0.67	0.67	1.17	1.17	1.17	23.38	23.38	23.38	23.38	19.00	19.00	19.00	19.00	6.00	6.00

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	22.69
Dispatch	23.38
Data deps.	0.00
Overall L1	23.38

Vectorization ratios

INT

all	42%
load	18%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	64%

all	77%
load	22%
store	100%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	52%
load	20%
store	100%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	70%

Vector efficiency ratios

INT

all	26%
load	25%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	33%

all	63%
load	32%
store	25%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	37%
load	27%
store	25%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	45%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 23.38 cycles. At this rate:

18% of peak load performance is reached (47.21 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (4.66 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source file and lines	par_lr_interp.c:1624-1628
Module	exec

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 24.25 to 14.50 cycles (1.67x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Your loop is partially vectorized. Only 42% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 24.25 to 16.50 cycles (1.47x speedup).

Details

56% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

25% of SSE/AVX loads are used in vector version.
71% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

Workaround

Read less array elements
Write less array elements
Provide more information to your compiler:
- hardcode the bounds of the corresponding 'for' loop
- use the 'restrict' C99 keyword

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

FMA

Presence of both ADD/SUB and MUL operations.

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible. For instance a + b*c is a valid FMA (MUL then ADD). However (a+b)* c cannot be translated into an FMA (ADD then MUL).

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

KMOVD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVQ: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVW: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1624
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KORTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
VPCMPEQQ: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Irregular (variable stride) or indirect: 1 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Vector unaligned load/store instructions

Detected 4 optimal vector unaligned load/store instructions.

Details

VMOVUPD: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627

Workaround

Use vector aligned instructions:

align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Type of elements and instruction set

12 AVX-512 instructions are processing arithmetic or math operations on double precision FP elements in vector mode (eight at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 64 FP arithmetical operations:

32: addition or subtraction
32: multiply

The binary loop is loading 1268 bytes (158 double precision FP elements). The binary loop is storing 114 bytes (14 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.05 FP operations per loaded or stored byte.

General properties

nb instructions	184
nb uops	192
loop length	1120
used x86 registers	7
used mmx registers	0
used xmm registers	26
used ymm registers	8
used zmm registers	18
nb stack references	8
ADD-SUB / MUL ratio	1.00

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	24.00 cycles
front end	24.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.67	0.67	0.67	1.33	1.33	1.33	24.25	24.25	24.25	24.25	20.50	20.50	20.50	20.50	6.50	6.50
cycles	0.67	0.67	0.67	1.33	1.33	1.33	24.25	24.25	24.25	24.25	20.50	20.50	20.50	20.50	6.50	6.50

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	24.00
Dispatch	24.25
Data deps.	0.00
Overall L1	24.25

Vectorization ratios

INT

all	43%
load	18%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	64%

all	83%
load	45%
store	100%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	56%
load	25%
store	100%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	71%

Vector efficiency ratios

INT

all	26%
load	25%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	33%

all	73%
load	52%
store	25%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	42%
load	31%
store	25%
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	47%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 24.25 cycles. At this rate:

20% of peak load performance is reached (52.29 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (4.70 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 27cd93

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
KORB %K5,%K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K0,%K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K3,%K1,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
ADD $0x20,%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPD %ZMM6,%ZMM23{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KMOVW 0x50(%RSP),%K4	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VMOVAPD %ZMM18,%ZMM25{%K2}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM17,%ZMM26{%K1}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM23,%ZMM6	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM25,%ZMM18	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM26,%ZMM17	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KORB %K4,%K6,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVAPD %ZMM16,%ZMM24{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM24,%ZMM16	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
CMP %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 27d1a0 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x23e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x88(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x90(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x98(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM25,%XMM30,%XMM30	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVAPS %XMM0,0x60(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x100(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xc0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x48(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x130(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x50(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x58(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x240(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVDQA64 0x50(%RSP),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0x130(%RSP),%XMM26,%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ 0x240(%RSP),%XMM23,%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM24	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x60(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ 0x60(%RSP),%XMM25,%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM21	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x68(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM25,%YMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVDQA64 0x100(%RSP),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0xc0(%RSP),%XMM30,%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM22	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x70(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM24,%XMM21,%XMM21	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM23,%YMM21,%YMM21	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM19	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x78(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM26,%YMM26	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM25,%ZMM26,%ZMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM25,%K4																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM20	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x80(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM22,%XMM19,%XMM19	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM20,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM19,%YMM2,%YMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM21,%ZMM2,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM2,%K1																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM7	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM4,%XMM11,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM7,%XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM4,%YMM1,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM5	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x40(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM3,%XMM10,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM31	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM5,%XMM31,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM1,%ZMM3,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM1,%K2																				vect (100.0%)
VMOVDQA64 0x380(%RSP),%ZMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VPCMPEQQ -0xc0(%RCX,%R8,8),%ZMM1,%K3	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x80(%RCX,%R8,8),%ZMM1,%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x40(%RCX,%R8,8),%ZMM1,%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ (%RCX,%R8,8),%ZMM1,%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM8,%XMM14,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
KORB %K4,%K3,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K1,%K6,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K2,%K7,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVQ (%R15,%RDX,8),%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM12,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM0,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM27	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM28	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV (%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM15,%XMM27,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM29	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ %XMM28,%XMM29,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM0,%ZMM3,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM0,%K5																				vect (100.0%)
KORB %K5,%K0,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K4,%K3,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKWD %K0,%K2,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KORTESTD %K0,%K0	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KMOVD %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
JE 27d0f8 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x2338>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VMOVUPD -0x40(%R10,%R8,8),%ZMM21{%K2}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVAPD 0x340(%RSP),%ZMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVUPD -0xc0(%R10,%R8,8),%ZMM19{%K4}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVUPD -0x80(%R10,%R8,8),%ZMM20{%K3}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVUPD (%R10,%R8,8),%ZMM22{%K1}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VPXOR %XMM4,%XMM4,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.25	vect (25.0%)
VMULPD %ZMM3,%ZMM21,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VMULPD %ZMM3,%ZMM19,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VMULPD %ZMM3,%ZMM20,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VMULPD %ZMM3,%ZMM22,%ZMM3	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VCMPPD $0x1,%ZMM4,%ZMM2,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VCMPPD $0x1,%ZMM4,%ZMM0,%K5	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VCMPPD $0x1,%ZMM4,%ZMM1,%K6	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
KMOVW %K0,0x60(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
VCMPPD $0x1,%ZMM4,%ZMM3,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
KMOVQ %K0,0x100(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVW %K6,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVQ %K5,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K5,%K6,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0x60(%RSP),%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KMOVD 0x50(%RSP),%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KUNPCKWD %K5,%K7,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ %K1,%K7	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ 0x100(%RSP),%K1	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VPMOVM2W %K5,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K5	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KTESTD %K6,%K5	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KANDNB %K4,%K0,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0xc0(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KNOTB %K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K3,%K6	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K3,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K7,%K1,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K7,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVW 0x60(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KANDNB %K2,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
JE 27cd40 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x1f80>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VADDPD %ZMM6,%ZMM19,%ZMM23	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPD %ZMM16,%ZMM20,%ZMM24	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPD %ZMM18,%ZMM21,%ZMM25	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPD %ZMM17,%ZMM22,%ZMM26	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
JMP 27cd40 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x1f80>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	1	N/A

Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source file and lines	par_lr_interp.c:1624-1628
Module	exec

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 24.25 to 14.00 cycles (1.73x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Your loop is partially vectorized. Only 40% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 24.25 to 15.87 cycles (1.53x speedup).

Details

55% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

25% of SSE/AVX loads are used in vector version.
71% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

Workaround

Read less array elements
Write less array elements
Provide more information to your compiler:
- hardcode the bounds of the corresponding 'for' loop
- use the 'restrict' C99 keyword

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

KMOVD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVQ: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVW: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1624
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KORTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
VPCMPEQQ: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Irregular (variable stride) or indirect: 1 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Vector unaligned load/store instructions

Detected 4 optimal vector unaligned load/store instructions.

Details

VMOVUPD: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627

Workaround

Use vector aligned instructions:

align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Type of elements and instruction set

8 AVX-512 instructions are processing arithmetic or math operations on double precision FP elements in vector mode (eight at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 32 FP arithmetical operations:

32: multiply

The binary loop is loading 1268 bytes (158 double precision FP elements). The binary loop is storing 114 bytes (14 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.02 FP operations per loaded or stored byte.

General properties

nb instructions	179
nb uops	187
loop length	1091
used x86 registers	7
used mmx registers	0
used xmm registers	26
used ymm registers	8
used zmm registers	18
nb stack references	8

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	23.38 cycles
front end	23.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.67	0.67	0.67	1.00	1.00	1.00	24.25	24.25	24.25	24.25	19.50	19.50	19.50	19.50	6.50	6.50
cycles	0.67	0.67	0.67	1.00	1.00	1.00	24.25	24.25	24.25	24.25	19.50	19.50	19.50	19.50	6.50	6.50

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	23.38
Dispatch	24.25
Data deps.	0.00
Overall L1	24.25

Vectorization ratios

INT

all	43%
load	18%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	64%

all	81%
load	45%
store	100%
mul	100%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	55%
load	25%
store	100%
mul	100%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	71%

Vector efficiency ratios

INT

all	26%
load	25%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	33%

all	70%
load	52%
store	25%
mul	100%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	40%
load	31%
store	25%
mul	100%
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	47%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 24.25 cycles. At this rate:

20% of peak load performance is reached (52.29 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (4.70 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 27cd93

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
KORB %K5,%K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K0,%K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K3,%K1,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
ADD $0x20,%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPD %ZMM6,%ZMM23{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KMOVW 0x50(%RSP),%K4	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VMOVAPD %ZMM18,%ZMM25{%K2}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM17,%ZMM26{%K1}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM23,%ZMM6	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM25,%ZMM18	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM26,%ZMM17	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KORB %K4,%K6,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVAPD %ZMM16,%ZMM24{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM24,%ZMM16	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
CMP %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 27d1a0 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x23e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x88(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x90(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x98(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM25,%XMM30,%XMM30	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVAPS %XMM0,0x60(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x100(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xc0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x48(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x130(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x50(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x58(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x240(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVDQA64 0x50(%RSP),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0x130(%RSP),%XMM26,%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ 0x240(%RSP),%XMM23,%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM24	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x60(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ 0x60(%RSP),%XMM25,%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM21	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x68(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM25,%YMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVDQA64 0x100(%RSP),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0xc0(%RSP),%XMM30,%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM22	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x70(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM24,%XMM21,%XMM21	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM23,%YMM21,%YMM21	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM19	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x78(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM26,%YMM26	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM25,%ZMM26,%ZMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM25,%K4																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM20	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x80(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM22,%XMM19,%XMM19	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM20,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM19,%YMM2,%YMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM21,%ZMM2,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM2,%K1																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM7	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM4,%XMM11,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM7,%XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM4,%YMM1,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM5	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x40(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM3,%XMM10,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM31	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM5,%XMM31,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM1,%ZMM3,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM1,%K2																				vect (100.0%)
VMOVDQA64 0x380(%RSP),%ZMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VPCMPEQQ -0xc0(%RCX,%R8,8),%ZMM1,%K3	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x80(%RCX,%R8,8),%ZMM1,%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x40(%RCX,%R8,8),%ZMM1,%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ (%RCX,%R8,8),%ZMM1,%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM8,%XMM14,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
KORB %K4,%K3,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K1,%K6,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K2,%K7,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVQ (%R15,%RDX,8),%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM12,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM0,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM27	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM28	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV (%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM15,%XMM27,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM29	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ %XMM28,%XMM29,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM0,%ZMM3,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM0,%K5																				vect (100.0%)
KORB %K5,%K0,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K4,%K3,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKWD %K0,%K2,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KORTESTD %K0,%K0	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KMOVD %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
JE 27d0f8 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x2338>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VMOVUPD -0x40(%R10,%R8,8),%ZMM21{%K2}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVAPD 0x340(%RSP),%ZMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVUPD -0xc0(%R10,%R8,8),%ZMM19{%K4}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVUPD -0x80(%R10,%R8,8),%ZMM20{%K3}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VMOVUPD (%R10,%R8,8),%ZMM22{%K1}{z}	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VPXOR %XMM4,%XMM4,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.25	vect (25.0%)
VMULPD %ZMM3,%ZMM21,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VMULPD %ZMM3,%ZMM19,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VMULPD %ZMM3,%ZMM20,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VMULPD %ZMM3,%ZMM22,%ZMM3	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50	vect (100.0%)
VCMPPD $0x1,%ZMM4,%ZMM2,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VCMPPD $0x1,%ZMM4,%ZMM0,%K5	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VCMPPD $0x1,%ZMM4,%ZMM1,%K6	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
KMOVW %K0,0x60(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
VCMPPD $0x1,%ZMM4,%ZMM3,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
KMOVQ %K0,0x100(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVW %K6,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVQ %K5,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K5,%K6,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0x60(%RSP),%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KMOVD 0x50(%RSP),%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KUNPCKWD %K5,%K7,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ %K1,%K7	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ 0x100(%RSP),%K1	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VPMOVM2W %K5,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K5	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KTESTD %K6,%K5	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KANDNB %K4,%K0,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0xc0(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KNOTB %K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K3,%K6	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K3,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K7,%K1,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K7,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVW 0x60(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KANDNB %K2,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
JE 27cd40 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x1f80>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source file and lines	par_lr_interp.c:1624-1628
Module	exec

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 22.50 to 12.75 cycles (1.76x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Your loop is partially vectorized. Only 34% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 22.50 to 14.47 cycles (1.55x speedup).

Details

50% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

16% of SSE/AVX loads are used in vector version.
69% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

Workaround

Read less array elements
Write less array elements
Provide more information to your compiler:
- hardcode the bounds of the corresponding 'for' loop
- use the 'restrict' C99 keyword

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

KMOVD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVQ: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVW: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1624
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KORTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
VPCMPEQQ: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Irregular (variable stride) or indirect: 1 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Type of elements and instruction set

4 AVX-512 instructions are processing arithmetic or math operations on double precision FP elements in vector mode (eight at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 32 FP arithmetical operations:

32: addition or subtraction

The binary loop is loading 948 bytes (118 double precision FP elements). The binary loop is storing 104 bytes (13 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.03 FP operations per loaded or stored byte.

General properties

nb instructions	168
nb uops	176
loop length	1009
used x86 registers	6
used mmx registers	0
used xmm registers	26
used ymm registers	8
used zmm registers	17
nb stack references	7

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	22.00 cycles
front end	22.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.67	0.67	0.67	1.33	1.33	1.33	22.50	22.50	22.50	22.50	18.50	18.50	18.50	18.50	5.50	5.50
cycles	0.67	0.67	0.67	1.33	1.33	1.33	22.50	22.50	22.50	22.50	18.50	18.50	18.50	18.50	5.50	5.50

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	22.00
Dispatch	22.50
Data deps.	0.00
Overall L1	22.50

Vectorization ratios

INT

all	42%
load	18%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	63%

all	75%
load	0%
store	100%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	50%
load	16%
store	100%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	69%

Vector efficiency ratios

INT

all	26%
load	25%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	33%

all	59%
load	12%
store	25%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	34%
load	23%
store	25%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	43%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 22.50 cycles. At this rate:

16% of peak load performance is reached (42.13 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (4.62 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 27cd93

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
KORB %K5,%K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K0,%K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K3,%K1,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
ADD $0x20,%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPD %ZMM6,%ZMM23{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KMOVW 0x50(%RSP),%K4	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VMOVAPD %ZMM18,%ZMM25{%K2}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM17,%ZMM26{%K1}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM23,%ZMM6	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM25,%ZMM18	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM26,%ZMM17	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KORB %K4,%K6,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVAPD %ZMM16,%ZMM24{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM24,%ZMM16	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
CMP %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 27d1a0 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x23e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x88(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x90(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x98(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM25,%XMM30,%XMM30	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVAPS %XMM0,0x60(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x100(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xc0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x48(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x130(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x50(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x58(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x240(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVDQA64 0x50(%RSP),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0x130(%RSP),%XMM26,%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ 0x240(%RSP),%XMM23,%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM24	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x60(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ 0x60(%RSP),%XMM25,%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM21	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x68(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM25,%YMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVDQA64 0x100(%RSP),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0xc0(%RSP),%XMM30,%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM22	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x70(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM24,%XMM21,%XMM21	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM23,%YMM21,%YMM21	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM19	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x78(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM26,%YMM26	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM25,%ZMM26,%ZMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM25,%K4																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM20	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x80(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM22,%XMM19,%XMM19	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM20,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM19,%YMM2,%YMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM21,%ZMM2,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM2,%K1																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM7	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM4,%XMM11,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM7,%XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM4,%YMM1,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM5	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x40(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM3,%XMM10,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM31	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM5,%XMM31,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM1,%ZMM3,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM1,%K2																				vect (100.0%)
VMOVDQA64 0x380(%RSP),%ZMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VPCMPEQQ -0xc0(%RCX,%R8,8),%ZMM1,%K3	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x80(%RCX,%R8,8),%ZMM1,%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x40(%RCX,%R8,8),%ZMM1,%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ (%RCX,%R8,8),%ZMM1,%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM8,%XMM14,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
KORB %K4,%K3,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K1,%K6,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K2,%K7,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVQ (%R15,%RDX,8),%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM12,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM0,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM27	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM28	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV (%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM15,%XMM27,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM29	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ %XMM28,%XMM29,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM0,%ZMM3,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM0,%K5																				vect (100.0%)
KORB %K5,%K0,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K4,%K3,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKWD %K0,%K2,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KORTESTD %K0,%K0	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KMOVD %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
JE 27d0f8 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x2338>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
KMOVW %K6,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVQ %K5,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K5,%K6,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0x60(%RSP),%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KMOVD 0x50(%RSP),%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KUNPCKWD %K5,%K7,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ %K1,%K7	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ 0x100(%RSP),%K1	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VPMOVM2W %K5,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K5	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KTESTD %K6,%K5	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KANDNB %K4,%K0,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0xc0(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KNOTB %K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K3,%K6	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K3,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K7,%K1,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K7,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVW 0x60(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KANDNB %K2,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
JE 27cd40 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x1f80>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
VADDPD %ZMM6,%ZMM19,%ZMM23	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPD %ZMM16,%ZMM20,%ZMM24	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPD %ZMM18,%ZMM21,%ZMM25	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPD %ZMM17,%ZMM22,%ZMM26	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
JMP 27cd40 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x1f80>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	1	N/A

Function	hypre_BoomerAMGBuildExtPIInterp.omp_outlined
Source file and lines	par_lr_interp.c:1624-1628
Module	exec

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 22.50 to 12.25 cycles (1.84x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Your loop is poorly vectorized. Only 32% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 22.50 to 13.83 cycles (1.63x speedup).

Details

48% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

16% of SSE/AVX loads are used in vector version.
69% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 22.50 to 21.38 cycles (1.05x speedup).

Workaround

Read less array elements
Write less array elements
Provide more information to your compiler:
- hardcode the bounds of the corresponding 'for' loop
- use the 'restrict' C99 keyword

Masked instructions

Detected masked instructions.

Details

Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround

If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

KMOVD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVQ: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KMOVW: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1624
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KORTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
KTESTD: 1 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
VPCMPEQQ: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_lr_interp.c:1627

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Irregular (variable stride) or indirect: 1 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

The binary loop does not contain any FP arithmetical operations. The binary loop is loading 948 bytes. The binary loop is storing 104 bytes.

General properties

nb instructions	163
nb uops	171
loop length	980
used x86 registers	6
used mmx registers	0
used xmm registers	26
used ymm registers	8
used zmm registers	13
nb stack references	7

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	21.38 cycles
front end	21.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.67	0.67	0.67	1.00	1.00	1.00	22.50	22.50	22.50	22.50	17.50	17.50	17.50	17.50	5.50	5.50
cycles	0.67	0.67	0.67	1.00	1.00	1.00	22.50	22.50	22.50	22.50	17.50	17.50	17.50	17.50	5.50	5.50

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	21.38
Dispatch	22.50
Data deps.	0.00
Overall L1	22.50

Vectorization ratios

INT

all	42%
load	18%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	63%

all	70%
load	0%
store	100%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	48%
load	16%
store	100%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	69%

Vector efficiency ratios

INT

all	26%
load	25%
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	33%

all	51%
load	12%
store	25%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	100%

INT+FP

all	32%
load	23%
store	25%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	43%

Cycles and memory resources usage

Detected masked instructions: assuming all mask elements are active. Assuming all data fit into the L1 cache, each iteration of the binary loop takes 22.50 cycles. At this rate:

16% of peak load performance is reached (42.13 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
3% of peak store performance is reached (4.62 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 27cd93

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
KORB %K5,%K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K0,%K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K3,%K1,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
ADD $0x20,%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPD %ZMM6,%ZMM23{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KMOVW 0x50(%RSP),%K4	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VMOVAPD %ZMM18,%ZMM25{%K2}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM17,%ZMM26{%K1}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM23,%ZMM6	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM25,%ZMM18	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM26,%ZMM17	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
KORB %K4,%K6,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVAPD %ZMM16,%ZMM24{%K4}	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPD %ZMM24,%ZMM16	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
CMP %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 27d1a0 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x23e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x88(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x90(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x98(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM25,%XMM30,%XMM30	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVAPS %XMM0,0x60(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xa8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xb8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x100(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0xc0(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x48(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x130(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVSD (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x50(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x58(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVAPS %XMM0,0x240(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMOVDQA64 0x50(%RSP),%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0x130(%RSP),%XMM26,%XMM26	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ 0x240(%RSP),%XMM23,%XMM23	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM24	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x60(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ 0x60(%RSP),%XMM25,%XMM25	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM21	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x68(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM25,%YMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVDQA64 0x100(%RSP),%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VPUNPCKLQDQ 0xc0(%RSP),%XMM30,%XMM30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM22	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x70(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM24,%XMM21,%XMM21	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM23,%YMM21,%YMM21	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM19	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x78(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI32X4 $0x1,%XMM30,%YMM26,%YMM26	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM25,%ZMM26,%ZMM25	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM25,%K4																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM20	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x80(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM22,%XMM19,%XMM19	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM20,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI32X4 $0x1,%XMM19,%YMM2,%YMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM21,%ZMM2,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM2,%K1																				vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM7	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM4,%XMM11,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM7,%XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM4,%YMM1,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM5	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV -0x40(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM3,%XMM10,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM31	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x38(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x30(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM5,%XMM31,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x28(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VINSERTI64X4 $0x1,%YMM1,%ZMM3,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM1,%K2																				vect (100.0%)
VMOVDQA64 0x380(%RSP),%ZMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
VPCMPEQQ -0xc0(%RCX,%R8,8),%ZMM1,%K3	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x80(%RCX,%R8,8),%ZMM1,%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ -0x40(%RCX,%R8,8),%ZMM1,%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VPCMPEQQ (%RCX,%R8,8),%ZMM1,%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0.50	0.50	0	0	0	6	0.50	vect (100.0%)
VMOVQ (%R15,%RDX,8),%XMM12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x20(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM8,%XMM14,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x18(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
KORB %K4,%K3,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K1,%K6,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KORB %K2,%K7,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VMOVQ (%R15,%RDX,8),%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x10(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM12,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM0,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VMOVQ (%R15,%RDX,8),%XMM27	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV 0x8(%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VMOVQ (%R15,%RDX,8),%XMM28	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
MOV (%RCX,%R8,8),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VPUNPCKLQDQ %XMM15,%XMM27,%XMM3	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VMOVQ (%R15,%RDX,8),%XMM29	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (12.5%)
VPUNPCKLQDQ %XMM28,%XMM29,%XMM4	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
VINSERTI128 $0x1,%XMM3,%YMM4,%YMM3	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VINSERTI64X4 $0x1,%YMM0,%ZMM3,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (50.0%)
VPCMPNLTQ %ZMM9,%ZMM0,%K5																				vect (100.0%)
KORB %K5,%K0,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K4,%K3,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKWD %K0,%K2,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KORTESTD %K0,%K0	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KMOVD %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
JE 27d0f8 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x2338>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A
KMOVW %K6,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVQ %K5,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KUNPCKBW %K5,%K6,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0x60(%RSP),%K7	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KMOVD 0x50(%RSP),%K6	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KUNPCKWD %K5,%K7,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ %K1,%K7	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVQ 0x100(%RSP),%K1	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
VPMOVM2W %K5,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VPMOVM2W %K1,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (6.3%)
VINSERTI32X4 $0x3,%XMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	vect (25.0%)
VPMOVW2M %ZMM0,%K5	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	6	0.50	vect (6.3%)
KTESTD %K6,%K5	2	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0.50	0.50	4	0.50	N/A
KANDNB %K4,%K0,%K5	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW 0xc0(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KNOTB %K4,%K4	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K3,%K6	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K3,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KANDNB %K7,%K1,%K3	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K7,%K1	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KMOVW %K0,0x50(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	6	0.50	N/A
KMOVW 0x60(%RSP),%K0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	0	0	6	0.50	N/A
KANDNB %K2,%K0,%K0	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
KNOTB %K2,%K2	1	0	0	0	0	0	0	0	0	0	0	0.50	0	0	0.50	0	0	1	0.50	N/A
JE 27cd40 <hypre_BoomerAMGBuildExtPIInterp.omp_outlined+0x1f80>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Report Configuration

Vectorization

Details

Workaround

Masked instructions

Details

Workaround

FMA

Workaround

Complex instructions

Details

Slow data structures access

Details

Workaround

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Masked instructions

Details

Workaround

FMA

Workaround

Complex instructions

Details

Slow data structures access

Details

Workaround

Vector unaligned load/store instructions

Details

Workaround

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Masked instructions

Details

Workaround

Complex instructions

Details

Slow data structures access

Details

Workaround

Vector unaligned load/store instructions

Details

Workaround

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

General properties

Front-end

Back-end