OV - exec - Loop 2666

	Loop Id: 2666	Module: exec	Source: ams.c:3662-3682 [...]	Coverage: NA%

0x4795d4 ADRP	X0,

0x4795d8 ADD	X0, X0, #920

0x4795dc ORR	W1, WZR, W20

0x4795e0 BL	40f910

0x4795e4 LDP	X20, X19, [SP, #160]    [1]

0x4795e8 LDP	X22, X21, [SP, #144]    [1]

0x4795ec LDP	X24, X23, [SP, #128]    [1]

0x4795f0 LDP	X26, X25, [SP, #112]    [1]

0x4795f4 LDP	X28, X27, [SP, #96]    [1]

0x4795f8 LDP	X29, X30, [SP, #80]    [1]

0x4795fc ADD	SP, SP, #176

0x479600 RET

0x479604 LDP	X16, X17, [X29, #112]    [16]

0x479608 LDP	X8, X0, [X29, #144]    [16]

0x47960c LDP	X18, X4, [X29, #128]    [16]

0x479610 ADD	X1, X1, #1

0x479614 LDP	X14, X15, [X29, #96]    [16]

0x479618 LDR	X10, [X24]    [9]

0x47961c ADD	X6, X10, #16

0x479620 LDR	X13, [X21]    [5]

0x479624 LDR	X16, [X16]    [12]

0x479628 LDR	X17, [X17]    [13]

0x47962c LDR	X11, [X23]    [7]

0x479630 LDR	X12, [X22]    [6]

0x479634 LDR	X14, [X14]    [8]

0x479638 LDR	X15, [X15]    [10]

0x47963c LDR	X18, [X18]    [14]

0x479640 LDR	X0, [X0]    [11]

0x479644 LDR	X4, [X4]    [17]

0x479648 ADD	X5, X13, #16

0x47964c ADD	X2, X13, #8

0x479650 ADD	X3, X10, #8

0x479654 ADD	X7, X16, #8

0x479658 STP	X5, X6, [SP, #32]    [2]

0x47965c ADD	X21, X17, #8

0x479660 ADD	X6, X17, #16

0x479664 ADD	X5, X16, #16

0x479668 STP	X5, X6, [SP, #16]    [2]

0x47966c B	4796a8

0x47969c ORR	X9, XZR, X24

0x4796a0 CMP	X24, X1

0x4796a4 B.EQ	4795d4

0x4796a8 LDR	X5, [X11, X9,LSL #3]    [4]

0x4796ac LDR	D1, [X19]    [15]

0x4796b0 LDR	D0, [X10, X5,LSL #3]    [3]

0x4796b4 FCMP	D0, D1

0x4796b8 B.EQ	4796e0

0x4796e0 ADD	X24, X9, #1

0x4796e4 B	47969c

/home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c: 3662 - 3682

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

3662: #pragma omp parallel for private(i,ii,jj,res) HYPRE_SMP_SCHEDULE

3663: #endif

3664:       for (i = 0; i < n; i++)

3665:       {

3666:          /*-----------------------------------------------------------

3667:           * If diagonal is nonzero, relax point i; otherwise, skip it.

3668:           *-----------------------------------------------------------*/

3669:          if (A_diag_data[A_diag_i[i]] != zero)

[...]

3682:             u_data[i] += (relax_weight*res)/l1_norms[i];

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	8.67
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.08
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.41
Bottlenecks	P12, P13, P14,
Function	hypre_ParCSRRelaxThreads.omp_outlined.28
Source	ams.c:3662-3669,ams.c:3682-3682
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	8.67
CQA cycles if no scalar integer	1.00
CQA cycles if FP arith vectorized	8.67
CQA cycles if fully vectorized	8.00
Front-end cycles	6.13
P0 cycles	3.00
P1 cycles	3.00
P2 cycles	2.92
P3 cycles	2.92
P4 cycles	2.83
P5 cycles	2.83
P6 cycles	2.75
P7 cycles	2.75
P8 cycles	1.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	8.67
P13 cycles	8.67
P14 cycles	8.67
P15 cycles	1.00
P16 cycles	1.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	49.00
Nb uops	49.00
Nb loads	NA
Nb stores	2.00
Nb stack references	0.00
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	35.08
Bytes prefetched	0.00
Bytes loaded	272.00
Bytes stored	32.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	60.29
Vector-efficiency ratio load	66.67
Vector-efficiency ratio store	100.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	50.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	41.67

Metric	Value
CQA speedup if no scalar integer	8.67
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.08
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.41
Bottlenecks	P12, P13, P14,
Function	hypre_ParCSRRelaxThreads.omp_outlined.28
Source	ams.c:3662-3669,ams.c:3682-3682
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	8.67
CQA cycles if no scalar integer	1.00
CQA cycles if FP arith vectorized	8.67
CQA cycles if fully vectorized	8.00
Front-end cycles	6.13
P0 cycles	3.00
P1 cycles	3.00
P2 cycles	2.92
P3 cycles	2.92
P4 cycles	2.83
P5 cycles	2.83
P6 cycles	2.75
P7 cycles	2.75
P8 cycles	1.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	8.67
P13 cycles	8.67
P14 cycles	8.67
P15 cycles	1.00
P16 cycles	1.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	49.00
Nb uops	49.00
Nb loads	NA
Nb stores	2.00
Nb stack references	0.00
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	35.08
Bytes prefetched	0.00
Bytes loaded	272.00
Bytes stored	32.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	60.29
Vector-efficiency ratio load	66.67
Vector-efficiency ratio store	100.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	50.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	41.67

Path /

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

Function	hypre_ParCSRRelaxThreads.omp_outlined.28
Source file and lines	ams.c:3662-3682
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3662-3669,3682.

The related source loop is not unrolled or unrolled with no peel/tail loop.
Warnings:

Ignoring paths for analysis
Failed to get the number of paths
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

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 8.67 to 1.00 cycles (8.67x speedup).

Workaround

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

Vectorization

Your loop is not vectorized. Only 60% of vector register length is used (average across all VPU instructions). By vectorizing your loop, you can lower the cost of an iteration from 8.67 to 8.00 cycles (1.08x speedup).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a 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)

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 8.67 to 0.50 cycles (17.33x speedup).

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 272 bytes. The binary loop is storing 32 bytes.

General properties

nb instructions	49
loop length	196
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

6.13 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	3.00	3.00	2.92	2.92	2.83	2.83	2.75	2.75	1.00	0.00	0.00	0.00	8.67	8.67	8.67	1.00	1.00
cycles	3.00	3.00	2.92	2.92	2.83	2.83	2.75	2.75	1.00	0.00	0.00	0.00	8.67	8.67	8.67	1.00	1.00

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6: ALU
P7: ALU
P8 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P9 (128 bits): VPU, ALU, FP store data
P10 (128 bits): VPU, ALU, DIV/SQRT
P11 (128 bits): ALU, VPU
P12 (256 bits): store address, load
P13 (256 bits): store address, load
P14 (256 bits): load
P15 (64 bits): store data
P16 (64 bits): store data

Cycles executing div or sqrt instructions

Cycles summary

Front-end	6.13
Overall L1	8.67

Vectorization ratios

INT

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

all	0%
load	NA (no load vectorizable/vectorized instructions)
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)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

INT+FP

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.67 cycles. At this rate:

32% of peak load performance is reached (31.38 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
7% of peak store performance is reached (3.69 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 47969c

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P12	P13	P14	P15	P16	Latency	Recip. throughput	Vectorization
ADRP X0, <4de5d4>	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD X0, X0, #920	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ORR W1, WZR, W20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (25.0%)
BL 40f910 <@plt_start@+0x520>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X20, X19, [SP, #160]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X22, X21, [SP, #144]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X24, X23, [SP, #128]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X26, X25, [SP, #112]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
LDP X28, X27, [SP, #96]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
LDP X29, X30, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
ADD SP, SP, #176	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X16, X17, [X29, #112]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X8, X0, [X29, #144]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X18, X4, [X29, #128]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
ADD X1, X1, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
LDP X14, X15, [X29, #96]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDR X10, [X24]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD X6, X10, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
LDR X13, [X21]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDR X16, [X16]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X17, [X17]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X11, [X23]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X12, [X22]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDR X14, [X14]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X15, [X15]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X18, [X18]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X0, [X0]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X4, [X4]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD X5, X13, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD X2, X13, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X3, X10, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X7, X16, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
STP X5, X6, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
ADD X21, X17, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD X6, X17, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X5, X16, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
STP X5, X6, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
B 4796a8 <hypre_ParCSRRelaxThreads.omp_outlined.28+0x168>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR X9, XZR, X24	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
CMP X24, X1	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	1	0.33	scal (50.0%)
B.EQ 4795d4 <hypre_ParCSRRelaxThreads.omp_outlined.28+0x94>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X5, [X11, X9,LSL #3]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR D1, [X19]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (50.0%)
LDR D0, [X10, X5,LSL #3]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (50.0%)
FCMP D0, D1	1	0	0	0	0	0	0	0	0	1	0	0	0	0	0	2	1	scal (50.0%)
B.EQ 4796e0 <hypre_ParCSRRelaxThreads.omp_outlined.28+0x1a0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD X24, X9, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
B 47969c <hypre_ParCSRRelaxThreads.omp_outlined.28+0x15c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	hypre_ParCSRRelaxThreads.omp_outlined.28
Source file and lines	ams.c:3662-3682
Module	exec

Ignoring paths for analysis
Failed to get the number of paths
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

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 8.67 to 1.00 cycles (8.67x speedup).

Workaround

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

Vectorization

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a 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)

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 8.67 to 0.50 cycles (17.33x speedup).

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 272 bytes. The binary loop is storing 32 bytes.

General properties

nb instructions	49
loop length	196
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

6.13 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	3.00	3.00	2.92	2.92	2.83	2.83	2.75	2.75	1.00	0.00	0.00	0.00	8.67	8.67	8.67	1.00	1.00
cycles	3.00	3.00	2.92	2.92	2.83	2.83	2.75	2.75	1.00	0.00	0.00	0.00	8.67	8.67	8.67	1.00	1.00

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6: ALU
P7: ALU
P8 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P9 (128 bits): VPU, ALU, FP store data
P10 (128 bits): VPU, ALU, DIV/SQRT
P11 (128 bits): ALU, VPU
P12 (256 bits): store address, load
P13 (256 bits): store address, load
P14 (256 bits): load
P15 (64 bits): store data
P16 (64 bits): store data

Cycles executing div or sqrt instructions

Cycles summary

Front-end	6.13
Overall L1	8.67

Vectorization ratios

INT

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

all	0%
load	NA (no load vectorizable/vectorized instructions)
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)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

INT+FP

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.67 cycles. At this rate:

32% of peak load performance is reached (31.38 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
7% of peak store performance is reached (3.69 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 47969c

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P12	P13	P14	P15	P16	Latency	Recip. throughput	Vectorization
ADRP X0, <4de5d4>	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
ADD X0, X0, #920	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ORR W1, WZR, W20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (25.0%)
BL 40f910 <@plt_start@+0x520>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X20, X19, [SP, #160]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X22, X21, [SP, #144]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X24, X23, [SP, #128]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X26, X25, [SP, #112]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
LDP X28, X27, [SP, #96]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	scal (100.0%)
LDP X29, X30, [SP, #80]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
ADD SP, SP, #176	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X16, X17, [X29, #112]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X8, X0, [X29, #144]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDP X18, X4, [X29, #128]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
ADD X1, X1, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
LDP X14, X15, [X29, #96]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
LDR X10, [X24]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD X6, X10, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
LDR X13, [X21]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDR X16, [X16]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X17, [X17]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X11, [X23]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X12, [X22]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDR X14, [X14]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X15, [X15]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X18, [X18]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X0, [X0]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X4, [X4]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD X5, X13, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD X2, X13, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X3, X10, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X7, X16, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
STP X5, X6, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
ADD X21, X17, #8	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD X6, X17, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X5, X16, #16	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
STP X5, X6, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (100.0%)
B 4796a8 <hypre_ParCSRRelaxThreads.omp_outlined.28+0x168>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR X9, XZR, X24	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
CMP X24, X1	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	1	0.33	scal (50.0%)
B.EQ 4795d4 <hypre_ParCSRRelaxThreads.omp_outlined.28+0x94>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X5, [X11, X9,LSL #3]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR D1, [X19]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (50.0%)
LDR D0, [X10, X5,LSL #3]	1	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (50.0%)
FCMP D0, D1	1	0	0	0	0	0	0	0	0	1	0	0	0	0	0	2	1	scal (50.0%)
B.EQ 4796e0 <hypre_ParCSRRelaxThreads.omp_outlined.28+0x1a0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD X24, X9, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
B 47969c <hypre_ParCSRRelaxThreads.omp_outlined.28+0x15c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

▶Scalability Runs Description

Run 1x1	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_NUM_THREADS: 1OMP_PLACES: threads
Run 1x2	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 2OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x4	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 4OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x8	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 8OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x16	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 16OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x24	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 24OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x32	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 32OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x40	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 40OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x48	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 48OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x56	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 56OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x64	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 64OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x72	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 72OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x80	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 80OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x88	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 88OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x96	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 96OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads

(1x1) Efficiency	(1x1) Potential Speed-Up (%)	(1x2) Efficiency	(1x2) Potential Speed-Up (%)	(1x4) Efficiency	(1x4) Potential Speed-Up (%)	(1x8) Efficiency	(1x8) Potential Speed-Up (%)	(1x16) Efficiency	(1x16) Potential Speed-Up (%)	(1x24) Efficiency	(1x24) Potential Speed-Up (%)	(1x32) Efficiency	(1x32) Potential Speed-Up (%)	(1x40) Efficiency	(1x40) Potential Speed-Up (%)	(1x48) Efficiency	(1x48) Potential Speed-Up (%)	(1x56) Efficiency	(1x56) Potential Speed-Up (%)	(1x64) Efficiency	(1x64) Potential Speed-Up (%)	(1x72) Efficiency	(1x72) Potential Speed-Up (%)	(1x80) Efficiency	(1x80) Potential Speed-Up (%)	(1x88) Efficiency	(1x88) Potential Speed-Up (%)	(1x96) Efficiency	(1x96) Potential Speed-Up (%)
		1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0	1	0

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency

Run	Number of threads	Efficiency (ideal is 1)	Speedup	Ideal Speedup	Time (s)	Coverage (%)
1x1
1x2	2	1	1	2	1.6100000143051	1.1301068067551
1x4	4	1	1	4	0.84999990463257	1.0495953559875
1x8	8	1	1	8	0.57500004768372	1.1720380783081
1x16	16	1	1	16	0.34499996900558	1.0622457265854
1x24	24	1	1	24	0.31499993801117	1.1014844179153
1x32	32	1	1	32	0.30499997735023	1.1525838375092
1x40	40	1	1	40	0.30500000715256	1.2903789281845
1x48	48	1	1	48	0.35499992966652	1.2483825683594
1x56	56	1	1	56	0.31000000238419	1.2399953603745
1x64	64	1	1	64	0.28999999165535	1.2308721542358
1x72	72	1	1	72	0.28000000119209	1.2027974128723
1x80	80	1	1	80	0.27000001072884	1.1922084093094
1x88	88	1	1	88	0.27000001072884	1.2020003795624
1x96	96	1	1	96	0.27500000596046	1.2024736404419

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Expensive FP math instructions/calls

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

Code clean check

Workaround

Vectorization

Details

Workaround

Expensive FP math instructions/calls

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

▶Scalability Runs Description

▼Columns Filter

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency