OV - exec - Loop 2052

	Loop Id: 2052	Module: exec	Source: ams.c:3672-3675	Coverage: 40.41%

0x4efad0 MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efad4 VMOVSD	(%R8,%RAX,8),%XMM8    [1]

0x4efada VFNMADD231SD	(%RSI,%RDX,8),%XMM8,%XMM0    [3]

0x4efae0 LEA	0x2(%RCX),%RDX

0x4efae4 MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efae8 VMOVSD	(%R8,%RAX,8),%XMM9    [1]

0x4efaee VFNMADD231SD	(%RSI,%RDX,8),%XMM9,%XMM0    [3]

0x4efaf4 LEA	0x3(%RCX),%RDX

0x4efaf8 MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efafc VMOVSD	(%R8,%RAX,8),%XMM10    [1]

0x4efb02 VFNMADD231SD	(%RSI,%RDX,8),%XMM10,%XMM0    [3]

0x4efb08 LEA	0x4(%RCX),%RDX

0x4efb0c MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efb10 VMOVSD	(%R8,%RAX,8),%XMM11    [1]

0x4efb16 VFNMADD231SD	(%RSI,%RDX,8),%XMM11,%XMM0    [3]

0x4efb1c LEA	0x5(%RCX),%RDX

0x4efb20 MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efb24 VMOVSD	(%R8,%RAX,8),%XMM12    [1]

0x4efb2a VFNMADD231SD	(%RSI,%RDX,8),%XMM12,%XMM0    [3]

0x4efb30 LEA	0x6(%RCX),%RDX

0x4efb34 MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efb38 VMOVSD	(%R8,%RAX,8),%XMM13    [1]

0x4efb3e VFNMADD231SD	(%RSI,%RDX,8),%XMM13,%XMM0    [3]

0x4efb44 LEA	0x7(%RCX),%RDX

0x4efb48 MOV	(%RDI,%RDX,8),%RAX    [2]

0x4efb4c ADD	$0x8,%RCX

0x4efb50 VMOVSD	(%R8,%RAX,8),%XMM14    [1]

0x4efb56 VMOVSD	(%RSI,%RCX,8),%XMM1    [3]

0x4efb5b VFNMADD231SD	(%RSI,%RDX,8),%XMM14,%XMM0    [3]

0x4efb61 MOV	(%RDI,%RCX,8),%RAX    [2]

0x4efb65 LEA	0x1(%RCX),%RDX

0x4efb69 VFNMADD231SD	(%R8,%RAX,8),%XMM1,%XMM0    [1]

0x4efb6f CMP	%R13,%RDX

0x4efb72 JNE	4efad0

/home/eoseret/qaas_runs_GNR/173-927-0874/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c: 3672 - 3675

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

3672:             for (jj = A_diag_i[i]; jj < A_diag_i[i+1]; jj++)

3673:             {

3674:                ii = A_diag_j[jj];

3675:                res -= A_diag_data[jj] * Vtemp_data[ii];

Coverage (%)	Name	Source Location	Module
○97.72	gomp_thread_start	team.c:130	libgomp.so.1.0.0
○2.28	GOMP_parallel	libgomp.h:982	libgomp.so.1.0.0

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	2.31
CQA speedup if fully vectorized	8.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	4.00
Bottlenecks
Function	hypre_ParCSRRelaxThreads._omp_fn.1
Source	ams.c:3672-3675
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	32.00
CQA cycles if no scalar integer	32.00
CQA cycles if FP arith vectorized	13.88
CQA cycles if fully vectorized	4.00
Front-end cycles	6.83
P0 cycles	4.00
P1 cycles	4.00
P2 cycles	8.00
P3 cycles	8.00
P4 cycles	0.00
P5 cycles	0.00
P6 cycles	1.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	8.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	32
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	34.00
Nb uops	33.00
Nb loads	24.00
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	0.50
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	8.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	6.00
Bytes prefetched	0.00
Bytes loaded	192.00
Bytes stored	0.00
Stride 0	0.00
Stride 1	0.00
Stride n	2.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	12.50
Vector-efficiency ratio load	12.50
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	12.50
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	NA

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	2.31
CQA speedup if fully vectorized	8.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	4.00
Bottlenecks
Function	hypre_ParCSRRelaxThreads._omp_fn.1
Source	ams.c:3672-3675
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	32.00
CQA cycles if no scalar integer	32.00
CQA cycles if FP arith vectorized	13.88
CQA cycles if fully vectorized	4.00
Front-end cycles	6.83
P0 cycles	4.00
P1 cycles	4.00
P2 cycles	8.00
P3 cycles	8.00
P4 cycles	0.00
P5 cycles	0.00
P6 cycles	1.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	8.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	32
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	34.00
Nb uops	33.00
Nb loads	24.00
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	0.50
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	8.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	6.00
Bytes prefetched	0.00
Bytes loaded	192.00
Bytes stored	0.00
Stride 0	0.00
Stride 1	0.00
Stride n	2.00
Stride unknown	0.00
Stride indirect	1.00
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	12.50
Vector-efficiency ratio load	12.50
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	12.50
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	NA

Path /

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

Function	hypre_ParCSRRelaxThreads._omp_fn.1
Source file and lines	ams.c:3672-3675
Module	exec

The loop is defined in /home/eoseret/qaas_runs_GNR/173-927-0874/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3672-3675.

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

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. 8 data elements could be processed at once in vector registers.

By vectorizing your loop, you can lower the cost of an iteration from 32.00 to 4.00 cycles (8.00x speedup).

Details

All SSE/AVX 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)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

FMA

Detected 8 FMA (fused multiply-add) operations.

Slow data structures access

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

Details

Constant non-unit stride: 2 occurrence(s)
Irregular (variable stride) or indirect: 1 occurrence(s)

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

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Type of elements and instruction set

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 16 FP arithmetical operations:

8: addition or subtraction (all inside FMA instructions)
8: multiply (all inside FMA instructions)

The binary loop is loading 192 bytes (24 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.

Unroll opportunity

Loop is data access bound.

Workaround

Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or by recompiling with -funroll-loops and/or -floop-unroll-and-jam. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma GCC unroll N

General properties

nb instructions	34
nb uops	33
loop length	168
used x86 registers	7
used mmx registers	0
used xmm registers	9
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	6.83 cycles
front end	6.83 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	4.00	4.00	8.00	8.00	0.00	0.00	1.00	0.00	0.00	0.00	0.00	8.00
cycles	4.00	4.00	8.00	8.00	0.00	0.00	1.00	0.00	0.00	0.00	0.00	8.00

Execution ports to units layout:

P0 (256 bits): VPU, BRU, ALU, DIV/SQRT
P1 (256 bits): ALU, VPU
P2 (512 bits): load
P3 (512 bits): load
P4 (512 bits): store data
P5 (512 bits): ALU, VPU
P6: ALU, BRU
P7: store address
P8: store address
P9 (512 bits): store data
P10: ALU
P11 (512 bits): load

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	32.00

Cycles summary

Front-end	6.83
Dispatch	8.00
Data deps.	32.00
Overall L1	32.00

Vectorization ratios

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

Vector efficiency ratios

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 32.00 cycles. At this rate:

3% of peak load performance is reached (6.00 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 4efb61

Instruction	Nb FU	P0	P1	P2	P3	P5	P6	P10	P11	Latency	Recip. throughput	Vectorization
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM8	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM8,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x2(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM9	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM9,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x3(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM10	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM10,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x4(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM11	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM11,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x5(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM12	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM12,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x6(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM13	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM13,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x7(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
ADD $0x8,%RCX	1	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVSD (%R8,%RAX,8),%XMM14	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VMOVSD (%RSI,%RCX,8),%XMM1	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM14,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
MOV (%RDI,%RCX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
LEA 0x1(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
VFNMADD231SD (%R8,%RAX,8),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
CMP %R13,%RDX	1	0.20	0.20	0	0	0.20	0.20	0.20	0	1	0.20	N/A
JNE 4efad0 <hypre_ParCSRRelaxThreads._omp_fn.1+0x190>	1	0.50	0	0	0	0	0.50	0	0	0	0.50	N/A

Function	hypre_ParCSRRelaxThreads._omp_fn.1
Source file and lines	ams.c:3672-3675
Module	exec

The loop is defined in /home/eoseret/qaas_runs_GNR/173-927-0874/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3672-3675.

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

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. 8 data elements could be processed at once in vector registers.

By vectorizing your loop, you can lower the cost of an iteration from 32.00 to 4.00 cycles (8.00x speedup).

Details

All SSE/AVX 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)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

FMA

Detected 8 FMA (fused multiply-add) operations.

Slow data structures access

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

Details

Constant non-unit stride: 2 occurrence(s)
Irregular (variable stride) or indirect: 1 occurrence(s)

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

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

Type of elements and instruction set

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 16 FP arithmetical operations:

8: addition or subtraction (all inside FMA instructions)
8: multiply (all inside FMA instructions)

The binary loop is loading 192 bytes (24 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.

Unroll opportunity

Loop is data access bound.

Workaround

General properties

nb instructions	34
nb uops	33
loop length	168
used x86 registers	7
used mmx registers	0
used xmm registers	9
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	6.83 cycles
front end	6.83 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11
uops	4.00	4.00	8.00	8.00	0.00	0.00	1.00	0.00	0.00	0.00	0.00	8.00
cycles	4.00	4.00	8.00	8.00	0.00	0.00	1.00	0.00	0.00	0.00	0.00	8.00

Execution ports to units layout:

P0 (256 bits): VPU, BRU, ALU, DIV/SQRT
P1 (256 bits): ALU, VPU
P2 (512 bits): load
P3 (512 bits): load
P4 (512 bits): store data
P5 (512 bits): ALU, VPU
P6: ALU, BRU
P7: store address
P8: store address
P9 (512 bits): store data
P10: ALU
P11 (512 bits): load

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	32.00

Cycles summary

Front-end	6.83
Dispatch	8.00
Data deps.	32.00
Overall L1	32.00

Vectorization ratios

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

Vector efficiency ratios

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 32.00 cycles. At this rate:

3% of peak load performance is reached (6.00 out of 192.00 bytes loaded per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 4efb61

Instruction	Nb FU	P0	P1	P2	P3	P5	P6	P10	P11	Latency	Recip. throughput	Vectorization
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM8	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM8,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x2(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM9	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM9,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x3(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM10	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM10,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x4(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM11	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM11,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x5(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM12	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM12,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x6(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
VMOVSD (%R8,%RAX,8),%XMM13	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM13,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
LEA 0x7(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RDI,%RDX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
ADD $0x8,%RCX	1	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVSD (%R8,%RAX,8),%XMM14	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VMOVSD (%RSI,%RCX,8),%XMM1	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	scal (12.5%)
VFNMADD231SD (%RSI,%RDX,8),%XMM14,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
MOV (%RDI,%RCX,8),%RAX	1	0	0	0.33	0.33	0	0	0	0.33	1	0.33	N/A
LEA 0x1(%RCX),%RDX	1	0	0	0	0	0	0	0	0	0	0.17	N/A
VFNMADD231SD (%R8,%RAX,8),%XMM1,%XMM0	1	0.50	0.50	0.33	0.33	0	0	0	0.33	4	0.50	scal (12.5%)
CMP %R13,%RDX	1	0.20	0.20	0	0	0.20	0.20	0.20	0	1	0.20	N/A
JNE 4efad0 <hypre_ParCSRRelaxThreads._omp_fn.1+0x190>	1	0.50	0	0	0	0	0.50	0	0	0	0.50	N/A

Report Configuration

Vectorization

Details

Workaround

Execution units bottlenecks

FMA

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

Unroll opportunity

Workaround

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Vectorization

Details

Workaround

Execution units bottlenecks

FMA

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

Unroll opportunity

Workaround

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code