OV - exec - Loop 306

	Loop Id: 306	Module: exec	Source: par_coarsen.c:2327-2334	Coverage: 0.07%

0x43dbd0 ADD	$0x40,%RDI

0x43dbd4 DEC	%RSI

0x43dbd7 JE	43db5d

0x43dbd9 MOV	-0x38(%RDI),%R8    [1]

0x43dbdd UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dbe3 JB	43dc50

0x43dbe5 MOV	-0x30(%RDI),%R8    [1]

0x43dbe9 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dbef JB	43dc64

0x43dbf1 MOV	-0x28(%RDI),%R8    [1]

0x43dbf5 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dbfb JB	43dc78

0x43dbfd MOV	-0x20(%RDI),%R8    [1]

0x43dc01 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc07 JB	43dc90

0x43dc0d MOV	-0x18(%RDI),%R8    [1]

0x43dc11 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc17 JB	43dca8

0x43dc1d MOV	-0x10(%RDI),%R8    [1]

0x43dc21 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc27 JB	43dcc0

0x43dc2d MOV	-0x8(%RDI),%R8    [1]

0x43dc31 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc37 JB	43dcd8

0x43dc3d MOV	(%RDI),%R8    [1]

0x43dc40 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc46 JAE	43dbd0

0x43dc48 JMP	43dcef

0x43dc50 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dc58 MOV	-0x30(%RDI),%R8    [1]

0x43dc5c UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc62 JAE	43dbf1

0x43dc64 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dc6c MOV	-0x28(%RDI),%R8    [1]

0x43dc70 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc76 JAE	43dbfd

0x43dc78 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dc80 MOV	-0x20(%RDI),%R8    [1]

0x43dc84 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dc8a JAE	43dc0d

0x43dc90 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dc98 MOV	-0x18(%RDI),%R8    [1]

0x43dc9c UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dca2 JAE	43dc1d

0x43dca8 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dcb0 MOV	-0x10(%RDI),%R8    [1]

0x43dcb4 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dcba JAE	43dc2d

0x43dcc0 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dcc8 MOV	-0x8(%RDI),%R8    [1]

0x43dccc UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dcd2 JAE	43dc3d

0x43dcd8 MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dce0 MOV	(%RDI),%R8    [1]

0x43dce3 UCOMISD	(%R14,%R8,8),%XMM0    [3]

0x43dce9 JAE	43dbd0

0x43dcef MOVQ	$0x1,(%R12,%R8,8)    [2]

0x43dcf7 JMP	43dbd0

/home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c: 2327 - 2334

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

2327: #pragma omp parallel for private(ig, i) HYPRE_SMP_SCHEDULE

2328: #endif

2329:         for (ig = 0; ig < graph_size; ig++)

2330:         {

2331:            i = graph_array[ig];

2332:            if (measure_array[i] > 1)

2333:            {

2334:               CF_marker[i] = 1;

Coverage (%)	Name	Source Location	Module
►94.19+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►2.85+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_launch_thread		libiomp5.so
○	_INTERNALea951ae6::__kmp_launc[...]		libiomp5.so
►2.79+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2327	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start		exec

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	1.27
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	10.13
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P6, P7, P8, P9,
Function	hypre_BoomerAMGCoarsenPMIS.extracted.25
Source	par_coarsen.c:2327-2334
Source loop unroll info	unrolled by 8
Source loop unroll confidence level	high
Unroll/vectorization loop type	main
Unroll factor	8
CQA cycles	9.50
CQA cycles if no scalar integer	7.50
CQA cycles if FP arith vectorized	9.50
CQA cycles if fully vectorized	0.94
Front-end cycles	9.13
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	6.00
P4 cycles	6.00
P5 cycles	6.00
P6 cycles	9.50
P7 cycles	9.50
P8 cycles	9.50
P9 cycles	9.50
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	7.50
P13 cycles	7.50
P14 cycles	7.50
P15 cycles	7.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	58.00
Nb uops	73.00
Nb loads	30.00
Nb stores	8.00
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	32.00
Bytes prefetched	0.00
Bytes loaded	240.00
Bytes stored	64.00
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	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	10.33
Vector-efficiency ratio load	12.50
Vector-efficiency ratio store	6.25
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	12.50

Metric	Value
CQA speedup if no scalar integer	1.27
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	10.13
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P6, P7, P8, P9,
Function	hypre_BoomerAMGCoarsenPMIS.extracted.25
Source	par_coarsen.c:2327-2334
Source loop unroll info	unrolled by 8
Source loop unroll confidence level	high
Unroll/vectorization loop type	main
Unroll factor	8
CQA cycles	9.50
CQA cycles if no scalar integer	7.50
CQA cycles if FP arith vectorized	9.50
CQA cycles if fully vectorized	0.94
Front-end cycles	9.13
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	6.00
P4 cycles	6.00
P5 cycles	6.00
P6 cycles	9.50
P7 cycles	9.50
P8 cycles	9.50
P9 cycles	9.50
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	7.50
P13 cycles	7.50
P14 cycles	7.50
P15 cycles	7.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	58.00
Nb uops	73.00
Nb loads	30.00
Nb stores	8.00
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	32.00
Bytes prefetched	0.00
Bytes loaded	240.00
Bytes stored	64.00
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	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	10.33
Vector-efficiency ratio load	12.50
Vector-efficiency ratio store	6.25
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	12.50

Path /

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

Function	hypre_BoomerAMGCoarsenPMIS.extracted.25
Source file and lines	par_coarsen.c:2327-2334
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2327-2334.

It is main loop of related source loop which is unrolled by 8 (including vectorization).
Warnings:

Ignoring paths for analysis
Too many paths. If you really need to analyze all of the 256 paths individually, rerun with max-paths=256
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 256 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

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 9.50 to 7.50 cycles (1.27x 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 10% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your loop, you can lower the cost of an iteration from 9.50 to 0.94 cycles (10.13x 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:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
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

No data for this section

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.

UCOMISD: 15 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332

Type of elements and instruction set

15 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 does not contain any FP arithmetical operations. The binary loop is loading 240 bytes (30 double precision FP elements). The binary loop is storing 64 bytes (8 double precision FP elements).

General properties

nb instructions	58
nb uops	73
loop length	297
used x86 registers	5
used mmx registers	0
used xmm registers	1
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	9.13 cycles
front end	9.13 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	6.00	6.00	6.00	9.50	9.50	9.50	9.50	0.00	0.00	7.50	7.50	7.50	7.50
cycles	0.67	0.67	0.67	6.00	6.00	6.00	9.50	9.50	9.50	9.50	0.00	0.00	7.50	7.50	7.50	7.50

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

Cycles summary

Front-end	9.13
Dispatch	9.50
Overall L1	9.50

Vectorization ratios

INT

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

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	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	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%

Vector efficiency ratios

INT

all	6%
load	NA (no load vectorizable/vectorized instructions)
store	6%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

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	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

INT+FP

all	10%
load	12%
store	6%
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	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 9.50 cycles. At this rate:

9% of peak load performance is reached (25.26 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
5% of peak store performance is reached (6.74 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: 43dbd9

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
ADD $0x40,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
DEC %RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 43db5d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x7d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x38(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc50 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x170>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x30(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc64 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x184>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x28(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc78 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x198>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x20(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc90 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1b0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x18(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dca8 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1c8>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x10(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dcc0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x8(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dcd8 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1f8>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV (%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbd0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0xf0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
JMP 43dcef <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x20f>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x30(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbf1 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x111>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x28(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbfd <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x11d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x20(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc0d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x12d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x18(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc1d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x13d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x10(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc2d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x14d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x8(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc3d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x15d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV (%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbd0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0xf0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
JMP 43dbd0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0xf0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	1	N/A

Function	hypre_BoomerAMGCoarsenPMIS.extracted.25
Source file and lines	par_coarsen.c:2327-2334
Module	exec

Ignoring paths for analysis
Too many paths. If you really need to analyze all of the 256 paths individually, rerun with max-paths=256
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

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 9.50 to 7.50 cycles (1.27x speedup).

Workaround

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

Vectorization

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:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
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

No data for this section

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.

UCOMISD: 15 occurrences►
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332
- /home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2332

Type of elements and instruction set

15 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

General properties

nb instructions	58
nb uops	73
loop length	297
used x86 registers	5
used mmx registers	0
used xmm registers	1
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	9.13 cycles
front end	9.13 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	6.00	6.00	6.00	9.50	9.50	9.50	9.50	0.00	0.00	7.50	7.50	7.50	7.50
cycles	0.67	0.67	0.67	6.00	6.00	6.00	9.50	9.50	9.50	9.50	0.00	0.00	7.50	7.50	7.50	7.50

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

Cycles summary

Front-end	9.13
Dispatch	9.50
Overall L1	9.50

Vectorization ratios

INT

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

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	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	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%

Vector efficiency ratios

INT

all	6%
load	NA (no load vectorizable/vectorized instructions)
store	6%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

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	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

INT+FP

all	10%
load	12%
store	6%
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	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 9.50 cycles. At this rate:

9% of peak load performance is reached (25.26 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
5% of peak store performance is reached (6.74 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: 43dbd9

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
ADD $0x40,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
DEC %RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 43db5d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x7d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x38(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc50 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x170>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x30(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc64 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x184>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x28(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc78 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x198>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x20(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dc90 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1b0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x18(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dca8 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1c8>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x10(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dcc0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV -0x8(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JB 43dcd8 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x1f8>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOV (%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbd0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0xf0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
JMP 43dcef <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x20f>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x30(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbf1 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x111>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x28(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbfd <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x11d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x20(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc0d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x12d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x18(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc1d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x13d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x10(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc2d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x14d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV -0x8(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dc3d <hypre_BoomerAMGCoarsenPMIS.extracted.25+0x15d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
MOV (%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
UCOMISD (%R14,%R8,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	7	1	scal (12.5%)
JAE 43dbd0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0xf0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A
MOVQ $0x1,(%R12,%R8,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	4	0.50	scal (6.3%)
JMP 43dbd0 <hypre_BoomerAMGCoarsenPMIS.extracted.25+0xf0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	1	N/A

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Complex instructions

Details

Type of elements and instruction set

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

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

Complex instructions

Details

Type of elements and instruction set

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code