OV - exec - Loop 102

	Loop Id: 102	Module: exec	Source: par_coarsen.c:2332-2334	Coverage: 0.07%

0x42fb90 MOV	(%RDX),%RBX    [4]

0x42fb93 LEA	(,%RBX,8),%R8

0x42fb9b VCOMISD	(%R9,%RBX,8),%XMM0    [1]

0x42fba1 JAE	42fbab

0x42fba3 MOVQ	$0x1,(%R10,%R8,1)    [2]

0x42fbab LEA	0x8(%RDX),%RCX

0x42fbaf MOV	0x8(%RDX),%RDX    [4]

0x42fbb3 LEA	(,%RDX,8),%RSI

0x42fbbb VCOMISD	(%R9,%RDX,8),%XMM0    [1]

0x42fbc1 JAE	42fbcb

0x42fbc3 MOVQ	$0x1,(%R10,%RSI,1)    [2]

0x42fbcb MOV	0x8(%RCX),%RAX    [3]

0x42fbcf LEA	(,%RAX,8),%RDI

0x42fbd7 VCOMISD	(%R9,%RAX,8),%XMM0    [1]

0x42fbdd JAE	42fbe7

0x42fbdf MOVQ	$0x1,(%R10,%RDI,1)    [2]

0x42fbe7 MOV	0x10(%RCX),%R12    [3]

0x42fbeb LEA	(,%R12,8),%RBX

0x42fbf3 VCOMISD	(%R9,%R12,8),%XMM0    [1]

0x42fbf9 JAE	42fc03

0x42fbfb MOVQ	$0x1,(%R10,%RBX,1)    [2]

0x42fc03 MOV	0x18(%RCX),%RDX    [3]

0x42fc07 LEA	(,%RDX,8),%R8

0x42fc0f VCOMISD	(%R9,%RDX,8),%XMM0    [1]

0x42fc15 JAE	42fc1f

0x42fc17 MOVQ	$0x1,(%R10,%R8,1)    [2]

0x42fc1f MOV	0x20(%RCX),%RSI    [3]

0x42fc23 LEA	(,%RSI,8),%RAX

0x42fc2b VCOMISD	(%R9,%RSI,8),%XMM0    [1]

0x42fc31 JAE	42fc3b

0x42fc33 MOVQ	$0x1,(%R10,%RAX,1)    [2]

0x42fc3b MOV	0x28(%RCX),%RDI    [3]

0x42fc3f LEA	(,%RDI,8),%R12

0x42fc47 VCOMISD	(%R9,%RDI,8),%XMM0    [1]

0x42fc4d JAE	42fc57

0x42fc4f MOVQ	$0x1,(%R10,%R12,1)    [2]

0x42fc57 MOV	0x30(%RCX),%RBX    [3]

0x42fc5b LEA	(,%RBX,8),%RDX

0x42fc63 VCOMISD	(%R9,%RBX,8),%XMM0    [1]

0x42fc69 JAE	42fc73

0x42fc6b MOVQ	$0x1,(%R10,%RDX,1)    [2]

0x42fc73 LEA	0x38(%RCX),%RDX

0x42fc77 CMP	%RDX,%R11

0x42fc7a JNE	42fb90

/home/eoseret/qaas_runs_ZEN5/173-989-6379/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c: 2332 - 2334

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

2332:            if (measure_array[i] > 1)

2333:            {

2334:               CF_marker[i] = 1;

Coverage (%)	Name	Source Location	Module
►97.04+	gomp_thread_start	team.c:130	libgomp.so.1.0.0
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►2.79+	GOMP_parallel	libgomp.h:982	libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2338	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	amg.c:253	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	1.59
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	12.75
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.06
Bottlenecks	micro-operation queue,
Function	hypre_BoomerAMGCoarsenPMIS._omp_fn.4
Source	par_coarsen.c:2332-2334
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	6.38
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	6.38
CQA cycles if fully vectorized	0.50
Front-end cycles	6.38
P0 cycles	3.17
P1 cycles	3.17
P2 cycles	3.17
P3 cycles	3.17
P4 cycles	3.17
P5 cycles	3.17
P6 cycles	6.00
P7 cycles	6.00
P8 cycles	6.00
P9 cycles	6.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	4.00
P13 cycles	4.00
P14 cycles	4.00
P15 cycles	4.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	44.00
Nb uops	51.00
Nb loads	16.00
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	30.12
Bytes prefetched	0.00
Bytes loaded	128.00
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	9.38
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	1.59
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	12.75
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.06
Bottlenecks	micro-operation queue,
Function	hypre_BoomerAMGCoarsenPMIS._omp_fn.4
Source	par_coarsen.c:2332-2334
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	6.38
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	6.38
CQA cycles if fully vectorized	0.50
Front-end cycles	6.38
P0 cycles	3.17
P1 cycles	3.17
P2 cycles	3.17
P3 cycles	3.17
P4 cycles	3.17
P5 cycles	3.17
P6 cycles	6.00
P7 cycles	6.00
P8 cycles	6.00
P9 cycles	6.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	4.00
P13 cycles	4.00
P14 cycles	4.00
P15 cycles	4.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	44.00
Nb uops	51.00
Nb loads	16.00
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	30.12
Bytes prefetched	0.00
Bytes loaded	128.00
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	9.38
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._omp_fn.4
Source file and lines	par_coarsen.c:2332-2334
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:2332-2334.

The related source loop is not unrolled or unrolled with no peel/tail loop.
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 6.38 to 4.00 cycles (1.59x 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 9% 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 6.38 to 0.50 cycles (12.75x 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.

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.

VCOMISD: 8 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

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 does not contain any FP arithmetical operations. The binary loop is loading 128 bytes (16 double precision FP elements). The binary loop is storing 64 bytes (8 double precision FP elements).

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	44
nb uops	51
loop length	240
used x86 registers	11
used mmx registers	0
used xmm registers	1
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	6.38 cycles
front end	6.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	3.17	3.17	3.17	3.17	3.17	3.17	6.00	6.00	6.00	6.00	0.00	0.00	4.00	4.00	4.00	4.00
cycles	3.17	3.17	3.17	3.17	3.17	3.17	6.00	6.00	6.00	6.00	0.00	0.00	4.00	4.00	4.00	4.00

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

Cycles summary

Front-end	6.38
Dispatch	6.00
Overall L1	6.38

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	9%
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 6.38 cycles. At this rate:

7% of peak load performance is reached (20.08 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
7% of peak store performance is reached (10.04 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Performance is limited by instruction throughput (loading/decoding program instructions to execution core) (front-end is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 6.38 to 6.00 cycles (1.06x speedup).

ASM code

In the binary file, the address of the loop is: 42fb90

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
MOV (%RDX),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RBX,8),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RBX,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	6	0.50	scal (12.5%)
JAE 42fbab <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x1ab>	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,(%R10,%R8,1)	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%)
LEA 0x8(%RDX),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0x8(%RDX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RDX,8),%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RDX,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	6	0.50	scal (12.5%)
JAE 42fbcb <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x1cb>	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,(%R10,%RSI,1)	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(%RCX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RAX,8),%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RAX,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	6	0.50	scal (12.5%)
JAE 42fbe7 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x1e7>	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,(%R10,%RDI,1)	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(%RCX),%R12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%R12,8),%RBX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%R12,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
JAE 42fc03 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x203>	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,(%R10,%RBX,1)	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(%RCX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RDX,8),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RDX,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	6	0.50	scal (12.5%)
JAE 42fc1f <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x21f>	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,(%R10,%R8,1)	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(%RCX),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RSI,8),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RSI,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	6	0.50	scal (12.5%)
JAE 42fc3b <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x23b>	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,(%R10,%RAX,1)	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(%RCX),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RDI,8),%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RDI,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	6	0.50	scal (12.5%)
JAE 42fc57 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x257>	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,(%R10,%R12,1)	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(%RCX),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RBX,8),%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RBX,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	6	0.50	scal (12.5%)
JAE 42fc73 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x273>	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,(%R10,%RDX,1)	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%)
LEA 0x38(%RCX),%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
CMP %RDX,%R11	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
JNE 42fb90 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x190>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Function	hypre_BoomerAMGCoarsenPMIS._omp_fn.4
Source file and lines	par_coarsen.c:2332-2334
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 6.38 to 4.00 cycles (1.59x 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
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.

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.

VCOMISD: 8 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

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

Unroll opportunity

Loop is data access bound.

Workaround

General properties

nb instructions	44
nb uops	51
loop length	240
used x86 registers	11
used mmx registers	0
used xmm registers	1
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	6.38 cycles
front end	6.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	3.17	3.17	3.17	3.17	3.17	3.17	6.00	6.00	6.00	6.00	0.00	0.00	4.00	4.00	4.00	4.00
cycles	3.17	3.17	3.17	3.17	3.17	3.17	6.00	6.00	6.00	6.00	0.00	0.00	4.00	4.00	4.00	4.00

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

Cycles summary

Front-end	6.38
Dispatch	6.00
Overall L1	6.38

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	9%
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 6.38 cycles. At this rate:

7% of peak load performance is reached (20.08 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
7% of peak store performance is reached (10.04 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the loop is: 42fb90

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
MOV (%RDX),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RBX,8),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RBX,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	6	0.50	scal (12.5%)
JAE 42fbab <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x1ab>	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,(%R10,%R8,1)	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%)
LEA 0x8(%RDX),%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
MOV 0x8(%RDX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RDX,8),%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RDX,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	6	0.50	scal (12.5%)
JAE 42fbcb <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x1cb>	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,(%R10,%RSI,1)	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(%RCX),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RAX,8),%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RAX,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	6	0.50	scal (12.5%)
JAE 42fbe7 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x1e7>	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,(%R10,%RDI,1)	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(%RCX),%R12	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%R12,8),%RBX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%R12,8),%XMM0	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0.50	0.50	6	0.50	scal (12.5%)
JAE 42fc03 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x203>	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,(%R10,%RBX,1)	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(%RCX),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RDX,8),%R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RDX,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	6	0.50	scal (12.5%)
JAE 42fc1f <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x21f>	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,(%R10,%R8,1)	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(%RCX),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RSI,8),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RSI,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	6	0.50	scal (12.5%)
JAE 42fc3b <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x23b>	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,(%R10,%RAX,1)	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(%RCX),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RDI,8),%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RDI,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	6	0.50	scal (12.5%)
JAE 42fc57 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x257>	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,(%R10,%R12,1)	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(%RCX),%RBX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	3	0.25	N/A
LEA (,%RBX,8),%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
VCOMISD (%R9,%RBX,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	6	0.50	scal (12.5%)
JAE 42fc73 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x273>	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,(%R10,%RDX,1)	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%)
LEA 0x38(%RCX),%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.25	N/A
CMP %RDX,%R11	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
JNE 42fb90 <hypre_BoomerAMGCoarsenPMIS._omp_fn.4+0x190>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Complex instructions

Details

Type of elements and instruction set

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

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

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Complex instructions

Details

Type of elements and instruction set

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

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