OV - exec - Loop 97

	Loop Id: 97	Module: exec	Source: ljForce.c:157-158 [...]	Coverage: 0.23%

0x40c160 VMOVUPS	%XMM0,(%R9)    [1]

0x40c165 MOVQ	$0,0x10(%R9)    [1]

0x40c16d VMOVUPS	%XMM0,0x18(%R9)    [1]

0x40c173 MOVQ	$0,0x28(%R9)    [1]

0x40c17b VMOVUPS	%XMM0,0x30(%R9)    [1]

0x40c181 MOVQ	$0,0x40(%R9)    [1]

0x40c189 VMOVUPS	%XMM0,0x48(%R9)    [1]

0x40c18f MOVQ	$0,0x58(%R9)    [1]

0x40c197 VMOVUPS	%XMM0,0x60(%R9)    [1]

0x40c19d MOVQ	$0,0x70(%R9)    [1]

0x40c1a5 VMOVUPS	%XMM0,0x78(%R9)    [1]

0x40c1ab MOVQ	$0,0x88(%R9)    [1]

0x40c1b6 VMOVUPS	%XMM0,0x90(%R9)    [1]

0x40c1bf MOVQ	$0,0xa0(%R9)    [1]

0x40c1ca VMOVUPS	%XMM0,0xa8(%R9)    [1]

0x40c1d3 MOVQ	$0,0xb8(%R9)    [1]

0x40c1de ADD	$0xc0,%R9

0x40c1e5 DEC	%R8

0x40c1e8 JNE	40c160

/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/ljForce.c: 157 - 158

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

157:    #pragma omp parallel for

158:    for (int ii=0; ii<fSize; ++ii)

/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/mytype.h: 23 - 23

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

23:    a[1] = 0.0;

Coverage (%)	Name	Source Location	Module
►50.77+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
►37.00+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
►11.14+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec

Coverage (%)	Name	Source Location	Module
►91.47+	__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
►6.64+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
►1.41+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6

Coverage (%)	Name	Source Location	Module
►93.76+	__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
►4.84+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
►1.31+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6

Coverage (%)	Name	Source Location	Module
►94.64+	__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
►3.87+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
►1.07+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6

Coverage (%)	Name	Source Location	Module
►95.61+	__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
►3.10+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
►1.02+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6

Coverage (%)	Name	Source Location	Module
►96.49+	__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
►2.48+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	exec

Coverage (%)	Name	Source Location	Module
►96.49+	__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
►2.19+	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
○	__kmp_fork_call		libiomp5.so
○	__kmpc_fork_call		libiomp5.so
○	ljForce	ljForce.c:157	exec
○	timestep	timestep.c:51	exec
○	main	CoMD.c:125	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	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.78
Bottlenecks	P6, P7, P8, P9, P14, P15,
Function	ljForce.extracted.27
Source	ljForce.c:157-158,mytype.h:23-23
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	4.00
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	4.00
CQA cycles if fully vectorized	1.00
Front-end cycles	2.25
P0 cycles	0.33
P1 cycles	0.33
P2 cycles	0.33
P3 cycles	0.33
P4 cycles	0.33
P5 cycles	0.33
P6 cycles	4.00
P7 cycles	4.00
P8 cycles	4.00
P9 cycles	4.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	4.00
P15 cycles	4.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	19.00
Nb uops	18.00
Nb loads	0.00
Nb stores	16.00
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	48.00
Bytes prefetched	0.00
Bytes loaded	0.00
Bytes stored	192.00
Stride 0	0.00
Stride 1	1.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	50.00
Vectorization ratio load	NA
Vectorization ratio store	50.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	15.63
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	15.63
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	NA

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.78
Bottlenecks	P6, P7, P8, P9, P14, P15,
Function	ljForce.extracted.27
Source	ljForce.c:157-158,mytype.h:23-23
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	4.00
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	4.00
CQA cycles if fully vectorized	1.00
Front-end cycles	2.25
P0 cycles	0.33
P1 cycles	0.33
P2 cycles	0.33
P3 cycles	0.33
P4 cycles	0.33
P5 cycles	0.33
P6 cycles	4.00
P7 cycles	4.00
P8 cycles	4.00
P9 cycles	4.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	4.00
P15 cycles	4.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	19.00
Nb uops	18.00
Nb loads	0.00
Nb stores	16.00
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	48.00
Bytes prefetched	0.00
Bytes loaded	0.00
Bytes stored	192.00
Stride 0	0.00
Stride 1	1.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	50.00
Vectorization ratio load	NA
Vectorization ratio store	50.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	15.63
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	15.63
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	NA

Path /

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

Function	ljForce.extracted.27
Source file and lines	ljForce.c:157-158
Module	exec

The loop is defined in:

/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/ljForce.c:157-158
/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/mytype.h:23

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

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 15% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 4.00 to 1.00 cycles (4.00x speedup).

Details

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

50% of SSE/AVX stores are used in vector version.

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

Workaround

Try another compiler or update/tune your current one:
- 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 writing data to caches/RAM (the store unit is a bottleneck).

Workaround

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

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This loop is probably writing/copying data or processing integer elements.

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 storing 192 bytes.

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 with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)

General properties

nb instructions	19
nb uops	18
loop length	142
used x86 registers	2
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	2.25 cycles
front end	2.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.33	0.33	0.33	0.33	0.33	0.33	4.00	4.00	4.00	4.00	0.00	0.00	0.00	0.00	4.00	4.00
cycles	0.33	0.33	0.33	0.33	0.33	0.33	4.00	4.00	4.00	4.00	0.00	0.00	0.00	0.00	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	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	2.25
Dispatch	4.00
Data deps.	1.00
Overall L1	4.00

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

INT+FP

all	50%
load	NA (no load vectorizable/vectorized instructions)
store	50%
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	NA (no other vectorizable/vectorized instructions)

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

INT+FP

all	15%
load	NA (no load vectorizable/vectorized instructions)
store	15%
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	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 4.00 cycles. At this rate:

37% of peak store performance is reached (48.00 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: 40c160

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P14	P15	Latency	Recip. throughput	Vectorization
VMOVUPS %XMM0,(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x10(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x18(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x28(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x30(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x40(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x48(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x58(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x60(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x70(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x78(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x88(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x90(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0xa0(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0xa8(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0xb8(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
ADD $0xc0,%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
DEC %R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JNE 40c160 <ljForce.extracted.27+0xa0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50-1	N/A

Function	ljForce.extracted.27
Source file and lines	ljForce.c:157-158
Module	exec

The loop is defined in:

/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/ljForce.c:157-158
/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/mytype.h:23

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

gain
potential
hint
expert

Vectorization

Details

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

50% of SSE/AVX stores are used in vector version.

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

Workaround

Try another compiler or update/tune your current one:
- 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 writing data to caches/RAM (the store unit is a bottleneck).

Workaround

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

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This loop is probably writing/copying data or processing integer elements.

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 storing 192 bytes.

Unroll opportunity

Loop is data access bound.

Workaround

General properties

nb instructions	19
nb uops	18
loop length	142
used x86 registers	2
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	2.25 cycles
front end	2.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.33	0.33	0.33	0.33	0.33	0.33	4.00	4.00	4.00	4.00	0.00	0.00	0.00	0.00	4.00	4.00
cycles	0.33	0.33	0.33	0.33	0.33	0.33	4.00	4.00	4.00	4.00	0.00	0.00	0.00	0.00	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	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	2.25
Dispatch	4.00
Data deps.	1.00
Overall L1	4.00

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

INT+FP

all	50%
load	NA (no load vectorizable/vectorized instructions)
store	50%
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	NA (no other vectorizable/vectorized instructions)

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

INT+FP

all	15%
load	NA (no load vectorizable/vectorized instructions)
store	15%
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	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 4.00 cycles. At this rate:

37% of peak store performance is reached (48.00 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: 40c160

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P14	P15	Latency	Recip. throughput	Vectorization
VMOVUPS %XMM0,(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x10(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x18(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x28(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x30(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x40(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x48(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x58(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x60(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x70(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x78(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0x88(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0x90(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0xa0(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVUPS %XMM0,0xa8(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	4	0.50	vect (25.0%)
MOVQ $0,0xb8(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
ADD $0xc0,%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
DEC %R8	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JNE 40c160 <ljForce.extracted.27+0xa0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50-1	N/A

▶Scalability Runs Description

Run 8x1	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 1I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads
Run 8x12	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 12I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads
Run 8x16	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 16I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads
Run 8x20	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 20I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads
Run 8x24	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 24I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads
Run 8x28	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 28I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads
Run 8x32	Number processes: 8Number nodes: 1Run Command: <executable> -x 200 -y 200 -z 200 --xproc=2 --yproc=2 --zproc=2MPI Command: mpirun -n <number_processes> Dataset: Run Directory: /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/run/oneview_runs/multicore/icx/oneview_run_1740438801OMP_NUM_THREADS: 32I_MPI_PIN_ORDER: bunchOMP_DISPLAY_AFFINITY: TRUEOMP_PROC_BIND: spreadOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEI_MPI_PIN_DOMAIN: autoI_MPI_DEBUG: 4OMP_PLACES: threads

(8x1) Efficiency	(8x1) Potential Speed-Up (%)	(8x12) Efficiency	(8x12) Potential Speed-Up (%)	(8x16) Efficiency	(8x16) Potential Speed-Up (%)	(8x20) Efficiency	(8x20) Potential Speed-Up (%)	(8x24) Efficiency	(8x24) Potential Speed-Up (%)	(8x28) Efficiency	(8x28) Potential Speed-Up (%)	(8x32) Efficiency	(8x32) Potential Speed-Up (%)
1	0	0.15	1.11	0.11	1.46	0.09	1.66	0.07	1.93	0.06	2.09	0.06	2.26

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency

Run	Number of threads	Efficiency (ideal is 1)	Speedup	Ideal Speedup	Time (s)	Coverage (%)
8x1	8	1	1	1	1.1699997186661	0.22979193925858
8x12	96	0.15	1.78	12	0.70999991893768	1.2978527545929
8x16	128	0.11	1.79	16	0.70500004291534	1.6469178199768
8x20	160	0.09	1.8	20	0.70999997854233	1.8238477706909
8x24	192	0.07	1.78	24	0.70499992370605	2.0807604789734
8x28	224	0.06	1.78	28	0.69999998807907	2.2350008487701
8x32	256	0.06	1.76	32	0.7399999499321	2.3938591480255

Report Configuration

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

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

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

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

▶Scalability Runs Description

▼Columns Filter

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency