OV - exec - Loop 62

	Loop Id: 62	Module: exec	Source: haloExchange.c:633-641	Coverage: 0.01%

0x408e70 MOV	(%RBX),%R8D    [3]

0x408e73 MOV	%R8D,(%RAX,%R15,4)    [2]

0x408e77 MOV	0x4(%RBX),%R8D    [3]

0x408e7b MOV	%R8D,(%RCX,%R15,4)    [4]

0x408e7f VMOVSD	0x8(%RBX),%XMM0    [3]

0x408e84 VMOVSD	%XMM0,(%RDX,%RDI,1)    [5]

0x408e89 VMOVUPS	0x10(%RBX),%XMM0    [3]

0x408e8e VMOVUPS	%XMM0,0x8(%RDX,%RDI,1)    [5]

0x408e94 VMOVSD	0x20(%RBX),%XMM0    [3]

0x408e99 VMOVSD	%XMM0,(%RSI,%RDI,1)    [1]

0x408e9e VMOVDQU	0x28(%RBX),%XMM0    [3]

0x408ea3 VMOVDQU	%XMM0,0x8(%RSI,%RDI,1)    [1]

0x408ea9 ADD	$0x18,%RDI

0x408ead INC	%R15

0x408eb0 ADD	$0x38,%RBX

0x408eb4 CMP	%R15,%R12

0x408eb7 JNE	408e70

/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c: 633 - 641

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

633:    for (int ii=begin, iTmp=0; ii<end; ++ii, ++iTmp)

634:    {

635:       atoms->gid[ii]   = tmp[iTmp].gid;

636:       atoms->iSpecies[ii] = tmp[iTmp].type;

637:       atoms->r[ii][0]  = tmp[iTmp].rx;

638:       atoms->r[ii][1]  = tmp[iTmp].ry;

639:       atoms->r[ii][2]  = tmp[iTmp].rz;

640:       atoms->p[ii][0]  = tmp[iTmp].px;

641:       atoms->p[ii][1]  = tmp[iTmp].py;

Coverage (%)	Name	Source Location	Module
►21.18+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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
►13.79+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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
►12.81+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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
►12.32+	redistributeAtoms.extracted	timestep.c:153	exec
○	__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
►10.84+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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
►9.36+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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
►7.88+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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
►6.90+	redistributeAtoms.extracted	timestep.c:154	exec
○	__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

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	1.50
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	6.86
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.50
Bottlenecks	P6, P7, P8, P9,
Function	sortAtomsInCell
Source	haloExchange.c:633-641
Source loop unroll info	unrolled by 16
Source loop unroll confidence level	max
Unroll/vectorization loop type	peel/tail
Unroll factor	1
CQA cycles	3.00
CQA cycles if no scalar integer	2.00
CQA cycles if FP arith vectorized	3.00
CQA cycles if fully vectorized	0.44
Front-end cycles	2.00
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	0.67
P4 cycles	0.67
P5 cycles	0.67
P6 cycles	3.00
P7 cycles	3.00
P8 cycles	3.00
P9 cycles	3.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	2.00
P15 cycles	2.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	17.00
Nb uops	16.00
Nb loads	6.00
Nb stores	6.00
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	37.33
Bytes prefetched	0.00
Bytes loaded	56.00
Bytes stored	56.00
Stride 0	0.00
Stride 1	5.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	33.33
Vectorization ratio load	33.33
Vectorization ratio store	33.33
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	14.58
Vector-efficiency ratio load	14.58
Vector-efficiency ratio store	14.58
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	1.50
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	6.86
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.50
Bottlenecks	P6, P7, P8, P9,
Function	sortAtomsInCell
Source	haloExchange.c:633-641
Source loop unroll info	unrolled by 16
Source loop unroll confidence level	max
Unroll/vectorization loop type	peel/tail
Unroll factor	1
CQA cycles	3.00
CQA cycles if no scalar integer	2.00
CQA cycles if FP arith vectorized	3.00
CQA cycles if fully vectorized	0.44
Front-end cycles	2.00
P0 cycles	0.67
P1 cycles	0.67
P2 cycles	0.67
P3 cycles	0.67
P4 cycles	0.67
P5 cycles	0.67
P6 cycles	3.00
P7 cycles	3.00
P8 cycles	3.00
P9 cycles	3.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	2.00
P15 cycles	2.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	17.00
Nb uops	16.00
Nb loads	6.00
Nb stores	6.00
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	37.33
Bytes prefetched	0.00
Bytes loaded	56.00
Bytes stored	56.00
Stride 0	0.00
Stride 1	5.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	33.33
Vectorization ratio load	33.33
Vectorization ratio store	33.33
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	14.58
Vector-efficiency ratio load	14.58
Vector-efficiency ratio store	14.58
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	sortAtomsInCell
Source file and lines	haloExchange.c:633-641
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:633-641.

It is peel/tail loop of related source loop which is unrolled by 16 (including vectorization).

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 3.00 to 2.00 cycles (1.50x speedup).

Workaround

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

Unrolling/vectorization cost

This loop is peel/tail of a unrolled/vectorized loop. If its cost is not negligible compared to the main (unrolled/vectorized) loop, unrolling/vectorization is counterproductive due to low trip count.

Details

The more iterations the main loop is processing, the higher the trip count must be to amortize peel/tail overhead.

Workaround

recompile with -fprofile-instr-generate, execute ,merge raw profiles with 'llvm-profdata merge -o default.profdata default.profraw' and recompile with -fprofile-instr-use (profile-guided optimization)
hardcode most frequent values of loop bounds by adding specialized paths.:
- For instance, replace for (i=0; i<n; i++) foo(i) with: switch (n) { case (4): for (i=0; i<4; i++) foo(i); break; case (6): for (i=0; i<6; i++) foo(i); break; default : for (i=0; i<n; i++) foo(i); break; }

Vectorization

Your loop is poorly vectorized. Only 14% of vector register length is used (average across all SSE/AVX instructions).

Details

33% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

33% of SSE/AVX loads are used in vector version.
33% of SSE/AVX stores are used in vector version.

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)

By removing all these bottlenecks, you can lower the cost of an iteration from 3.00 to 2.00 cycles (1.50x speedup).

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

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 loading 56 bytes. The binary loop is storing 56 bytes.

General properties

nb instructions	17
nb uops	16
loop length	73
used x86 registers	9
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	2.00 cycles
front end	2.00 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	0.67	0.67	0.67	3.00	3.00	3.00	3.00	0.00	0.00	0.00	0.00	2.00	2.00
cycles	0.67	0.67	0.67	0.67	0.67	0.67	3.00	3.00	3.00	3.00	0.00	0.00	0.00	0.00	2.00	2.00

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	2.00
Dispatch	3.00
Data deps.	1.00
Overall L1	3.00

Vectorization ratios

INT

all	33%
load	33%
store	33%
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	33%
load	33%
store	33%
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	33%
load	33%
store	33%
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	12%
load	12%
store	12%
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	16%
load	16%
store	16%
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	14%
load	14%
store	14%
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 3.00 cycles. At this rate:

7% of peak load performance is reached (18.67 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
14% of peak store performance is reached (18.67 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: 408e70

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P14	P15	Latency	Recip. throughput	Vectorization
MOV (%RBX),%R8D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R8D,(%RAX,%R15,4)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOV 0x4(%RBX),%R8D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R8D,(%RCX,%R15,4)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVSD 0x8(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVSD %XMM0,(%RDX,%RDI,1)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	1	0.50	scal (12.5%)
VMOVUPS 0x10(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
VMOVUPS %XMM0,0x8(%RDX,%RDI,1)	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%)
VMOVSD 0x20(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVSD %XMM0,(%RSI,%RDI,1)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	1	0.50	scal (12.5%)
VMOVDQU 0x28(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
VMOVDQU %XMM0,0x8(%RSI,%RDI,1)	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%)
ADD $0x18,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
INC %R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD $0x38,%RBX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
CMP %R15,%R12	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 408e70 <sortAtomsInCell+0x790>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50-1	N/A

Function	sortAtomsInCell
Source file and lines	haloExchange.c:633-641
Module	exec

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 3.00 to 2.00 cycles (1.50x speedup).

Workaround

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

Unrolling/vectorization cost

Details

The more iterations the main loop is processing, the higher the trip count must be to amortize peel/tail overhead.

Workaround

recompile with -fprofile-instr-generate, execute ,merge raw profiles with 'llvm-profdata merge -o default.profdata default.profraw' and recompile with -fprofile-instr-use (profile-guided optimization)
hardcode most frequent values of loop bounds by adding specialized paths.:
- For instance, replace for (i=0; i<n; i++) foo(i) with: switch (n) { case (4): for (i=0; i<4; i++) foo(i); break; case (6): for (i=0; i<6; i++) foo(i); break; default : for (i=0; i<n; i++) foo(i); break; }

Vectorization

Your loop is poorly vectorized. Only 14% of vector register length is used (average across all SSE/AVX instructions).

Details

33% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

33% of SSE/AVX loads are used in vector version.
33% of SSE/AVX stores are used in vector version.

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)

By removing all these bottlenecks, you can lower the cost of an iteration from 3.00 to 2.00 cycles (1.50x speedup).

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

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 loading 56 bytes. The binary loop is storing 56 bytes.

General properties

nb instructions	17
nb uops	16
loop length	73
used x86 registers	9
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	2.00 cycles
front end	2.00 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	0.67	0.67	0.67	3.00	3.00	3.00	3.00	0.00	0.00	0.00	0.00	2.00	2.00
cycles	0.67	0.67	0.67	0.67	0.67	0.67	3.00	3.00	3.00	3.00	0.00	0.00	0.00	0.00	2.00	2.00

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	2.00
Dispatch	3.00
Data deps.	1.00
Overall L1	3.00

Vectorization ratios

INT

all	33%
load	33%
store	33%
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	33%
load	33%
store	33%
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	33%
load	33%
store	33%
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	12%
load	12%
store	12%
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	16%
load	16%
store	16%
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	14%
load	14%
store	14%
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 3.00 cycles. At this rate:

7% of peak load performance is reached (18.67 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
14% of peak store performance is reached (18.67 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: 408e70

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P14	P15	Latency	Recip. throughput	Vectorization
MOV (%RBX),%R8D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R8D,(%RAX,%R15,4)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOV 0x4(%RBX),%R8D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R8D,(%RCX,%R15,4)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
VMOVSD 0x8(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVSD %XMM0,(%RDX,%RDI,1)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	1	0.50	scal (12.5%)
VMOVUPS 0x10(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
VMOVUPS %XMM0,0x8(%RDX,%RDI,1)	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%)
VMOVSD 0x20(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
VMOVSD %XMM0,(%RSI,%RDI,1)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	1	0.50	scal (12.5%)
VMOVDQU 0x28(%RBX),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
VMOVDQU %XMM0,0x8(%RSI,%RDI,1)	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%)
ADD $0x18,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
INC %R15	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
ADD $0x38,%RBX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
CMP %R15,%R12	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 408e70 <sortAtomsInCell+0x790>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50-1	N/A

Report Configuration

Code clean check

Workaround

Unrolling/vectorization cost

Details

Workaround

Vectorization

Details

Execution units bottlenecks

Workaround

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

Unrolling/vectorization cost

Details

Workaround

Vectorization

Details

Execution units bottlenecks

Workaround

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