OV - exec - Loop 87

	Loop Id: 87	Module: exec	Source: haloExchange.c:621-630	Coverage: 0.95%

0x40cb90 MOV	(%RAX,%R10,4),%R11D    [4]

0x40cb94 MOV	%R11D,(%R9)    [2]

0x40cb97 MOV	(%RCX,%R10,4),%R11D    [5]

0x40cb9b MOV	%R11D,0x4(%R9)    [2]

0x40cb9f MOVUPS	-0x10(%RDI,%RDX,4),%XMM0    [3]

0x40cba4 MOVUPS	%XMM0,0x8(%R9)    [2]

0x40cba9 MOVSD	(%RDI,%RDX,4),%XMM0    [3]

0x40cbae MOVSD	%XMM0,0x18(%R9)    [2]

0x40cbb4 MOVUPS	-0x10(%R8,%RDX,4),%XMM0    [1]

0x40cbba MOVUPS	%XMM0,0x20(%R9)    [2]

0x40cbbf MOVSD	(%R8,%RDX,4),%XMM0    [1]

0x40cbc5 MOVSD	%XMM0,0x30(%R9)    [2]

0x40cbcb INC	%R10

0x40cbce ADD	$0x6,%RDX

0x40cbd2 ADD	$0x38,%R9

0x40cbd6 CMP	%R12,%R10

0x40cbd9 JL	40cb90

/home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c: 621 - 630

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

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

622:    {

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

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

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

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

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

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

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

630:       tmp[iTmp].pz =   atoms->p[ii][2];

Coverage (%)	Name	Source Location	Module
►12.69+	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
►12.24+	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
►12.02+	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
►11.87+	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
►11.84+	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
►11.77+	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
►11.73+	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
►11.69+	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:621-630
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	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:621-630
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	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:621-630
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621-630.

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

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)

Vectorization

Your loop is poorly vectorized. Only 14% 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 3.00 to 0.44 cycles (6.86x speedup).

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.

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:

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

Vector unaligned load/store instructions

Detected 4 suboptimal vector unaligned load/store instructions.

Details

MOVUPS: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:625
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:625
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:628
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:628

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

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.

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	17
nb uops	16
loop length	75
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	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)
other	NA (no other vectorizable/vectorized instructions)

all	50%
load	50%
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)

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	6%
load	6%
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	18%
load	18%
store	18%
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: 40cb90

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P14	P15	Latency	Recip. throughput	Vectorization
MOV (%RAX,%R10,4),%R11D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R11D,(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOV (%RCX,%R10,4),%R11D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R11D,0x4(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVUPS -0x10(%RDI,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
MOVUPS %XMM0,0x8(%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%)
MOVSD (%RDI,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
MOVSD %XMM0,0x18(%R9)	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%)
MOVUPS -0x10(%R8,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
MOVUPS %XMM0,0x20(%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%)
MOVSD (%R8,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
MOVSD %XMM0,0x30(%R9)	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%)
INC %R10	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 $0x6,%RDX	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,%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
CMP %R12,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JL 40cb90 <sortAtomsInCell.A+0x70>	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:621-630
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)

Vectorization

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.

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:

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

Vector unaligned load/store instructions

Detected 4 suboptimal vector unaligned load/store instructions.

Details

MOVUPS: 4 occurrences►
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:625
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:625
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:628
- /home/eoseret/qaas_runs_ZEN5/174-043-3878/intel/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:628

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

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.

Unroll opportunity

Loop is data access bound.

Workaround

General properties

nb instructions	17
nb uops	16
loop length	75
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	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)
other	NA (no other vectorizable/vectorized instructions)

all	50%
load	50%
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)

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	6%
load	6%
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	18%
load	18%
store	18%
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: 40cb90

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P14	P15	Latency	Recip. throughput	Vectorization
MOV (%RAX,%R10,4),%R11D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R11D,(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOV (%RCX,%R10,4),%R11D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	scal (6.3%)
MOV %R11D,0x4(%R9)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (6.3%)
MOVUPS -0x10(%RDI,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
MOVUPS %XMM0,0x8(%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%)
MOVSD (%RDI,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
MOVSD %XMM0,0x18(%R9)	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%)
MOVUPS -0x10(%R8,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.50	vect (25.0%)
MOVUPS %XMM0,0x20(%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%)
MOVSD (%R8,%RDX,4),%XMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.50	scal (12.5%)
MOVSD %XMM0,0x30(%R9)	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%)
INC %R10	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 $0x6,%RDX	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,%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
CMP %R12,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JL 40cb90 <sortAtomsInCell.A+0x70>	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

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Vector unaligned load/store instructions

Details

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

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Vector unaligned load/store instructions

Details

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