OV - exec - Function sortAtomsInCell

	Function: sortAtomsInCell	Module: exec	Source: haloExchange.c:614-645	Coverage (incl. loops): 2.59% \| (excl. loops): 0.09%

/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c: 614 - 645

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

614: {

615:    int nAtoms = boxes->nAtoms[iBox];

616:

617:    AtomMsg tmp[nAtoms];

618:

619:    int begin = iBox*MAXATOMS;

620:    int end = begin + nAtoms;

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];

631:    }

632:    qsort(&tmp, nAtoms, sizeof(AtomMsg), sortAtomsById);

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;

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

643:    }

644:

645: }

0x408500 PUSH	%RBP

0x408501 MOV	%RSP,%RBP

0x408504 PUSH	%R15

0x408506 PUSH	%R14

0x408508 PUSH	%R13

0x40850a PUSH	%R12

0x40850c PUSH	%RBX

0x40850d SUB	$0x28,%RSP

0x408511 MOV	0x78(%RSI),%RAX

0x408515 MOVSXD	%EDX,%RCX

0x408518 MOV	%RSP,%RBX

0x40851b MOV	(%RAX,%RCX,4),%R9D

0x40851f IMUL	$0x38,%R9,%RCX

0x408523 LEA	0xf(%RCX),%RAX

0x408527 AND	$-0x10,%RAX

0x40852b SUB	%RAX,%RBX

0x40852e MOV	%RBX,%RSP

0x408531 TEST	%R9D,%R9D

0x408534 JLE	40873d

0x40853a SAL	$0x6,%EDX

0x40853d MOV	0x8(%RDI),%R11

0x408541 LEA	-0x4(%RCX,%RBX,1),%RCX

0x408546 MOV	0x10(%RDI),%R8

0x40854a MOV	%RDI,-0x50(%RBP)

0x40854e MOV	%R9,-0x48(%RBP)

0x408552 MOVSXD	%EDX,%R15

0x408555 LEA	(%R9,%RDX,1),%ESI

0x408559 LEA	(%R9,%R15,1),%R12

0x40855d LEA	0x1(%R15),%RAX

0x408561 CMP	%RAX,%R12

0x408564 CMOVLE	%RAX,%R12

0x408568 LEA	-0x4(%R11,%R12,4),%R10

0x40856d LEA	-0x1(%R12),%RDX

0x408572 LEA	(%R8,%R15,4),%R13

0x408576 MOV	%R12,-0x40(%RBP)

0x40857a LEA	-0x4(%R8,%R12,4),%R12

0x40857f MOV	%R11,-0x38(%RBP)

0x408583 CMP	%RBX,%R10

0x408586 LEA	(%R11,%R15,4),%R10

0x40858a SETAEB	-0x30(%RBP)

0x40858e CMP	%R10,%RCX

0x408591 SETAE	%R14B

0x408595 SUB	%R15,%RDX

0x408598 IMUL	$0x38,%RDX,%RCX

0x40859c LEA	0x34(%RBX,%RCX,1),%RDX

0x4085a1 MOV	0x18(%RDI),%RCX

0x4085a5 CMP	%R13,%RDX

0x4085a8 MOV	0x20(%RDI),%RDX

0x4085ac SETB	%R11B

0x4085b0 CMP	%RBX,%R12

0x4085b3 MOVSXD	%ESI,%R12

0x4085b6 SETB	%SIL

0x4085ba TEST	%R14B,-0x30(%RBP)

0x4085be JNE	408757

0x4085c4 OR	%SIL,%R11B

0x4085c7 JE	408757

0x4085cd MOV	-0x40(%RBP),%RSI

0x4085d1 SUB	%R15,%RSI

0x4085d4 MOV	%RSI,%RDI

0x4085d7 AND	$-0x4,%RDI

0x4085db JE	4087cf

0x4085e1 VPMOVSXBQ	0x17cbe(%RIP),%YMM0

0x4085ea VPMOVSXBQ	0x17cb9(%RIP),%YMM1

0x4085f3 VPMOVSXBQ	0x17cb4(%RIP),%YMM2

0x4085fc VPMOVSXBQ	0x17caf(%RIP),%YMM3

0x408605 MOV	%RDI,-0x30(%RBP)

0x408609 LEA	-0x1(%RDI),%R14

0x40860d MOV	-0x38(%RBP),%RDI

0x408611 MOV	%RSI,-0x40(%RBP)

0x408615 LEA	(,%R15,8),%RSI

0x40861d MOV	$0xf,%R9B

0x408620 KMOVD	%R9D,%K1

0x408625 LEA	(%RSI,%RSI,2),%R11

0x408629 XOR	%ESI,%ESI

0x40862b NOPL	(%RAX,%RAX,1)

(66) 0x408630 VMOVDQU	(%R10,%RSI,4),%XMM4

(66) 0x408636 VPBROADCASTQ	%RSI,%YMM5

(66) 0x40863c VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x408642 VPADDQ	%YMM0,%YMM5,%YMM5

(66) 0x408646 VPSLLQ	$0x3,%YMM5,%YMM6

(66) 0x40864b VPSUBQ	%YMM5,%YMM6,%YMM5

(66) 0x40864f VPSCATTERQD	%XMM4,(%RBX,%YMM5,8){%K2}

(66) 0x408656 VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x40865c VMOVDQU	(%R13,%RSI,4),%XMM4

(66) 0x408663 ADD	$0x4,%RSI

(66) 0x408667 VPSCATTERQD	%XMM4,0x4(%RBX,%YMM5,8){%K2}

(66) 0x40866f VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x408675 VMOVUPD	(%RCX,%R11,1),%ZMM6

(66) 0x40867c VMOVUPD	0x40(%RCX,%R11,1),%ZMM4{%K1}{z}

(66) 0x408684 VMOVAPD	%ZMM6,%ZMM7

(66) 0x40868a VPERMT2PD	%ZMM4,%ZMM1,%ZMM7

(66) 0x408690 VMOVAPD	%ZMM6,%ZMM8

(66) 0x408696 VPERMT2PD	%ZMM4,%ZMM2,%ZMM8

(66) 0x40869c VPERMT2PD	%ZMM4,%ZMM3,%ZMM6

(66) 0x4086a2 VSCATTERQPD	%YMM7,0x8(%RBX,%YMM5,8){%K2}

(66) 0x4086aa VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x4086b0 VSCATTERQPD	%YMM8,0x10(%RBX,%YMM5,8){%K2}

(66) 0x4086b8 VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x4086be VSCATTERQPD	%YMM6,0x18(%RBX,%YMM5,8){%K2}

(66) 0x4086c6 VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x4086cc VMOVUPD	(%RDX,%R11,1),%ZMM6

(66) 0x4086d3 VMOVUPD	0x40(%RDX,%R11,1),%ZMM4{%K1}{z}

(66) 0x4086db ADD	$0x60,%R11

(66) 0x4086df VMOVAPD	%ZMM6,%ZMM7

(66) 0x4086e5 VPERMT2PD	%ZMM4,%ZMM1,%ZMM7

(66) 0x4086eb VMOVAPD	%ZMM6,%ZMM8

(66) 0x4086f1 VPERMT2PD	%ZMM4,%ZMM2,%ZMM8

(66) 0x4086f7 VPERMT2PD	%ZMM4,%ZMM3,%ZMM6

(66) 0x4086fd VSCATTERQPD	%YMM7,0x20(%RBX,%YMM5,8){%K2}

(66) 0x408705 VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x40870b VSCATTERQPD	%YMM8,0x28(%RBX,%YMM5,8){%K2}

(66) 0x408713 VPCMPEQB	%XMM0,%XMM0,%K2

(66) 0x408719 VSCATTERQPD	%YMM6,0x30(%RBX,%YMM5,8){%K2}

(66) 0x408721 CMP	%R14,%RSI

(66) 0x408724 JLE	408630

0x40872a MOV	-0x30(%RBP),%R10

0x40872e CMP	%R10,-0x40(%RBP)

0x408732 JNE	4087d6

0x408738 JMP	40884d

0x40873d MOVSXD	%R9D,%RSI

0x408740 MOV	$0x38,%EDX

0x408745 MOV	$0x4088f0,%ECX

0x40874a MOV	%RBX,%RDI

0x40874d CALL	403210 <qsort@plt>

0x408752 JMP	4088d9

0x408757 MOV	-0x38(%RBP),%RDI

0x40875b LEA	(,%R15,8),%RSI

0x408763 CMP	%RAX,%R12

0x408766 MOV	%RBX,%R9

0x408769 MOV	%R15,%R10

0x40876c CMOVG	%R12,%RAX

0x408770 LEA	(%RSI,%RSI,2),%RSI

0x408774 NOPW	%CS:(%RAX,%RAX,1)

(63) 0x408780 MOV	(%RDI,%R10,4),%R11D

(63) 0x408784 MOV	%R11D,(%R9)

(63) 0x408787 MOV	(%R8,%R10,4),%R11D

(63) 0x40878b INC	%R10

(63) 0x40878e MOV	%R11D,0x4(%R9)

(63) 0x408792 VMOVSD	(%RCX,%RSI,1),%XMM0

(63) 0x408797 VMOVSD	%XMM0,0x8(%R9)

(63) 0x40879d VMOVUPS	0x8(%RCX,%RSI,1),%XMM0

(63) 0x4087a3 VMOVUPD	(%RDX,%RSI,1),%XMM1

(63) 0x4087a8 VMOVUPD	%XMM1,0x20(%R9)

(63) 0x4087ae VMOVUPS	%XMM0,0x10(%R9)

(63) 0x4087b4 VMOVQ	0x10(%RDX,%RSI,1),%XMM0

(63) 0x4087ba ADD	$0x18,%RSI

(63) 0x4087be VMOVQ	%XMM0,0x30(%R9)

(63) 0x4087c4 ADD	$0x38,%R9

(63) 0x4087c8 CMP	%R10,%RAX

(63) 0x4087cb JNE	408780

0x4087cd JMP	40884d

0x4087cf MOV	-0x38(%RBP),%RDI

0x4087d3 XOR	%R10D,%R10D

0x4087d6 CMP	%RAX,%R12

0x4087d9 CMOVG	%R12,%RAX

0x4087dd IMUL	$0x38,%R10,%RSI

0x4087e1 SUB	%R10,%RAX

0x4087e4 ADD	%R15,%R10

0x4087e7 LEA	(%R10,%R10,1),%R9

0x4087eb LEA	(%R8,%R10,4),%R8

0x4087ef LEA	(%RDI,%R10,4),%RDI

0x4087f3 SUB	%R15,%RAX

0x4087f6 XOR	%R10D,%R10D

0x4087f9 LEA	(%R9,%R9,2),%R9

0x4087fd ADD	%RBX,%RSI

(65) 0x408800 MOV	(%RDI,%R10,4),%R11D

(65) 0x408804 MOV	%R11D,(%RSI)

(65) 0x408807 MOV	(%R8,%R10,4),%R11D

(65) 0x40880b INC	%R10

(65) 0x40880e MOV	%R11D,0x4(%RSI)

(65) 0x408812 VMOVSD	(%RCX,%R9,4),%XMM0

(65) 0x408818 VMOVSD	%XMM0,0x8(%RSI)

(65) 0x40881d VMOVUPS	0x8(%RCX,%R9,4),%XMM0

(65) 0x408824 VMOVUPD	(%RDX,%R9,4),%XMM1

(65) 0x40882a VMOVUPD	%XMM1,0x20(%RSI)

(65) 0x40882f VMOVUPS	%XMM0,0x10(%RSI)

(65) 0x408834 VMOVQ	0x10(%RDX,%R9,4),%XMM0

(65) 0x40883b ADD	$0x6,%R9

(65) 0x40883f VMOVQ	%XMM0,0x30(%RSI)

(65) 0x408844 ADD	$0x38,%RSI

(65) 0x408848 CMP	%R10,%RAX

(65) 0x40884b JNE	408800

0x40884d MOV	-0x48(%RBP),%RSI

0x408851 MOV	$0x38,%EDX

0x408856 MOV	$0x4088f0,%ECX

0x40885b MOV	%RBX,%RDI

0x40885e VZEROUPPER

0x408861 CALL	403210 <qsort@plt>

0x408866 MOV	-0x50(%RBP),%RSI

0x40886a LEA	(%R15,%R15,2),%RDI

0x40886e LEA	0x10(,%RDI,8),%RDI

0x408876 MOV	0x8(%RSI),%RAX

0x40887a MOV	0x10(%RSI),%RCX

0x40887e MOV	0x18(%RSI),%RDX

0x408882 MOV	0x20(%RSI),%RSI

0x408886 NOPW	%CS:(%RAX,%RAX,1)

(64) 0x408890 MOV	(%RBX),%R8D

(64) 0x408893 MOV	%R8D,(%RAX,%R15,4)

(64) 0x408897 MOV	0x4(%RBX),%R8D

(64) 0x40889b MOV	%R8D,(%RCX,%R15,4)

(64) 0x40889f INC	%R15

(64) 0x4088a2 VMOVUPS	0x8(%RBX),%XMM0

(64) 0x4088a7 VMOVUPS	%XMM0,-0x10(%RDX,%RDI,1)

(64) 0x4088ad VMOVSD	0x18(%RBX),%XMM0

(64) 0x4088b2 VMOVSD	%XMM0,(%RDX,%RDI,1)

(64) 0x4088b7 VMOVUPS	0x20(%RBX),%XMM0

(64) 0x4088bc VMOVUPS	%XMM0,-0x10(%RSI,%RDI,1)

(64) 0x4088c2 VMOVQ	0x30(%RBX),%XMM0

(64) 0x4088c7 ADD	$0x38,%RBX

(64) 0x4088cb VMOVQ	%XMM0,(%RSI,%RDI,1)

(64) 0x4088d0 ADD	$0x18,%RDI

(64) 0x4088d4 CMP	%R12,%R15

(64) 0x4088d7 JL	408890

0x4088d9 LEA	-0x28(%RBP),%RSP

0x4088dd POP	%RBX

0x4088de POP	%R12

0x4088e0 POP	%R13

0x4088e2 POP	%R14

0x4088e4 POP	%R15

0x4088e6 POP	%RBP

0x4088e7 RET

0x4088e8 NOPL	(%RAX,%RAX,1)

Coverage (%)	Name	Source Location	Module
►13.13+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►12.91+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►12.78+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►12.66+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►12.25+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►12.15+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►12.01+	redistributeAtoms.extracted	timestep.c:154	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so
►11.56+	redistributeAtoms.extracted	timestep.c:153	exec
○	__kmp_invoke_microtask		libiomp5.so
○	__kmp_invoke_task_func		libiomp5.so

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 /

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

The code analyzed by CQA in that panel excludes loops and represents 0.09% of application time for run icx_0

Source file and lines	haloExchange.c:614-645
Module	exec

The function is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:614-645.

Warnings:
Ignoring paths for analysis

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 17.13 to 3.38 cycles (5.07x speedup).

Workaround

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

Vectorization

Your function is probably not vectorized. Only 11% of vector register length is used (average across all SSE/AVX instructions).

Details

Store and arithmetical SSE/AVX instructions are used in scalar version (process only one data element in vector registers).

Workaround

Read the "512-bits vectorization" report at "Potential" confidence level.

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

512-bits vectorization

On some x86 processors supporting 512-bits vectorization, compilers are often too conservative and limit vectorization to 256 bits. Performance can then be improved by enforcing 512-bits vectorization, especially with many vectorized and high trip count loops. 512-bits vectorization performance overhead (compared to 256-bits) is generally lower on newer processors.

Workaround

Recompile with -mprefer-vector-width=512

CALL instructions

Detected function call instructions.

Details

Calling (and then returning from) a function prevents many compiler optimizations (like vectorization), breaks control flow (which reduces pipeline performance) and executes extra instructions to save/restore the registers used inside it, which is very expensive (dozens of cycles). Consider to inline small functions.

qsort@plt: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:632
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:632

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.

CMOVG: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:632
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
CMOVLE: 1 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:623
KMOVD: 1 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
LEA: 4 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:623
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:624
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621

Conversion instructions

Detected expensive conversion instructions.

Details

VPMOVSXBQ: 4 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:625
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:626
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:627

Workaround

Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.

Type of elements and instruction set

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

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 157 bytes. The binary function is storing 49 bytes.

General properties

nb instructions	131
nb uops	137
loop length	531
used x86 registers	16
used mmx registers	0
used xmm registers	0
used ymm registers	4
used zmm registers	0
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	17.13 cycles
front end	17.13 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	13.00	13.00	13.00	13.00	13.00	13.00	9.00	9.00	9.00	9.00	0.00	2.50	2.50	0.00	0.00	0.00
cycles	13.00	13.00	13.00	13.00	13.00	13.00	9.00	9.00	9.00	9.00	0.00	2.50	2.50	0.00	0.00	0.00

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	17.13
Dispatch	13.00
Overall L1	17.13

Vectorization ratios

all	22%
load	80%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	9%

Vector efficiency ratios

all	11%
load	7%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	13%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 17.13 cycles. At this rate:

3% of peak load performance is reached (9.17 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (2.86 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 17.13 to 13.00 cycles (1.32x speedup).

ASM code

In the binary file, the address of the function is: 408500

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P11	P12	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R13	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R12	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB $0x28,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV 0x78(%RSI),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOVSXD %EDX,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RSP,%RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RAX,%RCX,4),%R9D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
IMUL $0x38,%R9,%RCX	1	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	3	0.33	N/A
LEA 0xf(%RCX),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
AND $-0x10,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
SUB %RAX,%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
MOV %RBX,%RSP	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
TEST %R9D,%R9D	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (6.3%)
JLE 40873d <sortAtomsInCell+0x23d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
SAL $0x6,%EDX	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
MOV 0x8(%RDI),%R11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
LEA -0x4(%RCX,%RBX,1),%RCX	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
MOV 0x10(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV %RDI,-0x50(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
MOV %R9,-0x48(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
MOVSXD %EDX,%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
LEA (%R9,%RDX,1),%ESI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R9,%R15,1),%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
LEA 0x1(%R15),%RAX	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 %RAX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMOVLE %RAX,%R12	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
LEA -0x4(%R11,%R12,4),%R10	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
LEA -0x1(%R12),%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
LEA (%R8,%R15,4),%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %R12,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
LEA -0x4(%R8,%R12,4),%R12	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
MOV %R11,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
CMP %RBX,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
LEA (%R11,%R15,4),%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
SETAEB -0x30(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	1	0.50	N/A
CMP %R10,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
SETAE %R14B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
SUB %R15,%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
IMUL $0x38,%RDX,%RCX	1	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	3	0.33	N/A
LEA 0x34(%RBX,%RCX,1),%RDX	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
MOV 0x18(%RDI),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
CMP %R13,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV 0x20(%RDI),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
SETB %R11B	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 %RBX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOVSXD %ESI,%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
SETB %SIL	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
TEST %R14B,-0x30(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	1	0.25	N/A
JNE 408757 <sortAtomsInCell+0x257>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
OR %SIL,%R11B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JE 408757 <sortAtomsInCell+0x257>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
MOV -0x40(%RBP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
SUB %R15,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RSI,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x4,%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
JE 4087cf <sortAtomsInCell+0x2cf>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
VPMOVSXBQ 0x17cbe(%RIP),%YMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
VPMOVSXBQ 0x17cb9(%RIP),%YMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
VPMOVSXBQ 0x17cb4(%RIP),%YMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
VPMOVSXBQ 0x17caf(%RIP),%YMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
MOV %RDI,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
LEA -0x1(%RDI),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV -0x38(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV %RSI,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
LEA (,%R15,8),%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0xf,%R9B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
KMOVD %R9D,%K1	2	0	0	0	0	0	0	0	0	0	0	0.50	0.50	2	1	N/A
LEA (%RSI,%RSI,2),%R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
XOR %ESI,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
MOV -0x30(%RBP),%R10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
CMP %R10,-0x40(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
JNE 4087d6 <sortAtomsInCell+0x2d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
JMP 40884d <sortAtomsInCell+0x34d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
MOVSXD %R9D,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x38,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x4088f0,%ECX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CALL 403210 <qsort@plt>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0	0	0.50	N/A
JMP 4088d9 <sortAtomsInCell+0x3d9>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
MOV -0x38(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
LEA (,%R15,8),%RSI	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 %RAX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV %RBX,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %R15,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CMOVG %R12,%RAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
LEA (%RSI,%RSI,2),%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
JMP 40884d <sortAtomsInCell+0x34d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
MOV -0x38(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
XOR %R10D,%R10D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CMP %RAX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMOVG %R12,%RAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
IMUL $0x38,%R10,%RSI	1	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	3	0.33	N/A
SUB %R10,%RAX	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 %R15,%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
LEA (%R10,%R10,1),%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
LEA (%R8,%R10,4),%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
LEA (%RDI,%R10,4),%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
SUB %R15,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
XOR %R10D,%R10D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%R9,%R9,2),%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
ADD %RBX,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV -0x48(%RBP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV $0x38,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x4088f0,%ECX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALL 403210 <qsort@plt>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0	0	0.50	N/A
MOV -0x50(%RBP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
LEA (%R15,%R15,2),%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
LEA 0x10(,%RDI,8),%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
MOV 0x8(%RSI),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV 0x10(%RSI),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV 0x18(%RSI),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV 0x20(%RSI),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
LEA -0x28(%RBP),%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
RET	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50	N/A
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A

The code analyzed by CQA in that panel excludes loops and represents 0.09% of application time for run icx_0

Source file and lines	haloExchange.c:614-645
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 17.13 to 3.38 cycles (5.07x speedup).

Workaround

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

Vectorization

Your function is probably not vectorized. Only 11% of vector register length is used (average across all SSE/AVX instructions).

Details

Store and arithmetical SSE/AVX instructions are used in scalar version (process only one data element in vector registers).

Workaround

Read the "512-bits vectorization" report at "Potential" confidence level.

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

512-bits vectorization

Workaround

Recompile with -mprefer-vector-width=512

CALL instructions

Detected function call instructions.

Details

qsort@plt: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:632
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:632

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.

CMOVG: 2 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:632
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
CMOVLE: 1 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:623
KMOVD: 1 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
LEA: 4 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:623
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:624
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621

Conversion instructions

Detected expensive conversion instructions.

Details

VPMOVSXBQ: 4 occurrences►
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:621
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:625
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:626
- /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz12.benchmarkcenter.megware.com/177-374-1600/CoMD/build/CoMD/CoMD/src-openmp/haloExchange.c:627

Workaround

Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.

Type of elements and instruction set

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

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 157 bytes. The binary function is storing 49 bytes.

General properties

nb instructions	131
nb uops	137
loop length	531
used x86 registers	16
used mmx registers	0
used xmm registers	0
used ymm registers	4
used zmm registers	0
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	17.13 cycles
front end	17.13 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	13.00	13.00	13.00	13.00	13.00	13.00	9.00	9.00	9.00	9.00	0.00	2.50	2.50	0.00	0.00	0.00
cycles	13.00	13.00	13.00	13.00	13.00	13.00	9.00	9.00	9.00	9.00	0.00	2.50	2.50	0.00	0.00	0.00

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	17.13
Dispatch	13.00
Overall L1	17.13

Vectorization ratios

all	22%
load	80%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	9%

Vector efficiency ratios

all	11%
load	7%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	13%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 17.13 cycles. At this rate:

3% of peak load performance is reached (9.17 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (2.86 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the function is: 408500

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P11	P12	Latency	Recip. throughput	Vectorization
PUSH %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
PUSH %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R13	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %R12	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
PUSH %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
SUB $0x28,%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV 0x78(%RSI),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOVSXD %EDX,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RSP,%RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV (%RAX,%RCX,4),%R9D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
IMUL $0x38,%R9,%RCX	1	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	3	0.33	N/A
LEA 0xf(%RCX),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
AND $-0x10,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
SUB %RAX,%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
MOV %RBX,%RSP	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
TEST %R9D,%R9D	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (6.3%)
JLE 40873d <sortAtomsInCell+0x23d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
SAL $0x6,%EDX	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
MOV 0x8(%RDI),%R11	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
LEA -0x4(%RCX,%RBX,1),%RCX	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
MOV 0x10(%RDI),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV %RDI,-0x50(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
MOV %R9,-0x48(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
MOVSXD %EDX,%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
LEA (%R9,%RDX,1),%ESI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
LEA (%R9,%R15,1),%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
LEA 0x1(%R15),%RAX	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 %RAX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMOVLE %RAX,%R12	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
LEA -0x4(%R11,%R12,4),%R10	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
LEA -0x1(%R12),%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
LEA (%R8,%R15,4),%R13	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %R12,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
LEA -0x4(%R8,%R12,4),%R12	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
MOV %R11,-0x38(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
CMP %RBX,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
LEA (%R11,%R15,4),%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
SETAEB -0x30(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	1	0.50	N/A
CMP %R10,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
SETAE %R14B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
SUB %R15,%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
IMUL $0x38,%RDX,%RCX	1	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	3	0.33	N/A
LEA 0x34(%RBX,%RCX,1),%RDX	2	0	0	0.50	0.50	0	0	0.25	0.25	0.25	0.25	0	0	2	0.50	N/A
MOV 0x18(%RDI),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
CMP %R13,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV 0x20(%RDI),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
SETB %R11B	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 %RBX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOVSXD %ESI,%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
SETB %SIL	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
TEST %R14B,-0x30(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	1	0.25	N/A
JNE 408757 <sortAtomsInCell+0x257>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
OR %SIL,%R11B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
JE 408757 <sortAtomsInCell+0x257>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
MOV -0x40(%RBP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
SUB %R15,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RSI,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x4,%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
JE 4087cf <sortAtomsInCell+0x2cf>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
VPMOVSXBQ 0x17cbe(%RIP),%YMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
VPMOVSXBQ 0x17cb9(%RIP),%YMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
VPMOVSXBQ 0x17cb4(%RIP),%YMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
VPMOVSXBQ 0x17caf(%RIP),%YMM3	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50	vect (6.3%)
MOV %RDI,-0x30(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
LEA -0x1(%RDI),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV -0x38(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV %RSI,-0x40(%RBP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	4	0.50	scal (12.5%)
LEA (,%R15,8),%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0xf,%R9B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
KMOVD %R9D,%K1	2	0	0	0	0	0	0	0	0	0	0	0.50	0.50	2	1	N/A
LEA (%RSI,%RSI,2),%R11	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
XOR %ESI,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
MOV -0x30(%RBP),%R10	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
CMP %R10,-0x40(%RBP)	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	1	0.25	scal (12.5%)
JNE 4087d6 <sortAtomsInCell+0x2d6>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.33-1	N/A
JMP 40884d <sortAtomsInCell+0x34d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
MOVSXD %R9D,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x38,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x4088f0,%ECX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CALL 403210 <qsort@plt>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0	0	0.50	N/A
JMP 4088d9 <sortAtomsInCell+0x3d9>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
MOV -0x38(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
LEA (,%R15,8),%RSI	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 %RAX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV %RBX,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %R15,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CMOVG %R12,%RAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
LEA (%RSI,%RSI,2),%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
JMP 40884d <sortAtomsInCell+0x34d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0.50	N/A
MOV -0x38(%RBP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
XOR %R10D,%R10D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CMP %RAX,%R12	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	scal (12.5%)
CMOVG %R12,%RAX	2	0.33	0.33	0.33	0.33	0.33	0.33	0	0	0	0	0	0	1	0.33	N/A
IMUL $0x38,%R10,%RSI	1	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	3	0.33	N/A
SUB %R10,%RAX	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 %R15,%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
LEA (%R10,%R10,1),%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
LEA (%R8,%R10,4),%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
LEA (%RDI,%R10,4),%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
SUB %R15,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
XOR %R10D,%R10D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
LEA (%R9,%R9,2),%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
ADD %RBX,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV -0x48(%RBP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV $0x38,%EDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV $0x4088f0,%ECX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALL 403210 <qsort@plt>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0	0	0.50	N/A
MOV -0x50(%RBP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
LEA (%R15,%R15,2),%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
LEA 0x10(,%RDI,8),%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
MOV 0x8(%RSI),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV 0x10(%RSI),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV 0x18(%RSI),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
MOV 0x20(%RSI),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	3	0.25	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
LEA -0x28(%RBP),%RSP	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	1	0.17	N/A
POP %RBX	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R12	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R13	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R14	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %R15	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
POP %RBP	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.25	N/A
RET	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	1	0.50	N/A
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A

Name	Coverage (%)	Time (s)
▼sortAtomsInCell–	2.59	0.19
○Loop 66 - haloExchange.c:621-630 - exec	2.39	0.17
○Loop 64 - haloExchange.c:633-642 - exec	0.09	0.02
○Loop 65 - haloExchange.c:621-630 - exec	0.03	0.01
○Loop 63 - haloExchange.c:621-630 - exec	0.00	0.00

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

512-bits vectorization

Workaround

CALL instructions

Details

Complex instructions

Details

Conversion instructions

Details

Workaround

Type of elements and instruction set

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

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

512-bits vectorization

Workaround

CALL instructions

Details

Complex instructions

Details

Conversion instructions

Details

Workaround

Type of elements and instruction set

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code