OV - exec - Function miniqmcreference::OneBodyJastrowRef<qmcplusplus::BsplineFunctor<double> >::recompute(qmcplusplus::ParticleSet&)

	Function: miniqmcreference::OneBodyJastrowRef<qmcplusplus::BsplineFunctor<double> >::recompute(qmcpl ...	Module: exec	Source: OneBodyJastrowRef.h:111-198 [...]	Coverage: 0.01%

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_numeric.h: 140 - 141

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

140:       for (; __first != __last; ++__first)

141: 	__init = _GLIBCXX_MOVE_IF_20(__init) + *__first;

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h: 1046 - 1046

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

1046: 	return *(this->_M_impl._M_start + __n);

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsMatrix.h: 214 - 214

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

214:   inline const Type_t* operator[](size_type i) const { return X.data() + i * D2; }

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/OneBodyJastrowRef.h: 111 - 198

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

111:   {

112:     const DistanceTableData& d_ie(*(P.DistTables[myTableID]));

113:     for (int iat = 0; iat < Nelec; ++iat)

114:     {

115:       computeU3(P, iat, d_ie.Distances[iat]);

116:       Vat[iat] = std::accumulate(U.begin(), U.begin() + Nions, valT());

117:       Lap[iat] = accumulateGL(dU.data(), d2U.data(), d_ie.Displacements[iat], Grad[iat]);

118:     }

119:   }

[...]

190:     for (int jat = 0; jat < Nions; ++jat)

191:       lap += d2u[jat] + lapfac * du[jat];

192:     for (int idim = 0; idim < OHMMS_DIM; ++idim)

193:     {

194:       const valT* restrict dX = displ.data(idim);

195:       valT s                  = valT();

196:       for (int jat = 0; jat < Nions; ++jat)

197:         s += du[jat] * dX[jat];

198:       grad[idim] = s;

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h: 223 - 249

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

223:     return X[i];

[...]

249:   inline const_pointer data() const { return X; }

0x4173c0 STP	X29, X30, [SP, #944]!

0x4173c4 STR	X25, [SP, #16]

0x4173c8 STP	X24, X23, [SP, #32]

0x4173cc STP	X22, X21, [SP, #48]

0x4173d0 STP	X20, X19, [SP, #64]

0x4173d4 ADD	X29, SP, #0

0x4173d8 LDR	W8, [X0, #152]

0x4173dc CMP	W8, #1

0x4173e0 B.LT	4177d8

0x4173e4 LDR	X8, [X1, #2576]

0x4173e8 LDRSW	X9, [X0, #144]

0x4173ec ORR	X19, XZR, X1

0x4173f0 ORR	X20, XZR, X0

0x4173f4 MOVZ	W23, #24

0x4173f8 MOVZ	W25, #40

0x4173fc LDR	X22, [X8, X9,LSL #3]

0x417400 ORR	X21, XZR, XZR

0x417404 CNTW	X24, ALL

0x417408 B	417430

    (326) 0x41740c MOVI	D0, #0

    (326) 0x417410 STP	XZR, XZR, [X8]

    (326) 0x417414 STR	XZR, [X8, #16]

    (326) 0x417418 LDR	X8, [X20, #432]

    (326) 0x41741c STR	D0, [X8, X21,LSL #3]

    (326) 0x417420 LDRSW	X8, [X20, #152]

    (326) 0x417424 ADD	X21, X21, #1

    (326) 0x417428 CMP	X21, X8

    (326) 0x41742c B.GE	4177d8

    (326) 0x417430 LDR	X9, [X22, #24]

    (326) 0x417434 LDR	X8, [X22, #64]

    (326) 0x417438 ORR	X0, XZR, X20

    (326) 0x41743c ORR	X1, XZR, X19

    (326) 0x417440 ORR	W2, WZR, W21

    (326) 0x417444 MADD	X9, X9, X21, XZR

    (326) 0x417448 ADD	X3, X8, X9,LSL #3

    (326) 0x41744c BL	4165a0

    (326) 0x417450 LDRSW	X9, [X20, #148]

    (326) 0x417454 CBZ	W9, 417484

0x417458 LDR	X8, [X20, #248]

0x41745c SUB	X10, X9, #1

0x417460 MOVI	D0, #0

0x417464 AND	X10, X10, #7936

0x417468 ADD	X10, X10, #1

0x41746c CMP	X10, X24

0x417470 B.CS	4174a0

0x417474 ORR	X10, XZR, X8

0x417478 PTRUE	P0.D, ALL

0x41747c FDUP	Z6.D, #0

0x417480 B	4174f0

    (326) 0x417484 MOVI	D0, #0

    (326) 0x417488 PTRUE	P0.D, ALL

    (326) 0x41748c FDUP	Z6.D, #0

    (326) 0x417490 B	417510

0x417494 HINT	#0

0x417498 HINT	#0

0x41749c HINT	#0

0x4174a0 UDIV	X11, X10, X24

0x4174a4 DUP	Z1.D, #0

0x4174a8 ORR	X12, XZR, XZR

0x4174ac ADDVL	X14, X8, #1

0x4174b0 PTRUE	P0.D, ALL

0x4174b4 ORR	Z0.D, Z1.D, Z1.D

0x4174b8 MADD	X13, X11, X24, XZR

0x4174bc SUB	X11, X10, X13

0x4174c0 ADD	X10, X8, X13,LSL #3

(336) 0x4174c4 LD1D	{Z2.D}, P0/Z, [X8, X12,LSL #3]

(336) 0x4174c8 LD1D	{Z3.D}, P0/Z, [X14, X12,LSL #3]

(336) 0x4174cc ADD	X12, X12, X24

(336) 0x4174d0 CMP	X13, X12

(336) 0x4174d4 FADD	Z0.D, Z2.D, Z0.D

(336) 0x4174d8 FADD	Z1.D, Z3.D, Z1.D

(336) 0x4174dc B.NE	4174c4

0x4174e0 FADD	Z0.D, Z1.D, Z0.D

0x4174e4 FDUP	Z6.D, #0

0x4174e8 FADDV	D0, P0, Z0.D

0x4174ec CBZ	X11, 417510

0x4174f0 ADD	X8, X8, X9,LSL #3

0x4174f4 HINT	#0

0x4174f8 HINT	#0

0x4174fc HINT	#0

(337) 0x417500 LDR	D1, [X10], #8

(337) 0x417504 FADD	D0, D1, D0

(337) 0x417508 CMP	X10, X8

(337) 0x41750c B.NE	417500

    (326) 0x417510 LDR	X8, [X20, #232]

    (326) 0x417514 CMP	W9, #1

    (326) 0x417518 STR	D0, [X8, X21,LSL #3]

    (326) 0x41751c LDR	X8, [X20, #392]

    (326) 0x417520 MADD	X8, X21, X23, X8

    (326) 0x417524 B.LT	41740c

  (332) 0x417528 AND	X10, X9, #6080

  (332) 0x41752c LDR	X9, [X20, #272]

  (332) 0x417530 LDR	X11, [X20, #296]

  (332) 0x417534 CMP	X24, X10

  (332) 0x417538 B.LS	417548

  (332) 0x41753c ORR	X12, XZR, XZR

  (332) 0x417540 MOVI	D0, #0

  (332) 0x417544 B	4175a0

  (332) 0x417548 UDIV	X12, X10, X24

  (332) 0x41754c DUP	Z1.D, #0

  (332) 0x417550 ORR	X14, XZR, XZR

  (332) 0x417554 ADDVL	X15, X11, #1

  (332) 0x417558 ADDVL	X16, X9, #1

  (332) 0x41755c ORR	Z0.D, Z1.D, Z1.D

  (332) 0x417560 UMSUBL	X13, W12, W24, X10

  (332) 0x417564 UMADDL	X12, W12, W24, XZR

    (327) 0x417568 LD1D	{Z2.D}, P0/Z, [X11, X14,LSL #3]

    (327) 0x41756c LD1D	{Z3.D}, P0/Z, [X15, X14,LSL #3]

    (327) 0x417570 FADD	Z0.D, Z2.D, Z0.D

    (327) 0x417574 LD1D	{Z4.D}, P0/Z, [X9, X14,LSL #3]

    (327) 0x417578 LD1D	{Z5.D}, P0/Z, [X16, X14,LSL #3]

    (327) 0x41757c FADD	Z1.D, Z3.D, Z1.D

    (327) 0x417580 ADD	X14, X14, X24

    (327) 0x417584 CMP	X12, X14

    (327) 0x417588 FMLA	Z0.D, P0/M, Z4.D, Z6.D

    (327) 0x41758c FMLA	Z1.D, P0/M, Z5.D, Z6.D

    (327) 0x417590 B.NE	417568

  (332) 0x417594 FADD	Z0.D, Z1.D, Z0.D

  (332) 0x417598 FADDV	D0, P0, Z0.D

  (332) 0x41759c CBZ	X13, 4175dc

  (332) 0x4175a0 UBFM	X14, X12, #61, #60

  (332) 0x4175a4 SUB	X13, X10, X12

  (332) 0x4175a8 ADD	X12, X9, X14

  (332) 0x4175ac ADD	X11, X11, X14

  (332) 0x4175b0 HINT	#0

  (332) 0x4175b4 HINT	#0

  (332) 0x4175b8 HINT	#0

  (332) 0x4175bc HINT	#0

    (335) 0x4175c0 LDR	D1, [X11], #8

    (335) 0x4175c4 LDR	D2, [X12], #8

    (335) 0x4175c8 FADD	D2, D2, D2

    (335) 0x4175cc FADD	D0, D1, D0

    (335) 0x4175d0 SUBS	X13, X13, #1

    (335) 0x4175d4 FADD	D0, D0, D2

    (335) 0x4175d8 B.NE	4175c0

  (332) 0x4175dc LDR	X11, [X22, #80]

  (332) 0x4175e0 CMP	X24, X10

  (332) 0x4175e4 MADD	X12, X21, X25, X11

  (332) 0x4175e8 LDR	X11, [X12, #24]

  (332) 0x4175ec B.LS	417600

  (332) 0x4175f0 ORR	X13, XZR, XZR

  (332) 0x4175f4 MOVI	D1, #0

  (332) 0x4175f8 B	417650

0x4175fc HINT	#0

  (332) 0x417600 UDIV	X13, X10, X24

  (332) 0x417604 DUP	Z2.D, #0

  (332) 0x417608 ORR	X15, XZR, XZR

  (332) 0x41760c ADDVL	X16, X9, #1

  (332) 0x417610 ADDVL	X17, X11, #1

  (332) 0x417614 ORR	Z1.D, Z2.D, Z2.D

  (332) 0x417618 UMSUBL	X14, W13, W24, X10

  (332) 0x41761c UMADDL	X13, W13, W24, XZR

    (328) 0x417620 LD1D	{Z3.D}, P0/Z, [X9, X15,LSL #3]

    (328) 0x417624 LD1D	{Z4.D}, P0/Z, [X16, X15,LSL #3]

    (328) 0x417628 LD1D	{Z5.D}, P0/Z, [X11, X15,LSL #3]

    (328) 0x41762c FMLA	Z1.D, P0/M, Z5.D, Z3.D

    (328) 0x417630 LD1D	{Z6.D}, P0/Z, [X17, X15,LSL #3]

    (328) 0x417634 FMLA	Z2.D, P0/M, Z6.D, Z4.D

    (328) 0x417638 ADD	X15, X15, X24

    (328) 0x41763c CMP	X13, X15

    (328) 0x417640 B.NE	417620

  (332) 0x417644 FADD	Z1.D, Z2.D, Z1.D

  (332) 0x417648 FADDV	D1, P0, Z1.D

  (332) 0x41764c CBZ	X14, 417674

  (332) 0x417650 UBFM	X15, X13, #61, #60

  (332) 0x417654 SUB	X14, X10, X13

  (332) 0x417658 ADD	X13, X11, X15

  (332) 0x41765c ADD	X15, X9, X15

    (334) 0x417660 LDR	D2, [X15], #8

    (334) 0x417664 LDR	D3, [X13], #8

    (334) 0x417668 FMADD	D1, D3, D2, D1

    (334) 0x41766c SUBS	X14, X14, #1

    (334) 0x417670 B.NE	417660

  (332) 0x417674 LDR	X12, [X12, #8]

  (332) 0x417678 CMP	X24, X10

  (332) 0x41767c STR	D1, [X8]

  (332) 0x417680 B.LS	4176a0

  (332) 0x417684 ORR	X13, XZR, XZR

  (332) 0x417688 MOVI	D1, #0

  (332) 0x41768c B	4176f4

0x417690 HINT	#0

0x417694 HINT	#0

0x417698 HINT	#0

0x41769c HINT	#0

  (332) 0x4176a0 UDIV	X13, X10, X24

  (332) 0x4176a4 ADD	X16, X11, X12,LSL #3

  (332) 0x4176a8 DUP	Z2.D, #0

  (332) 0x4176ac ORR	X15, XZR, XZR

  (332) 0x4176b0 ADDVL	X17, X9, #1

  (332) 0x4176b4 ADDVL	X18, X16, #1

  (332) 0x4176b8 ORR	Z1.D, Z2.D, Z2.D

  (332) 0x4176bc UMSUBL	X14, W13, W24, X10

  (332) 0x4176c0 UMADDL	X13, W13, W24, XZR

    (329) 0x4176c4 LD1D	{Z3.D}, P0/Z, [X9, X15,LSL #3]

    (329) 0x4176c8 LD1D	{Z4.D}, P0/Z, [X17, X15,LSL #3]

    (329) 0x4176cc LD1D	{Z5.D}, P0/Z, [X16, X15,LSL #3]

    (329) 0x4176d0 FMLA	Z1.D, P0/M, Z5.D, Z3.D

    (329) 0x4176d4 LD1D	{Z6.D}, P0/Z, [X18, X15,LSL #3]

    (329) 0x4176d8 FMLA	Z2.D, P0/M, Z6.D, Z4.D

    (329) 0x4176dc ADD	X15, X15, X24

    (329) 0x4176e0 CMP	X13, X15

    (329) 0x4176e4 B.NE	4176c4

  (332) 0x4176e8 FADD	Z1.D, Z2.D, Z1.D

  (332) 0x4176ec FADDV	D1, P0, Z1.D

  (332) 0x4176f0 CBZ	X14, 41771c

  (332) 0x4176f4 UBFM	X15, X13, #61, #60

  (332) 0x4176f8 SUB	X14, X10, X13

  (332) 0x4176fc ADD	X13, X15, X12,LSL #3

  (332) 0x417700 ADD	X15, X9, X15

  (332) 0x417704 ADD	X13, X11, X13

    (333) 0x417708 LDR	D2, [X15], #8

    (333) 0x41770c LDR	D3, [X13], #8

    (333) 0x417710 FMADD	D1, D3, D2, D1

    (333) 0x417714 SUBS	X14, X14, #1

    (333) 0x417718 B.NE	417708

  (332) 0x41771c CMP	X24, X10

  (332) 0x417720 STR	D1, [X8, #8]

  (332) 0x417724 B.LS	417740

  (332) 0x417728 ORR	X13, XZR, XZR

  (332) 0x41772c MOVI	D1, #0

  (332) 0x417730 B	417794

0x417734 HINT	#0

0x417738 HINT	#0

0x41773c HINT	#0

(330) 0x417740 UDIV	X13, X10, X24

(330) 0x417744 ADD	X16, X11, X12,LSL #4

(330) 0x417748 DUP	Z2.D, #0

(330) 0x41774c ORR	X15, XZR, XZR

(330) 0x417750 ADDVL	X17, X9, #1

(330) 0x417754 ADDVL	X18, X16, #1

(330) 0x417758 ORR	Z1.D, Z2.D, Z2.D

(330) 0x41775c UMSUBL	X14, W13, W24, X10

(330) 0x417760 UMADDL	X13, W13, W24, XZR

  (325) 0x417764 LD1D	{Z3.D}, P0/Z, [X9, X15,LSL #3]

  (325) 0x417768 LD1D	{Z4.D}, P0/Z, [X17, X15,LSL #3]

  (325) 0x41776c LD1D	{Z5.D}, P0/Z, [X16, X15,LSL #3]

  (325) 0x417770 FMLA	Z1.D, P0/M, Z5.D, Z3.D

  (325) 0x417774 LD1D	{Z6.D}, P0/Z, [X18, X15,LSL #3]

  (325) 0x417778 FMLA	Z2.D, P0/M, Z6.D, Z4.D

  (325) 0x41777c ADD	X15, X15, X24

  (325) 0x417780 CMP	X13, X15

  (325) 0x417784 B.NE	417764

(330) 0x417788 FADD	Z1.D, Z2.D, Z1.D

(330) 0x41778c FADDV	D1, P0, Z1.D

(330) 0x417790 CBZ	X14, 4177bc

  (332) 0x417794 SUB	X10, X10, X13

  (332) 0x417798 UBFM	X13, X13, #61, #60

  (332) 0x41779c ADD	X12, X13, X12,LSL #4

  (332) 0x4177a0 ADD	X9, X9, X13

  (332) 0x4177a4 ADD	X11, X11, X12

    (331) 0x4177a8 LDR	D2, [X9], #8

    (331) 0x4177ac LDR	D3, [X11], #8

    (331) 0x4177b0 FMADD	D1, D3, D2, D1

    (331) 0x4177b4 SUBS	X10, X10, #1

    (331) 0x4177b8 B.NE	4177a8

  (332) 0x4177bc STR	D1, [X8, #16]

  (332) 0x4177c0 LDR	X8, [X20, #432]

  (332) 0x4177c4 STR	D0, [X8, X21,LSL #3]

  (332) 0x4177c8 LDRSW	X8, [X20, #152]

  (332) 0x4177cc ADD	X21, X21, #1

  (332) 0x4177d0 CMP	X21, X8

  (332) 0x4177d4 B.LT	417430

0x4177d8 LDP	X20, X19, [SP, #64]

0x4177dc LDP	X22, X21, [SP, #48]

0x4177e0 LDP	X24, X23, [SP, #32]

0x4177e4 LDR	X25, [SP, #16]

0x4177e8 LDP	X29, X30, [SP], #80

0x4177ec RET

0x4177f0 HINT	#0

0x4177f4 HINT	#0

0x4177f8 HINT	#0

0x4177fc HINT	#0

Coverage (%)	Name	Source Location	Module
►100.00+	miniqmcreference::OneBodyJastr[...]	OneBodyJastrowRef.h:171	exec
○	miniqmcreference::OneBodyJastr[...]	OneBodyJastrowRef.h:126	exec
○	qmcplusplus::WaveFunction::eva[...]	WaveFunction.cpp:181	exec
○	.omp_outlined.	miniqmc.cpp:375	exec
○	__kmp_invoke_microtask		libomp.so

Path /

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

Source file and lines	OneBodyJastrowRef.h:111-198
Module	exec

The function is defined in:

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_numeric.h:140-141
/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:1046
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsMatrix.h:214
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/OneBodyJastrowRef.h:111-119,190-198
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,249

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 6.25 to 3.00 cycles (2.08x speedup).

Workaround

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

Vectorization

Your function is not vectorized. Only 35% of vector register length is used (average across all VPU instructions). By fully vectorizing your function, you can lower the cost of an iteration from 6.25 to 5.17 cycles (1.21x speedup).

Details

11% of VPU instructions are used in vector version (process two or more data elements in vector registers):

25% of VPU addition or subtraction instructions are used in vector version.
0% of VPU fused multiply-add instructions are used in vector version.
0% of VPU divide and square root instructions are used in vector version.
5% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

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

Workaround

Try another compiler or update/tune your current one
Make array accesses unit-stride:
- If your function 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)

No data for this section

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

The binary function is composed of 8 FP arithmetical operations:

8: addition or subtraction

The binary function is loading 104 bytes. The binary function is storing 72 bytes.

Arithmetic intensity

Arithmetic intensity is 0.05 FP operations per loaded or stored byte.

General properties

nb instructions	68
loop length	272
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

6.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.00	3.00	5.75	5.75	6.00	5.50	3.00	3.00	1.00	1.00	5.17	4.83	5.00	2.50	2.50
cycles	3.00	3.00	5.75	5.75	6.00	5.50	3.00	3.00	1.50	1.50	5.17	4.83	5.00	2.50	2.50

Cycles executing div or sqrt instructions

1.00-0.50

Cycles summary

Front-end	6.25
Overall L1	6.25

Vectorization ratios

INT

all	4%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	0%
other	6%

all	50%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

INT+FP

all	11%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	0%
div/sqrt	0%
other	5%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 6.25 cycles. At this rate:

17% of peak load performance is reached (16.64 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
24% of peak store performance is reached (11.52 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 4173c0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
STP X29, X30, [SP, #944]!	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STR X25, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
ADD X29, SP, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR W8, [X0, #152]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP W8, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.LT 4177d8 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x418>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X1, #2576]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDRSW X9, [X0, #144]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR X19, XZR, X1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X20, XZR, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W23, #24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W25, #40	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X22, [X8, X9,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR X21, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CNTW X24, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
B 417430 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x70>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X20, #248]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
SUB X10, X9, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVI D0, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
AND X10, X10, #7936	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X10, X10, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP X10, X24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.CS 4174a0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0xe0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR X10, XZR, X8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
PTRUE P0.D, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
FDUP Z6.D, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B 4174f0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x130>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0
HINT #0
HINT #0
UDIV X11, X10, X24	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-20	1-0.50
DUP Z1.D, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
ORR X12, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADDVL X14, X8, #1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
PTRUE P0.D, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
ORR Z0.D, Z1.D, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
MADD X13, X11, X24, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
SUB X11, X10, X13	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X10, X8, X13,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FADD Z0.D, Z1.D, Z0.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
FDUP Z6.D, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
FADDV D0, P0, Z0.D	1	0	0	0	0	0	0	0.75	0.75	0.25	0.25	0	0	0	0	0	9	2
CBZ X11, 417510 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ADD X8, X8, X9,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
LDP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDR X25, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDP X29, X30, [SP], #80	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0
HINT #0
HINT #0
HINT #0

Source file and lines	OneBodyJastrowRef.h:111-198
Module	exec

The function is defined in:

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_numeric.h:140-141
/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:1046
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsMatrix.h:214
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/OneBodyJastrowRef.h:111-119,190-198
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,249

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 6.25 to 3.00 cycles (2.08x speedup).

Workaround

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

Vectorization

Details

11% of VPU instructions are used in vector version (process two or more data elements in vector registers):

25% of VPU addition or subtraction instructions are used in vector version.
0% of VPU fused multiply-add instructions are used in vector version.
0% of VPU divide and square root instructions are used in vector version.
5% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

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

Workaround

Try another compiler or update/tune your current one
Make array accesses unit-stride:
- If your function 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)

No data for this section

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

The binary function is composed of 8 FP arithmetical operations:

8: addition or subtraction

The binary function is loading 104 bytes. The binary function is storing 72 bytes.

Arithmetic intensity

Arithmetic intensity is 0.05 FP operations per loaded or stored byte.

General properties

nb instructions	68
loop length	272
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

6.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.00	3.00	5.75	5.75	6.00	5.50	3.00	3.00	1.00	1.00	5.17	4.83	5.00	2.50	2.50
cycles	3.00	3.00	5.75	5.75	6.00	5.50	3.00	3.00	1.50	1.50	5.17	4.83	5.00	2.50	2.50

Cycles executing div or sqrt instructions

1.00-0.50

Cycles summary

Front-end	6.25
Overall L1	6.25

Vectorization ratios

INT

all	4%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	0%
other	6%

all	50%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

INT+FP

all	11%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	0%
div/sqrt	0%
other	5%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 6.25 cycles. At this rate:

17% of peak load performance is reached (16.64 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
24% of peak store performance is reached (11.52 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 4173c0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
STP X29, X30, [SP, #944]!	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STR X25, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
ADD X29, SP, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR W8, [X0, #152]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP W8, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.LT 4177d8 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x418>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X1, #2576]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDRSW X9, [X0, #144]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR X19, XZR, X1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X20, XZR, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W23, #24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W25, #40	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X22, [X8, X9,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR X21, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CNTW X24, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
B 417430 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x70>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X20, #248]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
SUB X10, X9, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVI D0, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
AND X10, X10, #7936	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X10, X10, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP X10, X24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.CS 4174a0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0xe0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR X10, XZR, X8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
PTRUE P0.D, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
FDUP Z6.D, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B 4174f0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x130>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0
HINT #0
HINT #0
UDIV X11, X10, X24	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-20	1-0.50
DUP Z1.D, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
ORR X12, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADDVL X14, X8, #1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
PTRUE P0.D, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
ORR Z0.D, Z1.D, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
MADD X13, X11, X24, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
SUB X11, X10, X13	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X10, X8, X13,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FADD Z0.D, Z1.D, Z0.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
FDUP Z6.D, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
FADDV D0, P0, Z0.D	1	0	0	0	0	0	0	0.75	0.75	0.25	0.25	0	0	0	0	0	9	2
CBZ X11, 417510 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9recomputeERNS1_11ParticleSetE+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ADD X8, X8, X9,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
HINT #0
LDP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDR X25, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDP X29, X30, [SP], #80	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0
HINT #0
HINT #0
HINT #0

Name	Coverage (%)	Time (s)
▼miniqmcreference::OneBodyJastrowRef >::recompute(qmcplusplus::ParticleSet&)–	0.01	0
○Loop 336 - stl_numeric.h:141-141 - exec	0.01	0
○Loop 337 - stl_numeric.h:140-141 - exec	0	0
▼Loop 330 - OneBodyJastrowRef.h:112-198 - exec–	0	0
▼Loop 332 - OneBodyJastrowRef.h:112-198 - exec–	0	0
○Loop 331 - OneBodyJastrowRef.h:196-197 - exec	0	0
○Loop 327 - OneBodyJastrowRef.h:190-191 - exec	0	0
○Loop 335 - OneBodyJastrowRef.h:190-191 - exec	0	0
○Loop 334 - OneBodyJastrowRef.h:196-197 - exec	0	0
○Loop 329 - OneBodyJastrowRef.h:196-197 - exec	0	0
○Loop 333 - OneBodyJastrowRef.h:196-197 - exec	0	0
○Loop 326 - OneBodyJastrowRef.h:112-198 - exec	0	0
○Loop 328 - OneBodyJastrowRef.h:196-197 - exec	0	0
○Loop 325 - OneBodyJastrowRef.h:196-197 - exec	0	0

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

Code clean check

Workaround

Vectorization

Details

Workaround

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code