OV - exec - Function miniqmcreference::OneBodyJastrowRef<qmcplusplus::BsplineFunctor<double> >::computeU3(qmcplusplus::ParticleSet&, int, double const*)

	Function: miniqmcreference::OneBodyJastrowRef<qmcplusplus::BsplineFunctor<double> >::computeU3(qmcpl ...	Module: exec	Source: OneBodyJastrowRef.h:210-245 [...]	Coverage: 0.03%

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h: 1046 - 1169

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

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

[...]

1169:       { return _M_data_ptr(this->_M_impl._M_start); }

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/OneBodyJastrowRef.h: 210 - 245

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

210:   {

211:     if (NumGroups > 0)

212:     { // ions are grouped

213:       constexpr valT czero(0);

214:       std::fill_n(U.data(), Nions, czero);

215:       std::fill_n(dU.data(), Nions, czero);

216:       std::fill_n(d2U.data(), Nions, czero);

217:

218:       for (int jg = 0; jg < NumGroups; ++jg)

219:       {

220:         if (F[jg] == nullptr)

221:           continue;

222:         F[jg]->evaluateVGL(-1,

223:                            Ions.first(jg),

[...]

235:       for (int c = 0; c < Nions; ++c)

236:       {

237:         int gid = Ions.GroupID[c];

238:         if (F[gid] != nullptr)

239:         {

240:           U[c] = F[gid]->evaluate(dist[c], dU[c], d2U[c]);

241:           dU[c] /= dist[c];

242:         }

243:       }

244:     }

245:   }

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h: 229 - 229

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

229:     return X[i];

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_algobase.h: 923 - 1115

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

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

924: 	*__first = __tmp;

[...]

1115:       if (__n <= 0)

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h: 302 - 305

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

302:   inline int first(int igroup) const { return SubPtcl[igroup]; }

303:

304:   /// return the last index of a group i

305:   inline int last(int igroup) const { return SubPtcl[igroup + 1]; }

0x4164c0 SUB	SP, SP, #80

0x4164c4 STP	X29, X30, [SP, #16]

0x4164c8 STR	X23, [SP, #32]

0x4164cc STP	X22, X21, [SP, #48]

0x4164d0 STP	X20, X19, [SP, #64]

0x4164d4 ADD	X29, SP, #16

0x4164d8 LDR	W22, [X0, #156]

0x4164dc ORR	X19, XZR, X3

0x4164e0 ORR	X20, XZR, X0

0x4164e4 CMP	W22, #0

0x4164e8 B.LE	416524

0x4164ec LDR	W8, [X20, #148]

0x4164f0 CMP	W8, #0

0x4164f4 B.LE	416650

0x4164f8 LDR	X0, [X20, #248]

0x4164fc RDVL	X23, #1

0x416500 DUP	Z0.B, #0

0x416504 UBFM	X21, X8, #61, #60

0x416508 CMP	X23, #17

0x41650c B.CS	4165a8

0x416510 ORR	W1, WZR, WZR

0x416514 ORR	X2, XZR, X21

0x416518 BL	403de0

0x41651c DUP	Z0.B, #0

0x416520 B	4165d0

0x416524 LDR	W8, [X20, #148]

0x416528 CMP	W8, #1

0x41652c B.LT	4166bc

0x416530 ORR	X21, XZR, XZR

0x416534 ORR	X22, XZR, XZR

0x416538 ORR	X23, XZR, XZR

0x41653c B	416554

(315) 0x416540 ADD	X23, X23, #1

(315) 0x416544 ADD	X22, X22, #8

(315) 0x416548 ADD	X21, X21, #4

(315) 0x41654c CMP	X23, W8,SXTW

(315) 0x416550 B.GE	4166bc

(315) 0x416554 LDR	X9, [X20, #160]

(315) 0x416558 LDR	X10, [X20, #448]

(315) 0x41655c LDR	X9, [X9, #1472]

(315) 0x416560 LDRSW	X9, [X9, X21]

(315) 0x416564 LDR	X0, [X10, X9,LSL #3]

(315) 0x416568 CBZ	X0, 416540

0x41656c LDR	X8, [X20, #272]

0x416570 LDR	D0, [X19, X22]

0x416574 ADD	X1, X8, X22

0x416578 LDR	X8, [X20, #296]

0x41657c ADD	X2, X8, X22

0x416580 BL	416d40

0x416584 LDR	X8, [X20, #248]

0x416588 STR	D0, [X8, X22]

0x41658c LDR	X8, [X20, #272]

0x416590 LDR	D0, [X19, X22]

0x416594 LDR	D1, [X8, X22]

0x416598 FDIV	D0, D1, D0

0x41659c STR	D0, [X8, X22]

0x4165a0 LDR	W8, [X20, #148]

0x4165a4 B	416540

0x4165a8 WHILELO	P0.B, XZR, X21

0x4165ac ORR	X8, XZR, XZR

0x4165b0 HINT	#0

0x4165b4 HINT	#0

0x4165b8 HINT	#0

0x4165bc HINT	#0

(316) 0x4165c0 ST1B	{Z0.B}, P0, [X0, X8]

(316) 0x4165c4 INCP	X8, P0.B

(316) 0x4165c8 WHILELO	P0.B, X8, X21

(316) 0x4165cc B.MI	4165c0

0x4165d0 LDR	X0, [X20, #272]

0x4165d4 CMP	X23, #17

0x4165d8 B.CS	4165f0

0x4165dc ORR	W1, WZR, WZR

0x4165e0 ORR	X2, XZR, X21

0x4165e4 BL	403de0

0x4165e8 DUP	Z0.B, #0

0x4165ec B	416610

0x4165f0 WHILELO	P0.B, XZR, X21

0x4165f4 ORR	X8, XZR, XZR

0x4165f8 HINT	#0

0x4165fc HINT	#0

(317) 0x416600 ST1B	{Z0.B}, P0, [X0, X8]

(317) 0x416604 INCP	X8, P0.B

(317) 0x416608 WHILELO	P0.B, X8, X21

(317) 0x41660c B.MI	416600

0x416610 LDR	X0, [X20, #296]

0x416614 CMP	X23, #17

0x416618 B.CS	41662c

0x41661c ORR	W1, WZR, WZR

0x416620 ORR	X2, XZR, X21

0x416624 BL	403de0

0x416628 B	416650

0x41662c WHILELO	P0.B, XZR, X21

0x416630 ORR	X8, XZR, XZR

0x416634 HINT	#0

0x416638 HINT	#0

0x41663c HINT	#0

(318) 0x416640 ST1B	{Z0.B}, P0, [X0, X8]

(318) 0x416644 INCP	X8, P0.B

(318) 0x416648 WHILELO	P0.B, X8, X21

(318) 0x41664c B.MI	416640

0x416650 ORR	X21, XZR, XZR

0x416654 B	4166a4

0x416658 HINT	#0

0x41665c HINT	#0

(314) 0x416660 LDR	X8, [X20, #160]

(314) 0x416664 LDR	X5, [X20, #248]

(314) 0x416668 LDR	X6, [X20, #272]

(314) 0x41666c ORR	X4, XZR, X19

(314) 0x416670 MOVN	W1, #0

(314) 0x416674 LDR	X7, [X20, #296]

(314) 0x416678 LDR	X9, [X20, #344]

(314) 0x41667c LDR	X8, [X8, #2712]

(314) 0x416680 LDR	W2, [X8, X21,LSL #2]

(314) 0x416684 ADD	X21, X21, #1

(314) 0x416688 LDR	W3, [X8, X21,LSL #2]

(314) 0x41668c LDR	X8, [X20, #320]

(314) 0x416690 STP	X8, X9, [SP]

(314) 0x416694 BL	4166e0

(314) 0x416698 LDR	W22, [X20, #156]

(314) 0x41669c CMP	X21, W22,SXTW

(314) 0x4166a0 B.GE	4166bc

(314) 0x4166a4 LDR	X8, [X20, #448]

(314) 0x4166a8 LDR	X0, [X8, X21,LSL #3]

(314) 0x4166ac CBNZ	X0, 416660

0x4166b0 ADD	X21, X21, #1

0x4166b4 CMP	X21, W22,SXTW

0x4166b8 B.LT	4166a4

0x4166bc LDP	X20, X19, [SP, #64]

0x4166c0 LDP	X22, X21, [SP, #48]

0x4166c4 LDR	X23, [SP, #32]

0x4166c8 LDP	X29, X30, [SP, #16]

0x4166cc ADD	SP, SP, #80

0x4166d0 RET

0x4166d4 HINT	#0

0x4166d8 HINT	#0

0x4166dc HINT	#0

Coverage (%)	Name	Source Location	Module
►66.67+	miniqmcreference::OneBodyJastr[...]	OneBodyJastrowRef.h:263	exec
○	qmcplusplus::WaveFunction::rat[...]	NewTimer.h:249	exec
○	.omp_outlined..62	miniqmc.cpp:437	exec
○	__kmp_invoke_microtask		libomp.so
►33.33+	miniqmcreference::OneBodyJastr[...]	OneBodyJastrowRef.h:116	exec
○	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:210-245
Module	exec

The function is defined in:

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:1046,1169
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/OneBodyJastrowRef.h:210-223,235-245
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:229
/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_algobase.h:923-924,1115
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:302-305

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 9.88 to 2.50 cycles (3.95x 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 19% of vector register length is used (average across all VPU instructions). By vectorizing your function, you can lower the cost of an iteration from 9.88 to 1.20 cycles (8.26x speedup).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a 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)

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 9.88 to 2.00 cycles (4.94x speedup).

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

The binary function is composed of 1 FP arithmetical operations:

1: divide

The binary function is loading 152 bytes. The binary function is storing 72 bytes.

Arithmetic intensity

Arithmetic intensity is 0.00 FP operations per loaded or stored byte.

General properties

nb instructions	93
loop length	372
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

9.88 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	9.00	9.00	9.25	9.25	9.25	9.25	2.50	2.50	1.00	0.00	8.00	8.00	8.00	2.00	2.00
cycles	9.00	9.00	9.25	9.25	9.25	9.25	2.50	2.50	1.00	0.00	8.00	8.00	8.00	2.00	2.00

Cycles executing div or sqrt instructions

1.00-0.50

Cycles summary

Front-end	9.88
Overall L1	9.88

Vectorization ratios

all	0%
load	0%
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	0%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 9.88 cycles. At this rate:

16% of peak load performance is reached (15.39 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
15% of peak store performance is reached (7.29 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 4164c0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
SUB SP, SP, #80	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STP X29, X30, [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
STR 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, #16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR W22, [X0, #156]	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, X3	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
CMP W22, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.LE 416524 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x64>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR W8, [X20, #148]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP W8, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.LE 416650 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x190>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X0, [X20, #248]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
RDVL X23, #1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
DUP Z0.B, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
UBFM X21, X8, #61, #60	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP X23, #17	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 4165a8 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0xe8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR W1, WZR, WZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X2, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 403de0 <@plt_start@+0x4b0>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50
DUP Z0.B, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B 4165d0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x110>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR W8, [X20, #148]	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 4166bc <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x1fc>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
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
ORR X22, 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
ORR X23, 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
B 416554 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x94>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X20, #272]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR D0, [X19, X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X1, X8, X22	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X8, [X20, #296]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X2, X8, X22	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 416d40 <_ZN11qmcplusplus14BsplineFunctorIdE8evaluateEdRdS2_>	1	0.50	0.50	0.50	0.50	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
STR D0, [X8, X22]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR X8, [X20, #272]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR D0, [X19, X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
LDR D1, [X8, X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
FDIV D0, D1, D0	1	0	0	0	0	0	0	0.50	0	0.50	0	0	0	0	0	0	7-15	1-0.50
STR D0, [X8, X22]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR W8, [X20, #148]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
B 416540 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x80>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
WHILELO P0.B, XZR, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2
ORR X8, 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
HINT #0
HINT #0
HINT #0
HINT #0
LDR X0, [X20, #272]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP X23, #17	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 4165f0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x130>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR W1, WZR, WZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X2, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 403de0 <@plt_start@+0x4b0>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50
DUP Z0.B, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B 416610 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
WHILELO P0.B, XZR, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2
ORR X8, 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
HINT #0
HINT #0
LDR X0, [X20, #296]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP X23, #17	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 41662c <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x16c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR W1, WZR, WZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X2, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 403de0 <@plt_start@+0x4b0>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50
B 416650 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x190>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
WHILELO P0.B, XZR, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2
ORR X8, 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
HINT #0
HINT #0
HINT #0
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
B 4166a4 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x1e4>	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
ADD X21, X21, #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 X21, W22,SXTW	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50
B.LT 4166a4 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x1e4>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
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
LDR X23, [SP, #32]	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, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
ADD SP, SP, #80	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
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

Source file and lines	OneBodyJastrowRef.h:210-245
Module	exec

The function is defined in:

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:1046,1169
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/OneBodyJastrowRef.h:210-223,235-245
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:229
/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_algobase.h:923-924,1115
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:302-305

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 9.88 to 2.50 cycles (3.95x speedup).

Workaround

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

Vectorization

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a 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)

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 9.88 to 2.00 cycles (4.94x speedup).

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

The binary function is composed of 1 FP arithmetical operations:

1: divide

The binary function is loading 152 bytes. The binary function is storing 72 bytes.

Arithmetic intensity

Arithmetic intensity is 0.00 FP operations per loaded or stored byte.

General properties

nb instructions	93
loop length	372
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

9.88 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	9.00	9.00	9.25	9.25	9.25	9.25	2.50	2.50	1.00	0.00	8.00	8.00	8.00	2.00	2.00
cycles	9.00	9.00	9.25	9.25	9.25	9.25	2.50	2.50	1.00	0.00	8.00	8.00	8.00	2.00	2.00

Cycles executing div or sqrt instructions

1.00-0.50

Cycles summary

Front-end	9.88
Overall L1	9.88

Vectorization ratios

all	0%
load	0%
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	0%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 9.88 cycles. At this rate:

16% of peak load performance is reached (15.39 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
15% of peak store performance is reached (7.29 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 4164c0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
SUB SP, SP, #80	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STP X29, X30, [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
STR 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, #16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR W22, [X0, #156]	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, X3	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
CMP W22, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.LE 416524 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x64>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR W8, [X20, #148]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP W8, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.LE 416650 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x190>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X0, [X20, #248]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
RDVL X23, #1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
DUP Z0.B, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
UBFM X21, X8, #61, #60	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP X23, #17	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 4165a8 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0xe8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR W1, WZR, WZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X2, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 403de0 <@plt_start@+0x4b0>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50
DUP Z0.B, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B 4165d0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x110>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR W8, [X20, #148]	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 4166bc <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x1fc>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
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
ORR X22, 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
ORR X23, 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
B 416554 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x94>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X20, #272]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR D0, [X19, X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X1, X8, X22	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X8, [X20, #296]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X2, X8, X22	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 416d40 <_ZN11qmcplusplus14BsplineFunctorIdE8evaluateEdRdS2_>	1	0.50	0.50	0.50	0.50	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
STR D0, [X8, X22]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR X8, [X20, #272]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR D0, [X19, X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
LDR D1, [X8, X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
FDIV D0, D1, D0	1	0	0	0	0	0	0	0.50	0	0.50	0	0	0	0	0	0	7-15	1-0.50
STR D0, [X8, X22]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR W8, [X20, #148]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
B 416540 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x80>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
WHILELO P0.B, XZR, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2
ORR X8, 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
HINT #0
HINT #0
HINT #0
HINT #0
LDR X0, [X20, #272]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP X23, #17	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 4165f0 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x130>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR W1, WZR, WZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X2, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 403de0 <@plt_start@+0x4b0>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50
DUP Z0.B, #0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B 416610 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
WHILELO P0.B, XZR, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2
ORR X8, 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
HINT #0
HINT #0
LDR X0, [X20, #296]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP X23, #17	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 41662c <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x16c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
ORR W1, WZR, WZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X2, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
BL 403de0 <@plt_start@+0x4b0>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50
B 416650 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x190>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
WHILELO P0.B, XZR, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2
ORR X8, 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
HINT #0
HINT #0
HINT #0
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
B 4166a4 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x1e4>	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
ADD X21, X21, #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 X21, W22,SXTW	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50
B.LT 4166a4 <_ZN16miniqmcreference17OneBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE9computeU3ERNS1_11ParticleSetEiPKd+0x1e4>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
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
LDR X23, [SP, #32]	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, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
ADD SP, SP, #80	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
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

Name	Coverage (%)	Time (s)
▼miniqmcreference::OneBodyJastrowRef >::computeU3(qmcplusplus::ParticleSet&, int, double const*)–	0.03	0.01
○Loop 317 - stl_algobase.h:924-924 - exec	0.03	0.01
○Loop 318 - stl_algobase.h:924-924 - exec	0	0
○Loop 315 - OneBodyJastrowRef.h:235-238 - exec	0	0
○Loop 316 - stl_algobase.h:924-924 - exec	0	0
○Loop 314 - OneBodyJastrowRef.h:218-223 - exec	0	0

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Expensive FP math instructions/calls

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

Expensive FP math instructions/calls

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