OV - exec - Function qmcplusplus::DistanceTableAA<double, 3u, 39>::evaluate(qmcplusplus::ParticleSet&, int)

	Function: qmcplusplus::DistanceTableAA<double, 3u, 39>::evaluate(qmcplusplus::ParticleSet&, int)	Module: exec	Source: DistanceTableAA.h:74-78 [...]	Coverage: 8.43%

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsMatrix.h: 217 - 217

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

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

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds.h: 167 - 217

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

167:     const T x0 = pos[0];

168:     const T y0 = pos[1];

169:     const T z0 = pos[2];

[...]

185: #pragma omp simd aligned(temp_r, px, py, pz, dx, dy, dz: QMC_SIMD_ALIGNMENT)

186:     for (int iat = first; iat < last; ++iat)

187:     {

188:       const T flip    = iat < flip_ind ? one : minusone;

189:       const T displ_0 = (px[iat] - x0) * flip;

190:       const T displ_1 = (py[iat] - y0) * flip;

191:       const T displ_2 = (pz[iat] - z0) * flip;

192:

193:       const T ar_0 = -std::floor(displ_0 * g00 + displ_1 * g10 + displ_2 * g20);

194:       const T ar_1 = -std::floor(displ_0 * g01 + displ_1 * g11 + displ_2 * g21);

195:       const T ar_2 = -std::floor(displ_0 * g02 + displ_1 * g12 + displ_2 * g22);

196:

197:       const T delx = displ_0 + ar_0 * r00 + ar_1 * r10 + ar_2 * r20;

198:       const T dely = displ_1 + ar_0 * r01 + ar_1 * r11 + ar_2 * r21;

199:       const T delz = displ_2 + ar_0 * r02 + ar_1 * r12 + ar_2 * r22;

200:

201:       T rmin = delx * delx + dely * dely + delz * delz;

202:       int ic = 0;

203: #pragma unroll(7)

204:       for (int c = 1; c < 8; ++c)

205:       {

206:         const T x  = delx + cellx[c];

207:         const T y  = dely + celly[c];

208:         const T z  = delz + cellz[c];

209:         const T r2 = x * x + y * y + z * z;

210:         ic         = (r2 < rmin) ? c : ic;

211:         rmin       = (r2 < rmin) ? r2 : rmin;

212:       }

213:

214:       temp_r[iat] = std::sqrt(rmin);

215:       dx[iat]     = flip * (delx + cellx[ic]);

216:       dy[iat]     = flip * (dely + celly[ic]);

217:       dz[iat]     = flip * (delz + cellz[ic]);

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h: 223 - 248

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

223:     return X[i];

[...]

248:   inline pointer data() { return X; }

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/VectorSoAContainer.h: 241 - 243

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

241:   T* restrict data(size_t i) { return myData + i * nGhosts; }

242:   ///return the const pointer of the i-th components

243:   const T* restrict data(size_t i) const { return myData + i * nGhosts; }

/usr/include/c++/11/bits/stl_vector.h: 1046 - 1046

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

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

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/DistanceTableAA.h: 74 - 78

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

74:   inline void evaluate(ParticleSet& P, IndexType jat)

75:   {

76:     DTD_BConds<T, D, SC>::computeDistances(P.R[jat], P.RSoA, Distances[jat], Displacements[jat], 0, Ntargets, jat);

77:     Distances[jat][jat] = std::numeric_limits<T>::max(); // assign a big number

78:   }

0x46c300 STP	X29, X30, [SP, #944]!

0x46c304 ADD	X29, SP, #0

0x46c308 MOVZ	W4, #40

0x46c30c MOVZ	W3, #24

0x46c310 ADD	X6, X1, #1528

0x46c314 LDR	X5, [X0, #80]

0x46c318 SMADDL	X7, W2, W3, XZR

0x46c31c STP	X21, X22, [SP, #32]

0x46c320 SBFM	X21, X2, #0, #31

0x46c324 LDR	X9, [X1, #1512]

0x46c328 SMADDL	X11, W2, W4, X5

0x46c32c LDR	X22, [X0, #24]

0x46c330 LDR	X12, [X11, #8]

0x46c334 LDR	X8, [X1, #1536]

0x46c338 ADD	X1, X9, X7

0x46c33c MADD	X22, X21, X22, XZR

0x46c340 LDR	X13, [X11, #24]

0x46c344 UBFM	X16, X12, #61, #60

0x46c348 LDR	X11, [X6, #24]

0x46c34c UBFM	X15, X8, #61, #60

0x46c350 LDR	W10, [X0, #200]

0x46c354 ADD	X14, X13, X16

0x46c358 LDR	D6, [X9, X7]

0x46c35c ADD	X9, X14, X16

0x46c360 ADD	X12, X11, X15

0x46c364 LDP	D7, D16, [X1, #8]

0x46c368 ADD	X8, X12, X15

0x46c36c STR	X23, [SP, #48]

0x46c370 LDR	X23, [X0, #64]

0x46c374 ADD	X15, X23, X22,LSL #3

0x46c378 CMP	W10, #0

0x46c37c B.LE	46c610

0x46c380 FMOV	D18, #-1.0000000

0x46c384 FMOV	D17, #1.0000000

0x46c388 MOVZ	X4, #0

0x46c38c ADD	X1, X0, #240

0x46c390 ADD	X7, X0, #384

0x46c394 STP	X19, X20, [SP, #16]

0x46c398 ADD	X6, X0, #448

0x46c39c ADD	X5, X0, #512

0x46c3a0 MOVZ	W20, #2

0x46c3a4 MOVZ	W19, #3

0x46c3a8 MOVZ	W30, #4

0x46c3ac MOVZ	W18, #5

0x46c3b0 STR	D10, [SP, #56]

0x46c3b4 MOVZ	W17, #6

0x46c3b8 MOVZ	W16, #7

0x46c3bc STP	D8, D9, [SP, #64]

(1205) 0x46c3c0 LDR	D20, [X12, X4,LSL #3]

(1205) 0x46c3c4 CMP	W2, W4

(1205) 0x46c3c8 LDR	D5, [X11, X4,LSL #3]

(1205) 0x46c3cc FCSEL	D19, D18, D17, #13

(1205) 0x46c3d0 LDR	D0, [X8, X4,LSL #3]

(1205) 0x46c3d4 FSUB	D1, D20, S7

(1205) 0x46c3d8 LDP	D28, D8, [X1, #32]

(1205) 0x46c3dc FSUB	D2, D5, S6

(1205) 0x46c3e0 LDP	D9, D4, [X1]

(1205) 0x46c3e4 FSUB	D3, D0, S16

(1205) 0x46c3e8 LDP	D27, D21, [X1, #16]

(1205) 0x46c3ec FMUL	D24, D1, D19

(1205) 0x46c3f0 LDP	D10, D22, [X1, #96]

(1205) 0x46c3f4 FMUL	D25, D2, D19

(1205) 0x46c3f8 LDR	D29, [X1, #80]

(1205) 0x46c3fc FMUL	D26, D3, D19

(1205) 0x46c400 LDR	D23, [X1, #64]

(1205) 0x46c404 FMUL	D30, D24, D28

(1205) 0x46c408 FMUL	D31, D24, D4

(1205) 0x46c40c LDR	D28, [X1, #56]

(1205) 0x46c410 FMADD	D20, D25, D21, D30

(1205) 0x46c414 FMADD	D5, D25, D9, D31

(1205) 0x46c418 LDR	D21, [X1, #128]

(1205) 0x46c41c FMUL	D31, D24, D28

(1205) 0x46c420 FMADD	D0, D26, D8, D20

(1205) 0x46c424 FRINTM	D2, D0

(1205) 0x46c428 FNMUL	D3, D2, D22

(1205) 0x46c42c FMADD	D9, D26, D27, D5

(1205) 0x46c430 FRINTM	D4, D9

(1205) 0x46c434 FNMUL	D22, D2, D21

(1205) 0x46c438 LDR	D20, [X1, #48]

(1205) 0x46c43c FNMUL	D1, D2, D29

(1205) 0x46c440 LDR	D29, [X1, #72]

(1205) 0x46c444 FMSUB	D27, D4, D10, D3

(1205) 0x46c448 LDP	D10, D8, [X1, #112]

(1205) 0x46c44c FMADD	D3, D25, D20, D31

(1205) 0x46c450 LDR	D9, [X1, #136]

(1205) 0x46c454 FMSUB	D30, D4, D29, D1

(1205) 0x46c458 LDR	D2, [X1, #88]

(1205) 0x46c45c FMSUB	D0, D4, D8, D22

(1205) 0x46c460 FMADD	D23, D26, D23, D3

(1205) 0x46c464 FRINTM	D5, D23

(1205) 0x46c468 FADD	D24, D24, D27

(1205) 0x46c46c FADD	D25, D25, D30

(1205) 0x46c470 LDP	D21, D30, [X6, #8]

(1205) 0x46c474 FMSUB	D20, D5, D10, D24

(1205) 0x46c478 FADD	D26, D26, D0

(1205) 0x46c47c LDP	D29, D10, [X5, #8]

(1205) 0x46c480 LDP	D24, D8, [X6, #24]

(1205) 0x46c484 FMSUB	D1, D5, D2, D25

(1205) 0x46c488 FMSUB	D0, D5, D9, D26

(1205) 0x46c48c FADD	D25, D20, D21

(1205) 0x46c490 FMUL	D4, D20, D20

(1205) 0x46c494 FADD	D3, D20, D30

(1205) 0x46c498 FADD	D2, D20, D24

(1205) 0x46c49c FADD	D23, D20, D8

(1205) 0x46c4a0 LDP	D8, D30, [X7, #24]

(1205) 0x46c4a4 FMUL	D5, D25, D25

(1205) 0x46c4a8 FMADD	D27, D1, D1, D4

(1205) 0x46c4ac LDR	D4, [X5, #24]

(1205) 0x46c4b0 FADD	D22, D0, D10

(1205) 0x46c4b4 FADD	D31, D0, D29

(1205) 0x46c4b8 LDP	D10, D21, [X7, #8]

(1205) 0x46c4bc FMUL	D26, D3, D3

(1205) 0x46c4c0 FMUL	D9, D2, D2

(1205) 0x46c4c4 FADD	D3, D1, D8

(1205) 0x46c4c8 FADD	D2, D1, D30

(1205) 0x46c4cc FMADD	D28, D0, D0, D27

(1205) 0x46c4d0 FMUL	D27, D23, D23

(1205) 0x46c4d4 FADD	D29, D0, D4

(1205) 0x46c4d8 FADD	D25, D1, D10

(1205) 0x46c4dc FADD	D24, D1, D21

(1205) 0x46c4e0 FMADD	D23, D25, D25, D5

(1205) 0x46c4e4 FMADD	D5, D24, D24, D26

(1205) 0x46c4e8 FMADD	D26, D3, D3, D9

(1205) 0x46c4ec FMADD	D9, D2, D2, D27

(1205) 0x46c4f0 FMADD	D27, D31, D31, D23

(1205) 0x46c4f4 FMINNM	D4, D28, D27

(1205) 0x46c4f8 FMADD	D22, D22, D22, D5

(1205) 0x46c4fc FMINNM	D31, D22, D4

(1205) 0x46c500 FMADD	D10, D29, D29, D26

(1205) 0x46c504 FMINNM	D29, D10, D31

(1205) 0x46c508 FCMPE	D28, D27

(1205) 0x46c50c LDP	D28, D21, [X5, #32]

(1205) 0x46c510 CSINC	W3, WZR, WZR, #13

(1205) 0x46c514 FCMPE	D22, D4

(1205) 0x46c518 CSEL	W3, W3, W20, #10

(1205) 0x46c51c LDR	D8, [X5, #48]

(1205) 0x46c520 FCMPE	D10, D31

(1205) 0x46c524 CSEL	W3, W3, W19, #10

(1205) 0x46c528 LDP	D24, D2, [X6, #40]

(1205) 0x46c52c FADD	D30, D0, D28

(1205) 0x46c530 FADD	D5, D0, D21

(1205) 0x46c534 LDR	D26, [X6, #56]

(1205) 0x46c538 FADD	D3, D0, D8

(1205) 0x46c53c LDP	D21, D8, [X7, #40]

(1205) 0x46c540 FADD	D4, D20, D2

(1205) 0x46c544 FMADD	D25, D30, D30, D9

(1205) 0x46c548 FADD	D9, D20, D24

(1205) 0x46c54c FMINNM	D23, D25, D29

(1205) 0x46c550 FADD	D27, D20, D26

(1205) 0x46c554 FADD	D30, D1, D21

(1205) 0x46c558 FMUL	D10, D4, D4

(1205) 0x46c55c FMUL	D22, D9, D9

(1205) 0x46c560 FMUL	D31, D27, D27

(1205) 0x46c564 FCMPE	D25, D29

(1205) 0x46c568 CSEL	W3, W3, W30, #10

(1205) 0x46c56c FADD	D25, D1, D8

(1205) 0x46c570 FMADD	D24, D30, D30, D22

(1205) 0x46c574 LDR	D29, [X5, #56]

(1205) 0x46c578 FMADD	D5, D5, D5, D24

(1205) 0x46c57c FMINNM	D26, D5, D23

(1205) 0x46c580 FMADD	D2, D25, D25, D10

(1205) 0x46c584 FADD	D28, D0, D29

(1205) 0x46c588 FMADD	D3, D3, D3, D2

(1205) 0x46c58c FMINNM	D9, D3, D26

(1205) 0x46c590 FCMPE	D5, D23

(1205) 0x46c594 LDR	D23, [X7, #56]

(1205) 0x46c598 CSEL	W3, W3, W18, #10

(1205) 0x46c59c FCMPE	D3, D26

(1205) 0x46c5a0 CSEL	W3, W3, W17, #10

(1205) 0x46c5a4 FADD	D4, D1, D23

(1205) 0x46c5a8 FMADD	D27, D4, D4, D31

(1205) 0x46c5ac FMADD	D22, D28, D28, D27

(1205) 0x46c5b0 FMINNM	D10, D22, D9

(1205) 0x46c5b4 FSQRT	D31, D10

(1205) 0x46c5b8 FCMPE	D22, D9

(1205) 0x46c5bc CSEL	W3, W3, W16, #10

(1205) 0x46c5c0 ADD	X3, X0, W3,SXTW #3

(1205) 0x46c5c4 LDR	D29, [X3, #384]

(1205) 0x46c5c8 STR	D31, [X15, X4,LSL #3]

(1205) 0x46c5cc FADD	D1, D1, D29

(1205) 0x46c5d0 FMUL	D28, D1, D19

(1205) 0x46c5d4 STR	D28, [X13, X4,LSL #3]

(1205) 0x46c5d8 LDR	D21, [X3, #448]

(1205) 0x46c5dc FADD	D20, D20, D21

(1205) 0x46c5e0 FMUL	D8, D20, D19

(1205) 0x46c5e4 STR	D8, [X14, X4,LSL #3]

(1205) 0x46c5e8 LDR	D30, [X3, #512]

(1205) 0x46c5ec FADD	D0, D0, D30

(1205) 0x46c5f0 FMUL	D19, D0, D19

(1205) 0x46c5f4 STR	D19, [X9, X4,LSL #3]

(1205) 0x46c5f8 ADD	X4, X4, #1

(1205) 0x46c5fc CMP	W10, W4

(1205) 0x46c600 B.GT	46c3c0

0x46c604 LDR	D10, [SP, #56]

0x46c608 LDP	X19, X20, [SP, #16]

0x46c60c LDP	D8, D9, [SP, #64]

0x46c610 ADD	X2, X21, X22

0x46c614 MOVN	X0, #32784

0x46c618 LDP	X21, X22, [SP, #32]

0x46c61c STR	X0, [X23, X2,LSL #3]

0x46c620 LDR	X23, [SP, #48]

0x46c624 LDP	X29, X30, [SP], #80

0x46c628 RET

0x46c62c HINT	#0

0x46c630 HINT	#0

0x46c634 HINT	#0

0x46c638 HINT	#0

0x46c63c HINT	#0

Coverage (%)	Name	Source Location	Module
►42.86+	qmcplusplus::ParticleSet::setA[...]	stl_vector.h:919	exec
○	main._omp_fn.1	refwrap.h:337	exec
○	GOMP_parallel		libomp.so
○	main	miniqmc.cpp:404	exec
○	__libc_start_main		libc-2.31.so
○	_start	miniqmc.cpp:546	exec
►35.71+	qmcplusplus::ParticleSet::setA[...]	stl_vector.h:919	exec
○	main._omp_fn.1	refwrap.h:337	exec
○	GOMP_parallel		libomp.so
○	main	miniqmc.cpp:402	exec
○	__libc_start_main		libc-2.31.so
○	_start	miniqmc.cpp:546	exec
►14.29+	qmcplusplus::ParticleSet::setA[...]	stl_vector.h:919	exec
○	main._omp_fn.1	refwrap.h:337	exec
○	GOMP_parallel		libomp.so
○	main	miniqmc.cpp:404	exec
○	__libc_start_main		libc-2.31.so
○	_start	miniqmc.cpp:546	exec
►7.14+	qmcplusplus::ParticleSet::setA[...]	stl_vector.h:919	exec
○	main._omp_fn.1	refwrap.h:337	exec
○	GOMP_parallel		libomp.so
○	main	miniqmc.cpp:402	exec
○	__libc_start_main		libc-2.31.so
○	_start	miniqmc.cpp:546	exec

Path /

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

Source file and lines	DistanceTableAA.h:74-78
Module	exec

The function is defined in:

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsMatrix.h:217
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds.h:167-169,185-217
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,248
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/VectorSoAContainer.h:241-243
/usr/include/c++/11/bits/stl_vector.h:1046
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/DistanceTableAA.h:74-78

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 8.17 to 2.67 cycles (3.06x 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 23% 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 8.17 to 1.52 cycles (5.37x speedup).

Details

13% of VPU instructions are used in vector version (process two or more data elements in vector registers):

50% of VPU loads are used in vector version.
50% of VPU stores are used in vector version.
0% of VPU addition or subtraction instructions are used in vector version.
0% 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 does not contain any FP arithmetical operations. The binary function is loading 172 bytes. The binary function is storing 88 bytes.

General properties

nb instructions	63
loop length	252
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

7.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	1.00	1.00	7.75	7.75	7.75	7.75	1.00	1.00	1.00	1.00	8.17	7.83	8.00	2.50	2.50
cycles	1.00	1.00	7.75	7.75	7.75	7.75	1.00	1.00	1.00	1.00	8.17	7.83	8.00	2.50	2.50

Cycles executing div or sqrt instructions

Cycles summary

Front-end	7.25
Overall L1	8.17

Vectorization ratios

all	13%
load	50%
store	50%
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 8.17 cycles. At this rate:

21% of peak load performance is reached (21.06 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
22% of peak store performance is reached (10.78 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 46c300

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
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
MOVZ W4, #40	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W3, #24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X6, X1, #1528	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X5, [X0, #80]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
SMADDL X7, W2, W3, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
STP X21, X22, [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
SBFM X21, X2, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X9, [X1, #1512]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
SMADDL X11, W2, W4, X5	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
LDR X22, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X12, [X11, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X8, [X1, #1536]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X1, X9, X7	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MADD X22, X21, X22, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
LDR X13, [X11, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
UBFM X16, X12, #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
LDR X11, [X6, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
UBFM X15, 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
LDR W10, [X0, #200]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X14, X13, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR D6, [X9, X7]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X9, X14, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X12, X11, X15	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDP D7, D16, [X1, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X8, X12, X15	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR X23, [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
LDR X23, [X0, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X15, X23, X22,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
CMP W10, #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 46c610 <_ZN11qmcplusplus15DistanceTableAAIdLj3ELi39EE8evaluateERNS_11ParticleSetEi+0x310>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
FMOV D18, #-1.0000000	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
FMOV D17, #1.0000000	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
MOVZ X4, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X1, X0, #240	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X7, X0, #384	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STP X19, X20, [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
ADD X6, X0, #448	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X5, X0, #512	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W20, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W19, #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W30, #4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W18, #5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR D10, [SP, #56]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
MOVZ W17, #6	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W16, #7	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STP D8, D9, [SP, #64]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D10, [SP, #56]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP D8, D9, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X2, X21, X22	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVN X0, #32784	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
STR X0, [X23, X2,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
LDR X23, [SP, #48]	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
HINT #0

Source file and lines	DistanceTableAA.h:74-78
Module	exec

The function is defined in:

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsMatrix.h:217
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds.h:167-169,185-217
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,248
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/VectorSoAContainer.h:241-243
/usr/include/c++/11/bits/stl_vector.h:1046
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/DistanceTableAA.h:74-78

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 8.17 to 2.67 cycles (3.06x speedup).

Workaround

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

Vectorization

Details

13% of VPU instructions are used in vector version (process two or more data elements in vector registers):

50% of VPU loads are used in vector version.
50% of VPU stores are used in vector version.
0% of VPU addition or subtraction instructions are used in vector version.
0% 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 does not contain any FP arithmetical operations. The binary function is loading 172 bytes. The binary function is storing 88 bytes.

General properties

nb instructions	63
loop length	252
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

7.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	1.00	1.00	7.75	7.75	7.75	7.75	1.00	1.00	1.00	1.00	8.17	7.83	8.00	2.50	2.50
cycles	1.00	1.00	7.75	7.75	7.75	7.75	1.00	1.00	1.00	1.00	8.17	7.83	8.00	2.50	2.50

Cycles executing div or sqrt instructions

Cycles summary

Front-end	7.25
Overall L1	8.17

Vectorization ratios

all	13%
load	50%
store	50%
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 8.17 cycles. At this rate:

21% of peak load performance is reached (21.06 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
22% of peak store performance is reached (10.78 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 46c300

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
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
MOVZ W4, #40	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W3, #24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X6, X1, #1528	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X5, [X0, #80]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
SMADDL X7, W2, W3, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
STP X21, X22, [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
SBFM X21, X2, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X9, [X1, #1512]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
SMADDL X11, W2, W4, X5	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
LDR X22, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X12, [X11, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X8, [X1, #1536]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X1, X9, X7	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MADD X22, X21, X22, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
LDR X13, [X11, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
UBFM X16, X12, #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
LDR X11, [X6, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
UBFM X15, 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
LDR W10, [X0, #200]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X14, X13, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR D6, [X9, X7]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X9, X14, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X12, X11, X15	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDP D7, D16, [X1, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X8, X12, X15	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR X23, [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
LDR X23, [X0, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X15, X23, X22,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
CMP W10, #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 46c610 <_ZN11qmcplusplus15DistanceTableAAIdLj3ELi39EE8evaluateERNS_11ParticleSetEi+0x310>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
FMOV D18, #-1.0000000	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
FMOV D17, #1.0000000	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
MOVZ X4, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X1, X0, #240	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X7, X0, #384	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STP X19, X20, [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
ADD X6, X0, #448	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X5, X0, #512	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W20, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W19, #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W30, #4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W18, #5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR D10, [SP, #56]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
MOVZ W17, #6	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVZ W16, #7	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STP D8, D9, [SP, #64]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D10, [SP, #56]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP D8, D9, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X2, X21, X22	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVN X0, #32784	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDP X21, X22, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
STR X0, [X23, X2,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
LDR X23, [SP, #48]	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
HINT #0

Name	Coverage (%)	Time (s)
▼qmcplusplus::DistanceTableAA::evaluate(qmcplusplus::ParticleSet&, int)–	8.43	0.07
○Loop 1205 - ParticleBConds.h:188-217 - exec	8.43	0.07

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

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

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code