OV - exec - Function qmcplusplus::DistanceTableBA<double, 3u, 39>::move(qmcplusplus::ParticleSet const&, qmcplusplus::TinyVector<double, 3u> const&, int)

	Function: qmcplusplus::DistanceTableBA<double, 3u, 39>::move(qmcplusplus::ParticleSet const&, qmcplu ...	Module: exec	Source: DistanceTableBA.h:117-120 [...]	Coverage: 2.41%

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/TinyVector.h: 177 - 177

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

177:   inline const Type_t& operator[](unsigned int i) const { return X[i]; }

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds.h: 168 - 217

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

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/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; }

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/DistanceTableBA.h: 117 - 120

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

117:   inline void move(const ParticleSet& P, const PosType& rnew, IndexType iat)

118:   {

119:     DTD_BConds<T, D, SC>::computeDistances(rnew, Origin->RSoA, Temp_r.data(), Temp_dr, 0, Nsources);

120:   }

0x470fc0 LDR	X1, [X0, #8]

0x470fc4 LDR	X8, [X0, #136]

0x470fc8 LDR	X12, [X0, #152]

0x470fcc LDR	X7, [X1, #1536]

0x470fd0 UBFM	X3, X8, #61, #60

0x470fd4 LDR	X10, [X1, #1552]

0x470fd8 ADD	X13, X12, X3

0x470fdc LDR	W9, [X0, #196]

0x470fe0 ADD	X8, X13, X3

0x470fe4 UBFM	X4, X7, #61, #60

0x470fe8 LDP	D6, D7, [X2]

0x470fec ADD	X11, X10, X4

0x470ff0 ADD	X7, X11, X4

0x470ff4 LDR	D16, [X2, #16]

0x470ff8 LDR	X14, [X0, #104]

0x470ffc CMP	W9, #0

0x471000 B.LE	471278

0x471004 STP	X29, X30, [SP, #992]!

0x471008 ADD	X29, SP, #0

0x47100c MOVZ	X3, #0

0x471010 ADD	X1, X0, #240

0x471014 ADD	X6, X0, #384

0x471018 ADD	X5, X0, #448

0x47101c ADD	X4, X0, #512

0x471020 MOVZ	W30, #3

0x471024 STR	X19, [SP, #16]

0x471028 MOVZ	W18, #4

0x47102c MOVZ	W17, #5

0x471030 MOVZ	W16, #6

0x471034 MOVZ	W15, #7

0x471038 MOVZ	W19, #2

0x47103c HINT	#0

(1249) 0x471040 LDR	D1, [X11, X3,LSL #3]

(1249) 0x471044 LDR	D2, [X10, X3,LSL #3]

(1249) 0x471048 LDP	D17, D5, [X1]

(1249) 0x47104c FSUB	D21, D7, S1

(1249) 0x471050 LDR	D0, [X7, X3,LSL #3]

(1249) 0x471054 FSUB	D22, D6, S2

(1249) 0x471058 LDR	D4, [X1, #32]

(1249) 0x47105c LDR	D3, [X1, #56]

(1249) 0x471060 FMUL	D24, D21, D5

(1249) 0x471064 FSUB	D23, D16, S0

(1249) 0x471068 LDP	D19, D20, [X1, #16]

(1249) 0x47106c LDR	D18, [X1, #96]

(1249) 0x471070 FMUL	D25, D21, D4

(1249) 0x471074 LDP	D1, D2, [X1, #40]

(1249) 0x471078 FMUL	D29, D21, D3

(1249) 0x47107c FMADD	D26, D22, D17, D24

(1249) 0x471080 LDR	D0, [X1, #104]

(1249) 0x471084 FMADD	D27, D22, D20, D25

(1249) 0x471088 LDR	D3, [X1, #64]

(1249) 0x47108c FMADD	D28, D23, D19, D26

(1249) 0x471090 FRINTM	D30, D28

(1249) 0x471094 LDR	D20, [X1, #80]

(1249) 0x471098 FMSUB	D31, D30, D18, D21

(1249) 0x47109c FMADD	D4, D22, D2, D29

(1249) 0x4710a0 LDR	D21, [X1, #72]

(1249) 0x4710a4 FMADD	D17, D23, D1, D27

(1249) 0x4710a8 FRINTM	D5, D17

(1249) 0x4710ac LDP	D19, D18, [X1, #112]

(1249) 0x4710b0 FMSUB	D24, D5, D0, D31

(1249) 0x4710b4 FMADD	D25, D23, D3, D4

(1249) 0x4710b8 FRINTM	D26, D25

(1249) 0x4710bc LDR	D29, [X1, #88]

(1249) 0x4710c0 FMSUB	D22, D30, D21, D22

(1249) 0x4710c4 FMSUB	D23, D30, D18, D23

(1249) 0x4710c8 LDP	D30, D1, [X1, #128]

(1249) 0x4710cc FMSUB	D31, D26, D19, D24

(1249) 0x4710d0 LDP	D19, D24, [X5, #24]

(1249) 0x4710d4 FMSUB	D27, D5, D20, D22

(1249) 0x4710d8 LDP	D3, D20, [X5, #8]

(1249) 0x4710dc FMSUB	D21, D5, D30, D23

(1249) 0x4710e0 LDP	D5, D22, [X4, #8]

(1249) 0x4710e4 FMSUB	D2, D26, D29, D27

(1249) 0x4710e8 FMSUB	D0, D26, D1, D21

(1249) 0x4710ec LDR	D21, [X4, #24]

(1249) 0x4710f0 FMUL	D28, D31, D31

(1249) 0x4710f4 FADD	D18, D31, D3

(1249) 0x4710f8 FADD	D23, D31, D24

(1249) 0x4710fc FADD	D26, D31, D20

(1249) 0x471100 FADD	D27, D31, D19

(1249) 0x471104 LDR	D19, [X6, #24]

(1249) 0x471108 FMADD	D17, D2, D2, D28

(1249) 0x47110c FMUL	D28, D18, D18

(1249) 0x471110 FMUL	D1, D23, D23

(1249) 0x471114 FMUL	D29, D26, D26

(1249) 0x471118 FADD	D18, D2, D19

(1249) 0x47111c FADD	D4, D0, D5

(1249) 0x471120 FMUL	D30, D27, D27

(1249) 0x471124 FMADD	D25, D0, D0, D17

(1249) 0x471128 LDP	D17, D5, [X6, #8]

(1249) 0x47112c FADD	D22, D0, D22

(1249) 0x471130 FADD	D27, D0, D21

(1249) 0x471134 FADD	D3, D2, D17

(1249) 0x471138 FADD	D20, D2, D5

(1249) 0x47113c FMADD	D24, D3, D3, D28

(1249) 0x471140 FMADD	D26, D20, D20, D29

(1249) 0x471144 FMADD	D28, D18, D18, D30

(1249) 0x471148 FMADD	D4, D4, D4, D24

(1249) 0x47114c FMINNM	D23, D25, D4

(1249) 0x471150 FMADD	D29, D22, D22, D26

(1249) 0x471154 FMADD	D21, D27, D27, D28

(1249) 0x471158 FMINNM	D30, D29, D23

(1249) 0x47115c FMINNM	D20, D30, D21

(1249) 0x471160 FCMPE	D25, D4

(1249) 0x471164 LDR	D25, [X6, #32]

(1249) 0x471168 CSINC	W2, WZR, WZR, #13

(1249) 0x47116c FCMPE	D29, D23

(1249) 0x471170 CSEL	W2, W2, W19, #10

(1249) 0x471174 LDR	D5, [X4, #32]

(1249) 0x471178 FCMPE	D30, D21

(1249) 0x47117c CSEL	W2, W2, W30, #9

(1249) 0x471180 LDP	D29, D23, [X5, #40]

(1249) 0x471184 FADD	D17, D2, D25

(1249) 0x471188 LDR	D19, [X4, #48]

(1249) 0x47118c FADD	D3, D0, D5

(1249) 0x471190 LDR	D30, [X5, #56]

(1249) 0x471194 LDR	D22, [X4, #40]

(1249) 0x471198 FADD	D21, D31, D29

(1249) 0x47119c FMADD	D1, D17, D17, D1

(1249) 0x4711a0 FADD	D27, D0, D19

(1249) 0x4711a4 LDR	D18, [X4, #56]

(1249) 0x4711a8 FADD	D17, D31, D23

(1249) 0x4711ac LDR	D19, [X6, #56]

(1249) 0x4711b0 FADD	D5, D31, D30

(1249) 0x4711b4 FADD	D4, D0, D22

(1249) 0x4711b8 FMUL	D25, D21, D21

(1249) 0x4711bc FMADD	D24, D3, D3, D1

(1249) 0x4711c0 FMINNM	D26, D20, D24

(1249) 0x4711c4 FADD	D28, D0, D18

(1249) 0x4711c8 FMUL	D3, D17, D17

(1249) 0x4711cc FADD	D18, D2, D19

(1249) 0x4711d0 FMUL	D1, D5, D5

(1249) 0x4711d4 FMADD	D17, D18, D18, D1

(1249) 0x4711d8 FCMPE	D20, D24

(1249) 0x4711dc LDP	D24, D20, [X6, #40]

(1249) 0x4711e0 CSEL	W2, W2, W18, #9

(1249) 0x4711e4 FMADD	D28, D28, D28, D17

(1249) 0x4711e8 FADD	D22, D2, D24

(1249) 0x4711ec FADD	D29, D2, D20

(1249) 0x4711f0 FMADD	D23, D22, D22, D25

(1249) 0x4711f4 FMADD	D30, D29, D29, D3

(1249) 0x4711f8 FMADD	D5, D4, D4, D23

(1249) 0x4711fc FMADD	D27, D27, D27, D30

(1249) 0x471200 FMINNM	D21, D26, D5

(1249) 0x471204 FMINNM	D4, D21, D27

(1249) 0x471208 FMINNM	D25, D28, D4

(1249) 0x47120c FSQRT	D3, D25

(1249) 0x471210 FCMPE	D26, D5

(1249) 0x471214 CSEL	W2, W2, W17, #9

(1249) 0x471218 FCMPE	D21, D27

(1249) 0x47121c CSEL	W2, W2, W16, #9

(1249) 0x471220 FCMPE	D28, D4

(1249) 0x471224 CSEL	W2, W2, W15, #10

(1249) 0x471228 ADD	X2, X0, W2,SXTW #3

(1249) 0x47122c LDR	D26, [X2, #384]

(1249) 0x471230 FADD	D2, D2, D26

(1249) 0x471234 STR	D3, [X14, X3,LSL #3]

(1249) 0x471238 FNEG	D1, D2

(1249) 0x47123c STR	D1, [X12, X3,LSL #3]

(1249) 0x471240 LDR	D24, [X2, #448]

(1249) 0x471244 FADD	D31, D31, D24

(1249) 0x471248 FNEG	D20, D31

(1249) 0x47124c STR	D20, [X13, X3,LSL #3]

(1249) 0x471250 LDR	D19, [X2, #512]

(1249) 0x471254 FADD	D0, D0, D19

(1249) 0x471258 FNEG	D22, D0

(1249) 0x47125c STR	D22, [X8, X3,LSL #3]

(1249) 0x471260 ADD	X3, X3, #1

(1249) 0x471264 CMP	W9, W3

(1249) 0x471268 B.GT	471040

0x47126c LDR	X19, [SP, #16]

0x471270 LDP	X29, X30, [SP], #32

0x471274 RET

0x471278 RET

0x47127c HINT	#0

Coverage (%)	Name	Source Location	Module
►25.00+	qmcplusplus::ParticleSet::make[...]	ParticleSet.cpp:290	exec
○	main._omp_fn.1	stl_vector.h:1043	exec
○	GOMP_parallel		libomp.so
○	main	miniqmc.cpp:402	exec
○	__libc_start_main		libc-2.31.so
○	_start	miniqmc.cpp:546	exec
►25.00+	qmcplusplus::ParticleSet::make[...]	ParticleSet.cpp:290	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
►25.00+	qmcplusplus::ParticleSet::make[...]	ParticleSet.cpp:290	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
►25.00+	qmcplusplus::ParticleSet::make[...]	ParticleSet.cpp:290	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

Path /

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

Source file and lines	DistanceTableBA.h:117-120
Module	exec

The function is defined in:

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/TinyVector.h:177
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds.h:168-169,185-217
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/VectorSoAContainer.h:241-243
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/DistanceTableBA.h:117-120

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 5.00 to 0.67 cycles (7.50x 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 21% 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 5.00 to 0.84 cycles (5.93x speedup).

Details

5% 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.
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 100 bytes. The binary function is storing 24 bytes.

General properties

nb instructions	37
loop length	148
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

4.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	1.50	1.50	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	4.33	4.33	4.33	1.00	1.00
cycles	1.50	1.50	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	4.33	4.33	4.33	1.00	1.00

Cycles executing div or sqrt instructions

Cycles summary

Front-end	4.38
Overall L1	5.00

Vectorization ratios

all	5%
load	50%
store	NA (no store vectorizable/vectorized instructions)
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 5.00 cycles. At this rate:

20% of peak load performance is reached (20.00 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
10% of peak store performance is reached (4.80 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 470fc0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
LDR X1, [X0, #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, [X0, #136]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X12, [X0, #152]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X7, [X1, #1536]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
UBFM X3, 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 X10, [X1, #1552]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X13, X12, X3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR W9, [X0, #196]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X8, X13, X3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
UBFM X4, X7, #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
LDP D6, D7, [X2]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X11, X10, X4	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, X11, X4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR D16, [X2, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
LDR X14, [X0, #104]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP W9, #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 471278 <_ZN11qmcplusplus15DistanceTableBAIdLj3ELi39EE4moveERKNS_11ParticleSetERKNS_10TinyVectorIdLj3EEEi+0x2b8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
STP X29, X30, [SP, #992]!	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 X3, #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 X6, 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
ADD X5, 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 X4, 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 W30, #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR X19, [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
MOVZ W18, #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 W17, #5	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, #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 W15, #7	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, #2	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
LDR X19, [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], #32	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
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0

Source file and lines	DistanceTableBA.h:117-120
Module	exec

The function is defined in:

/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/TinyVector.h:177
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds.h:168-169,185-217
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/VectorSoAContainer.h:241-243
/home/hbollore/qaas/qaas-runs/169-816-7419/intel/miniqmc/build/miniqmc/src/Particle/DistanceTableBA.h:117-120

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 5.00 to 0.67 cycles (7.50x speedup).

Workaround

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

Vectorization

Details

5% 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.
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 100 bytes. The binary function is storing 24 bytes.

General properties

nb instructions	37
loop length	148
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

4.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	1.50	1.50	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	4.33	4.33	4.33	1.00	1.00
cycles	1.50	1.50	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	4.33	4.33	4.33	1.00	1.00

Cycles executing div or sqrt instructions

Cycles summary

Front-end	4.38
Overall L1	5.00

Vectorization ratios

all	5%
load	50%
store	NA (no store vectorizable/vectorized instructions)
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 5.00 cycles. At this rate:

20% of peak load performance is reached (20.00 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
10% of peak store performance is reached (4.80 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 470fc0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
LDR X1, [X0, #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, [X0, #136]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X12, [X0, #152]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X7, [X1, #1536]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
UBFM X3, 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 X10, [X1, #1552]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X13, X12, X3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR W9, [X0, #196]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X8, X13, X3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
UBFM X4, X7, #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
LDP D6, D7, [X2]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
ADD X11, X10, X4	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, X11, X4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR D16, [X2, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
LDR X14, [X0, #104]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
CMP W9, #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 471278 <_ZN11qmcplusplus15DistanceTableBAIdLj3ELi39EE4moveERKNS_11ParticleSetERKNS_10TinyVectorIdLj3EEEi+0x2b8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
STP X29, X30, [SP, #992]!	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 X3, #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 X6, 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
ADD X5, 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 X4, 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 W30, #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR X19, [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
MOVZ W18, #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 W17, #5	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, #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 W15, #7	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, #2	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
LDR X19, [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], #32	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
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0

Name	Coverage (%)	Time (s)
▼qmcplusplus::DistanceTableBA::move(qmcplusplus::ParticleSet const&, qmcplusplus::TinyVector const&, int)–	2.41	0.02
○Loop 1249 - ParticleBConds.h:189-217 - exec	2.41	0.02

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