OV - exec - Loop 3456

	Loop Id: 3456	Module: exec	Source: ParticleBConds3DSoa.h:237-255	Coverage: 0.6%

0x4f7590 LD1D	{Z21.D}, P1/Z, [X12, X0,LSL #3]    [4]

0x4f7594 LD1D	{Z22.D}, P1/Z, [X11, X0,LSL #3]    [6]

0x4f7598 LD1D	{Z4.D}, P1/Z, [X6, X0,LSL #3]    [2]

0x4f759c LD1RD	{Z20.D}, P0/Z, [X13, #13]    [3]

0x4f75a0 LD1RD	{Z3.D}, P0/Z, [X13, #12]    [3]

0x4f75a4 LD1RD	{Z6.D}, P0/Z, [X13, #10]    [3]

0x4f75a8 LD1RD	{Z2.D}, P0/Z, [X13, #9]    [3]

0x4f75ac LD1RD	{Z7.D}, P0/Z, [X13, #16]    [3]

0x4f75b0 LD1RD	{Z23.D}, P0/Z, [X13, #15]    [3]

0x4f75b4 FSUB	Z27.D, Z21.D, Z30.D

0x4f75b8 FSUB	Z28.D, Z22.D, Z31.D

0x4f75bc FSUB	Z26.D, Z4.D, Z29.D

0x4f75c0 LD1RD	{Z5.D}, P0/Z, [X13, #14]    [3]

0x4f75c4 LD1RD	{Z24.D}, P0/Z, [X13, #11]    [3]

0x4f75c8 LD1RD	{Z17.D}, P0/Z, [X13, #17]    [3]

0x4f75cc FMUL	Z1.D, Z20.D, Z27.D

0x4f75d0 FMUL	Z0.D, Z6.D, Z27.D

0x4f75d4 FMLA	Z1.D, P0/M, Z3.D, Z28.D

0x4f75d8 FMLA	Z0.D, P0/M, Z2.D, Z28.D

0x4f75dc FMAD	Z5.D, P0/M, Z26.D, Z1.D

0x4f75e0 FMAD	Z24.D, P0/M, Z26.D, Z0.D

0x4f75e4 FMUL	Z25.D, Z7.D, Z27.D

0x4f75e8 FMLA	Z25.D, P0/M, Z23.D, Z28.D

0x4f75ec FMAD	Z17.D, P0/M, Z26.D, Z25.D

0x4f75f0 LD1RD	{Z16.D}, P0/Z, [X13, #1]    [3]

0x4f75f4 LD1RD	{Z1.D}, P0/Z, [X13]    [3]

0x4f75f8 LD1RD	{Z3.D}, P0/Z, [X13, #2]    [3]

0x4f75fc MOVPRFX	Z18, Z5

0x4f7600 FRINTA	Z18.D, P0/M, Z5.D

0x4f7604 MOVPRFX	Z21, Z24

0x4f7608 FRINTA	Z21.D, P0/M, Z24.D

0x4f760c FSUB	Z19.D, Z5.D, Z18.D

0x4f7610 FSUB	Z22.D, Z24.D, Z21.D

0x4f7614 MOVPRFX	Z20, Z17

0x4f7618 FRINTA	Z20.D, P0/M, Z17.D

0x4f761c FMUL	Z27.D, Z16.D, Z19.D

0x4f7620 FSUB	Z28.D, Z17.D, Z20.D

0x4f7624 FMLA	Z27.D, P0/M, Z1.D, Z22.D

0x4f7628 FMAD	Z3.D, P0/M, Z28.D, Z27.D

0x4f762c ST1D	{Z3.D}, P1, [X20, X0,LSL #3]    [1]

0x4f7630 LD1RD	{Z4.D}, P0/Z, [X13, #3]    [3]

0x4f7634 LD1RD	{Z26.D}, P0/Z, [X13, #4]    [3]

0x4f7638 LD1RD	{Z5.D}, P0/Z, [X13, #5]    [3]

0x4f763c FMUL	Z6.D, Z26.D, Z19.D

0x4f7640 FMLA	Z6.D, P0/M, Z4.D, Z22.D

0x4f7644 FMAD	Z5.D, P0/M, Z28.D, Z6.D

0x4f7648 ST1D	{Z5.D}, P1, [X2, X0,LSL #3]    [5]

0x4f764c LD1RD	{Z0.D}, P0/Z, [X13, #7]    [3]

0x4f7650 LD1RD	{Z2.D}, P0/Z, [X13, #6]    [3]

0x4f7654 LD1RD	{Z24.D}, P0/Z, [X13, #8]    [3]

0x4f7658 FMUL	Z7.D, Z0.D, Z19.D

0x4f765c FMLA	Z7.D, P0/M, Z2.D, Z22.D

0x4f7660 FMAD	Z24.D, P0/M, Z28.D, Z7.D

0x4f7664 ST1D	{Z24.D}, P1, [X28, X0,LSL #3]    [7]

0x4f7668 LD1D	{Z25.D}, P1/Z, [X2, X0,LSL #3]    [5]

0x4f766c LD1D	{Z23.D}, P1/Z, [X20, X0,LSL #3]    [1]

0x4f7670 FMUL	Z17.D, Z25.D, Z25.D

0x4f7674 FMLA	Z17.D, P0/M, Z23.D, Z23.D

0x4f7678 FMLA	Z17.D, P0/M, Z24.D, Z24.D

0x4f767c FSQRT	Z17.D, P0/M, Z17.D

0x4f7680 ST1D	{Z17.D}, P1, [X8, X0,LSL #3]    [8]

0x4f7684 ADD	X0, X0, X9

0x4f7688 WHILELO	P1.D, W0, W10

0x4f768c B.NE	4f7590

/home/hbollore/qaas-runs/171-275-7410/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds3DSoa.h: 237 - 255

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

237:       T displ_0 = px[iat] - x0;

238:       T displ_1 = py[iat] - y0;

239:       T displ_2 = pz[iat] - z0;

240:

241:       T ar_0 = displ_0 * g00 + displ_1 * g10 + displ_2 * g20;

242:       T ar_1 = displ_0 * g01 + displ_1 * g11 + displ_2 * g21;

243:       T ar_2 = displ_0 * g02 + displ_1 * g12 + displ_2 * g22;

244:

245:       //put them in the box

246:       ar_0 -= round(ar_0);

247:       ar_1 -= round(ar_1);

248:       ar_2 -= round(ar_2);

249:

250:       //unit2cart

251:       dx[iat] = ar_0 * r00 + ar_1 * r10 + ar_2 * r20;

252:       dy[iat] = ar_0 * r01 + ar_1 * r11 + ar_2 * r21;

253:       dz[iat] = ar_0 * r02 + ar_1 * r12 + ar_2 * r22;

254:

255:       temp_r[iat] = std::sqrt(dx[iat] * dx[iat] + dy[iat] * dy[iat] + dz[iat] * dz[iat]);

Coverage (%)	Name	Source Location	Module

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	NA
Bottlenecks	P6, P7,
Function	qmcplusplus::SoaDistanceTableAAOMPTarget::move(qmcplusplus::ParticleSet const&, qmcplusplus::TinyVector const&, int, bool)
Source	ParticleBConds3DSoa.h:237-255
Source loop unroll info	multi-versionned
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	17.00
CQA cycles if no scalar integer	17.00
CQA cycles if FP arith vectorized	17.00
CQA cycles if fully vectorized	17.00
Front-end cycles	8.00
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.25
P3 cycles	1.00
P4 cycles	0.25
P5 cycles	17.00
P6 cycles	17.00
P7 cycles	13.50
P8 cycles	13.50
P9 cycles	13.50
P10 cycles	13.50
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	6.99 - 14.08
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	64.00
Nb uops	64.00
Nb loads	NA
Nb stores	4.00
Nb stack references	0.00
FLOP/cycle	9.88
Nb FLOP add-sub	24.00
Nb FLOP mul	28.00
Nb FLOP fma	56.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	4.00
Nb FLOP rsqrt	0.00
Bytes/cycle	25.41
Bytes prefetched	0.00
Bytes loaded	304.00
Bytes stored	128.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	65.57
Vectorization ratio load	21.74
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	100.00
Vectorization ratio div_sqrt	100.00
Vectorization ratio other	50.00
Vector-efficiency ratio all	77.87
Vector-efficiency ratio load	41.30
Vector-efficiency ratio store	100.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	100.00
Vector-efficiency ratio div_sqrt	100.00
Vector-efficiency ratio other	100.00

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	NA
Bottlenecks	P6, P7,
Function	qmcplusplus::SoaDistanceTableAAOMPTarget::move(qmcplusplus::ParticleSet const&, qmcplusplus::TinyVector const&, int, bool)
Source	ParticleBConds3DSoa.h:237-255
Source loop unroll info	multi-versionned
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	17.00
CQA cycles if no scalar integer	17.00
CQA cycles if FP arith vectorized	17.00
CQA cycles if fully vectorized	17.00
Front-end cycles	8.00
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.25
P3 cycles	1.00
P4 cycles	0.25
P5 cycles	17.00
P6 cycles	17.00
P7 cycles	13.50
P8 cycles	13.50
P9 cycles	13.50
P10 cycles	13.50
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	6.99 - 14.08
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	64.00
Nb uops	64.00
Nb loads	NA
Nb stores	4.00
Nb stack references	0.00
FLOP/cycle	9.88
Nb FLOP add-sub	24.00
Nb FLOP mul	28.00
Nb FLOP fma	56.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	4.00
Nb FLOP rsqrt	0.00
Bytes/cycle	25.41
Bytes prefetched	0.00
Bytes loaded	304.00
Bytes stored	128.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	65.57
Vectorization ratio load	21.74
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	100.00
Vectorization ratio div_sqrt	100.00
Vectorization ratio other	50.00
Vector-efficiency ratio all	77.87
Vector-efficiency ratio load	41.30
Vector-efficiency ratio store	100.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	100.00
Vector-efficiency ratio div_sqrt	100.00
Vector-efficiency ratio other	100.00

Path /

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

Function	qmcplusplus::SoaDistanceTableAAOMPTarget::move(qmcplusplus::ParticleSet const&, qmcplusplus::TinyVector const&, int, bool)
Source file and lines	ParticleBConds3DSoa.h:237-255
Module	exec

The loop is defined in /home/hbollore/qaas-runs/171-275-7410/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds3DSoa.h:237-255.

The related source loop is multi-versionned.

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 77% of vector register length is used (average across all VPU instructions).

Details

65% of VPU instructions are used in vector version (process two or more data elements in vector registers):

21% of VPU loads are used in vector version.
50% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

FMA

Detected 56 FMA (fused multiply-add) operations. Presence of both ADD/SUB and MUL operations.

Workaround

Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible. For instance a + b*c is a valid FMA (MUL then ADD). However (a+b)* c cannot be translated into an FMA (ADD then MUL).

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

The binary loop is composed of 168 FP arithmetical operations:

80: addition or subtraction (56 inside FMA instructions)
84: multiply (56 inside FMA instructions)
4: square root

The binary loop is loading 304 bytes. The binary loop is storing 128 bytes.

Arithmetic intensity

Arithmetic intensity is 0.39 FP operations per loaded or stored byte.

General properties

nb instructions	64
loop length	256
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

8.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	0.50	0.25	1.00	0.25	17.00	17.00	2.00	2.00	13.50	13.50	0.00	0.00	0.00
cycles	0.50	0.50	0.50	0.25	1.00	0.25	17.00	17.00	13.50	13.50	13.50	13.50	0.00	0.00	0.00

Execution ports to units layout:

P0:
P1:
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P7 (128 bits): VPU, ALU, FP store data
P8 (128 bits): VPU, ALU, DIV/SQRT
P9 (128 bits): ALU, VPU
P10 (256 bits): store address, load
P11 (256 bits): store address, load
P12 (256 bits): load
P13 (64 bits): store data
P14 (64 bits): store data

Cycles executing div or sqrt instructions	6.99-14.08
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	8.00
Data deps.	1.00
Overall L1	17.00

Vectorization ratios

INT

all	30%
load	21%
store	100%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	0%

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

INT+FP

all	65%
load	21%
store	100%
mul	100%
add-sub	100%
fma	100%
div/sqrt	100%
other	50%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 17.00 cycles. At this rate:

18% of peak load performance is reached (17.88 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
15% of peak store performance is reached (7.53 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 4f7590

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput
LD1D {Z21.D}, P1/Z, [X12, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z22.D}, P1/Z, [X11, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z4.D}, P1/Z, [X6, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z20.D}, P0/Z, [X13, #13]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z3.D}, P0/Z, [X13, #12]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z6.D}, P0/Z, [X13, #10]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z2.D}, P0/Z, [X13, #9]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z7.D}, P0/Z, [X13, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z23.D}, P0/Z, [X13, #15]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FSUB Z27.D, Z21.D, Z30.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FSUB Z28.D, Z22.D, Z31.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FSUB Z26.D, Z4.D, Z29.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
LD1RD {Z5.D}, P0/Z, [X13, #14]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z24.D}, P0/Z, [X13, #11]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z17.D}, P0/Z, [X13, #17]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z1.D, Z20.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMUL Z0.D, Z6.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z1.D, P0/M, Z3.D, Z28.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMLA Z0.D, P0/M, Z2.D, Z28.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z5.D, P0/M, Z26.D, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z24.D, P0/M, Z26.D, Z0.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMUL Z25.D, Z7.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z25.D, P0/M, Z23.D, Z28.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z17.D, P0/M, Z26.D, Z25.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
LD1RD {Z16.D}, P0/Z, [X13, #1]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z1.D}, P0/Z, [X13]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z3.D}, P0/Z, [X13, #2]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
MOVPRFX Z18, Z5	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FRINTA Z18.D, P0/M, Z5.D	1	0	0	0	0	0	0	1	0	0	0	0	0	3	1
MOVPRFX Z21, Z24	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FRINTA Z21.D, P0/M, Z24.D	1	0	0	0	0	0	0	1	0	0	0	0	0	3	1
FSUB Z19.D, Z5.D, Z18.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FSUB Z22.D, Z24.D, Z21.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
MOVPRFX Z20, Z17	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FRINTA Z20.D, P0/M, Z17.D	1	0	0	0	0	0	0	1	0	0	0	0	0	3	1
FMUL Z27.D, Z16.D, Z19.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FSUB Z28.D, Z17.D, Z20.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FMLA Z27.D, P0/M, Z1.D, Z22.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z3.D, P0/M, Z28.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
ST1D {Z3.D}, P1, [X20, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
LD1RD {Z4.D}, P0/Z, [X13, #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z26.D}, P0/Z, [X13, #4]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z5.D}, P0/Z, [X13, #5]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z6.D, Z26.D, Z19.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z6.D, P0/M, Z4.D, Z22.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z5.D, P0/M, Z28.D, Z6.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
ST1D {Z5.D}, P1, [X2, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
LD1RD {Z0.D}, P0/Z, [X13, #7]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z2.D}, P0/Z, [X13, #6]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z24.D}, P0/Z, [X13, #8]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z7.D, Z0.D, Z19.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z7.D, P0/M, Z2.D, Z22.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z24.D, P0/M, Z28.D, Z7.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
ST1D {Z24.D}, P1, [X28, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
LD1D {Z25.D}, P1/Z, [X2, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z23.D}, P1/Z, [X20, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z17.D, Z25.D, Z25.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z17.D, P0/M, Z23.D, Z23.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMLA Z17.D, P0/M, Z24.D, Z24.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FSQRT Z17.D, P0/M, Z17.D	1	0	0	0	0	0	0	1	0	0	0	0	0	7-16	6.99-14.08
ST1D {Z17.D}, P1, [X8, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
ADD X0, X0, X9	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25
WHILELO P1.D, W0, W10	1	0	0	0	0	1	0	0	0	0	0	0	0	3	2
B.NE 4f7590 <_ZN11qmcplusplus27SoaDistanceTableAAOMPTargetIdLj3ELi40EE4moveERKNS_11ParticleSetERKNS_10TinyVectorIdLj3EEEib+0x2b0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1	0.50

Function	qmcplusplus::SoaDistanceTableAAOMPTarget::move(qmcplusplus::ParticleSet const&, qmcplusplus::TinyVector const&, int, bool)
Source file and lines	ParticleBConds3DSoa.h:237-255
Module	exec

The loop is defined in /home/hbollore/qaas-runs/171-275-7410/intel/miniqmc/build/miniqmc/src/Particle/Lattice/ParticleBConds3DSoa.h:237-255.

The related source loop is multi-versionned.

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 77% of vector register length is used (average across all VPU instructions).

Details

65% of VPU instructions are used in vector version (process two or more data elements in vector registers):

21% of VPU loads are used in vector version.
50% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

FMA

Detected 56 FMA (fused multiply-add) operations. Presence of both ADD/SUB and MUL operations.

Workaround

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

The binary loop is composed of 168 FP arithmetical operations:

80: addition or subtraction (56 inside FMA instructions)
84: multiply (56 inside FMA instructions)
4: square root

The binary loop is loading 304 bytes. The binary loop is storing 128 bytes.

Arithmetic intensity

Arithmetic intensity is 0.39 FP operations per loaded or stored byte.

General properties

nb instructions	64
loop length	256
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

8.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	0.50	0.25	1.00	0.25	17.00	17.00	2.00	2.00	13.50	13.50	0.00	0.00	0.00
cycles	0.50	0.50	0.50	0.25	1.00	0.25	17.00	17.00	13.50	13.50	13.50	13.50	0.00	0.00	0.00

Execution ports to units layout:

P0:
P1:
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P7 (128 bits): VPU, ALU, FP store data
P8 (128 bits): VPU, ALU, DIV/SQRT
P9 (128 bits): ALU, VPU
P10 (256 bits): store address, load
P11 (256 bits): store address, load
P12 (256 bits): load
P13 (64 bits): store data
P14 (64 bits): store data

Cycles executing div or sqrt instructions	6.99-14.08
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	8.00
Data deps.	1.00
Overall L1	17.00

Vectorization ratios

INT

all	30%
load	21%
store	100%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	0%

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

INT+FP

all	65%
load	21%
store	100%
mul	100%
add-sub	100%
fma	100%
div/sqrt	100%
other	50%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 17.00 cycles. At this rate:

18% of peak load performance is reached (17.88 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
15% of peak store performance is reached (7.53 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 4f7590

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput
LD1D {Z21.D}, P1/Z, [X12, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z22.D}, P1/Z, [X11, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z4.D}, P1/Z, [X6, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z20.D}, P0/Z, [X13, #13]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z3.D}, P0/Z, [X13, #12]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z6.D}, P0/Z, [X13, #10]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z2.D}, P0/Z, [X13, #9]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z7.D}, P0/Z, [X13, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z23.D}, P0/Z, [X13, #15]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FSUB Z27.D, Z21.D, Z30.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FSUB Z28.D, Z22.D, Z31.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FSUB Z26.D, Z4.D, Z29.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
LD1RD {Z5.D}, P0/Z, [X13, #14]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z24.D}, P0/Z, [X13, #11]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z17.D}, P0/Z, [X13, #17]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z1.D, Z20.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMUL Z0.D, Z6.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z1.D, P0/M, Z3.D, Z28.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMLA Z0.D, P0/M, Z2.D, Z28.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z5.D, P0/M, Z26.D, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z24.D, P0/M, Z26.D, Z0.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMUL Z25.D, Z7.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z25.D, P0/M, Z23.D, Z28.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z17.D, P0/M, Z26.D, Z25.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
LD1RD {Z16.D}, P0/Z, [X13, #1]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z1.D}, P0/Z, [X13]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z3.D}, P0/Z, [X13, #2]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
MOVPRFX Z18, Z5	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FRINTA Z18.D, P0/M, Z5.D	1	0	0	0	0	0	0	1	0	0	0	0	0	3	1
MOVPRFX Z21, Z24	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FRINTA Z21.D, P0/M, Z24.D	1	0	0	0	0	0	0	1	0	0	0	0	0	3	1
FSUB Z19.D, Z5.D, Z18.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FSUB Z22.D, Z24.D, Z21.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
MOVPRFX Z20, Z17	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FRINTA Z20.D, P0/M, Z17.D	1	0	0	0	0	0	0	1	0	0	0	0	0	3	1
FMUL Z27.D, Z16.D, Z19.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FSUB Z28.D, Z17.D, Z20.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FMLA Z27.D, P0/M, Z1.D, Z22.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z3.D, P0/M, Z28.D, Z27.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
ST1D {Z3.D}, P1, [X20, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
LD1RD {Z4.D}, P0/Z, [X13, #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z26.D}, P0/Z, [X13, #4]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z5.D}, P0/Z, [X13, #5]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z6.D, Z26.D, Z19.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z6.D, P0/M, Z4.D, Z22.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z5.D, P0/M, Z28.D, Z6.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
ST1D {Z5.D}, P1, [X2, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
LD1RD {Z0.D}, P0/Z, [X13, #7]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z2.D}, P0/Z, [X13, #6]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1RD {Z24.D}, P0/Z, [X13, #8]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z7.D, Z0.D, Z19.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z7.D, P0/M, Z2.D, Z22.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMAD Z24.D, P0/M, Z28.D, Z7.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
ST1D {Z24.D}, P1, [X28, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
LD1D {Z25.D}, P1/Z, [X2, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z23.D}, P1/Z, [X20, X0,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z17.D, Z25.D, Z25.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
FMLA Z17.D, P0/M, Z23.D, Z23.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FMLA Z17.D, P0/M, Z24.D, Z24.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FSQRT Z17.D, P0/M, Z17.D	1	0	0	0	0	0	0	1	0	0	0	0	0	7-16	6.99-14.08
ST1D {Z17.D}, P1, [X8, X0,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
ADD X0, X0, X9	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25
WHILELO P1.D, W0, W10	1	0	0	0	0	1	0	0	0	0	0	0	0	3	2
B.NE 4f7590 <_ZN11qmcplusplus27SoaDistanceTableAAOMPTargetIdLj3ELi40EE4moveERKNS_11ParticleSetERKNS_10TinyVectorIdLj3EEEib+0x2b0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1	0.50

Report Configuration

Vectorization

Details

FMA

Workaround

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

Vectorization

Details

FMA

Workaround

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code