OV - exec - Loop 367

0x419020 LD1D	{Z3.D}, P0/Z, [X8, X1,LSL #3]    [5]

0x419024 LD1D	{Z4.D}, P0/Z, [X2, X1,LSL #3]    [1]

0x419028 LD1D	{Z5.D}, P0/Z, [X14, X1,LSL #3]    [4]

0x41902c FMUL	Z3.D, Z5.D, Z3.D

0x419030 LD1D	{Z6.D}, P0/Z, [X3, X1,LSL #3]    [3]

0x419034 FMUL	Z4.D, Z6.D, Z4.D

0x419038 LD1D	{Z5.D}, P0/Z, [X10, X1,LSL #3]    [9]

0x41903c LD1D	{Z6.D}, P0/Z, [X4, X1,LSL #3]    [6]

0x419040 LD1D	{Z7.D}, P0/Z, [X16, X1,LSL #3]    [8]

0x419044 LD1D	{Z16.D}, P0/Z, [X5, X1,LSL #3]    [7]

0x419048 LD1D	{Z17.D}, P0/Z, [X13, X1,LSL #3]    [2]

0x41904c LD1D	{Z18.D}, P0/Z, [X6, X1,LSL #3]    [10]

0x419050 FADD	Z1.D, Z3.D, Z1.D

0x419054 FADD	Z2.D, Z4.D, Z2.D

0x419058 FNMSB	Z5.D, P0/M, Z7.D, Z3.D

0x41905c FNMSB	Z6.D, P0/M, Z16.D, Z4.D

0x419060 FADD	Z5.D, Z5.D, Z17.D

0x419064 FADD	Z6.D, Z6.D, Z18.D

0x419068 ST1D	{Z5.D}, P0, [X13, X1,LSL #3]    [2]

0x41906c ST1D	{Z6.D}, P0, [X6, X1,LSL #3]    [10]

0x419070 ADD	X1, X1, X11

0x419074 CMP	X18, X1

0x419078 B.NE	419020

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h: 340 - 345

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

340:     for (int jat = 0; jat < N; jat++)

341:     {

342:       const valT newg = cur_du_pt[jat] * new_dX[jat];

343:       const valT dg   = newg - old_du_pt[jat] * old_dX[jat];

344:       save_g[jat] -= dg;

345:       cur_g += newg;

Coverage (%)	Name	Source Location	Module
►100.00+	qmcplusplus::WaveFunction::acc[...]	NewTimer.h:249	exec
○	.omp_outlined..62	miniqmc.cpp:445	exec
○	__kmp_invoke_microtask		libomp.so

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	1.50
Bottlenecks	P10, P11,
Function	miniqmcreference::TwoBodyJastrowRef >::acceptMove(qmcplusplus::ParticleSet&, int)
Source	TwoBodyJastrowRef.h:340-345
Source loop unroll info	unrolled by 8
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	8
CQA cycles	6.00
CQA cycles if no scalar integer	6.00
CQA cycles if FP arith vectorized	6.00
CQA cycles if fully vectorized	6.00
Front-end cycles	2.88
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.50
P5 cycles	4.00
P6 cycles	4.00
P7 cycles	4.00
P8 cycles	4.00
P9 cycles	6.00
P10 cycles	6.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	0.00
Inter-iter dependencies cycles	2
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	23.00
Nb uops	23.00
Nb loads	NA
Nb stores	2.00
Nb stack references	0.00
FLOP/cycle	6.67
Nb FLOP add-sub	16.00
Nb FLOP mul	8.00
Nb FLOP fma	8.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	64.00
Bytes prefetched	0.00
Bytes loaded	320.00
Bytes stored	64.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	100.00
Vectorization ratio load	100.00
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	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	100.00
Vector-efficiency ratio load	100.00
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	NA
Vector-efficiency ratio other	NA

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	1.50
Bottlenecks	P10, P11,
Function	miniqmcreference::TwoBodyJastrowRef >::acceptMove(qmcplusplus::ParticleSet&, int)
Source	TwoBodyJastrowRef.h:340-345
Source loop unroll info	unrolled by 8
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	8
CQA cycles	6.00
CQA cycles if no scalar integer	6.00
CQA cycles if FP arith vectorized	6.00
CQA cycles if fully vectorized	6.00
Front-end cycles	2.88
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.50
P5 cycles	4.00
P6 cycles	4.00
P7 cycles	4.00
P8 cycles	4.00
P9 cycles	6.00
P10 cycles	6.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	0.00
Inter-iter dependencies cycles	2
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	23.00
Nb uops	23.00
Nb loads	NA
Nb stores	2.00
Nb stack references	0.00
FLOP/cycle	6.67
Nb FLOP add-sub	16.00
Nb FLOP mul	8.00
Nb FLOP fma	8.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	64.00
Bytes prefetched	0.00
Bytes loaded	320.00
Bytes stored	64.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	100.00
Vectorization ratio load	100.00
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	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	100.00
Vector-efficiency ratio load	100.00
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	NA
Vector-efficiency ratio other	NA

Path /

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

Function	miniqmcreference::TwoBodyJastrowRef >::acceptMove(qmcplusplus::ParticleSet&, int)
Source file and lines	TwoBodyJastrowRef.h:340-345
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:340-345.

The related source loop is unrolled by 8 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is fully vectorized, using full register length.

Details

All VPU instructions are used in vector version (process two or more data elements in vector registers).

FMA

Detected 8 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 40 FP arithmetical operations:

24: addition or subtraction (8 inside FMA instructions)
16: multiply (8 inside FMA instructions)

The binary loop is loading 320 bytes. The binary loop is storing 64 bytes.

Arithmetic intensity

Arithmetic intensity is 0.10 FP operations per loaded or stored byte.

General properties

nb instructions	23
loop length	92
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

2.88 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	0.50	0.50	0.50	4.00	4.00	1.00	1.00	6.00	6.00	0.00	0.00	0.00
cycles	0.50	0.50	0.50	0.50	0.50	0.50	4.00	4.00	4.00	4.00	6.00	6.00	0.00	0.00	0.00

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	2.00

Cycles summary

Front-end	2.88
Data deps.	2.00
Overall L1	6.00

Vectorization ratios

INT

all	100%
load	100%
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	NA (no other vectorizable/vectorized instructions)

FP

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	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 6.00 cycles. At this rate:

55% of peak load performance is reached (53.33 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
22% of peak store performance is reached (10.67 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 419020

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput
LD1D {Z3.D}, P0/Z, [X8, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z4.D}, P0/Z, [X2, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z5.D}, P0/Z, [X14, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z3.D, Z5.D, Z3.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
LD1D {Z6.D}, P0/Z, [X3, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z4.D, Z6.D, Z4.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
LD1D {Z5.D}, P0/Z, [X10, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z6.D}, P0/Z, [X4, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z7.D}, P0/Z, [X16, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z16.D}, P0/Z, [X5, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z17.D}, P0/Z, [X13, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z18.D}, P0/Z, [X6, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FADD Z1.D, Z3.D, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FADD Z2.D, Z4.D, Z2.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FNMSB Z5.D, P0/M, Z7.D, Z3.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FNMSB Z6.D, P0/M, Z16.D, Z4.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FADD Z5.D, Z5.D, Z17.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FADD Z6.D, Z6.D, Z18.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
ST1D {Z5.D}, P0, [X13, X1,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
ST1D {Z6.D}, P0, [X6, X1,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 X1, X1, X11	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25
CMP X18, X1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33
B.NE 419020 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE10acceptMoveERNS1_11ParticleSetEi+0x720>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	1	0.50

Function	miniqmcreference::TwoBodyJastrowRef >::acceptMove(qmcplusplus::ParticleSet&, int)
Source file and lines	TwoBodyJastrowRef.h:340-345
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:340-345.

The related source loop is unrolled by 8 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is fully vectorized, using full register length.

Details

All VPU instructions are used in vector version (process two or more data elements in vector registers).

FMA

Detected 8 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 40 FP arithmetical operations:

24: addition or subtraction (8 inside FMA instructions)
16: multiply (8 inside FMA instructions)

The binary loop is loading 320 bytes. The binary loop is storing 64 bytes.

Arithmetic intensity

Arithmetic intensity is 0.10 FP operations per loaded or stored byte.

General properties

nb instructions	23
loop length	92
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

2.88 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	0.50	0.50	0.50	4.00	4.00	1.00	1.00	6.00	6.00	0.00	0.00	0.00
cycles	0.50	0.50	0.50	0.50	0.50	0.50	4.00	4.00	4.00	4.00	6.00	6.00	0.00	0.00	0.00

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	2.00

Cycles summary

Front-end	2.88
Data deps.	2.00
Overall L1	6.00

Vectorization ratios

INT

all	100%
load	100%
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	NA (no other vectorizable/vectorized instructions)

FP

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	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 6.00 cycles. At this rate:

55% of peak load performance is reached (53.33 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
22% of peak store performance is reached (10.67 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 419020

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	Latency	Recip. throughput
LD1D {Z3.D}, P0/Z, [X8, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z4.D}, P0/Z, [X2, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z5.D}, P0/Z, [X14, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z3.D, Z5.D, Z3.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
LD1D {Z6.D}, P0/Z, [X3, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FMUL Z4.D, Z6.D, Z4.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	3	0.50
LD1D {Z5.D}, P0/Z, [X10, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z6.D}, P0/Z, [X4, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z7.D}, P0/Z, [X16, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z16.D}, P0/Z, [X5, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z17.D}, P0/Z, [X13, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
LD1D {Z18.D}, P0/Z, [X6, X1,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	6	0.50
FADD Z1.D, Z3.D, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FADD Z2.D, Z4.D, Z2.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FNMSB Z5.D, P0/M, Z7.D, Z3.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FNMSB Z6.D, P0/M, Z16.D, Z4.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	4	0.50
FADD Z5.D, Z5.D, Z17.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
FADD Z6.D, Z6.D, Z18.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	2	0.50
ST1D {Z5.D}, P0, [X13, X1,LSL #3]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	2	0.50
ST1D {Z6.D}, P0, [X6, X1,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 X1, X1, X11	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25
CMP X18, X1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.33
B.NE 419020 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE10acceptMoveERNS1_11ParticleSetEi+0x720>	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