OV - exec - Loop 137

0x40d6f0 LDR	X3, [X22]    [6]

0x40d6f4 LDR	X21, [X3, X19,LSL #3]    [1]

0x40d6f8 ORR	X0, XZR, X21

0x40d6fc BL	479f20

0x40d700 LDR	X8, [X20]    [4]

0x40d704 ORR	X3, XZR, X23

0x40d708 ORR	W2, WZR, W24

0x40d70c ORR	X1, XZR, X25

0x40d710 LDR	X0, [X8, X19,LSL #3]    [5]

0x40d714 LDR	X9, [X0]    [3]

0x40d718 LDR	X10, [X9, #32]    [2]

0x40d71c BLR	X10

0x40d720 FMUL	D8, D8, D0

0x40d724 ORR	X0, XZR, X21

0x40d728 ADD	X19, X19, #1

0x40d72c BL	47a180

0x40d730 LDP	X11, X12, [X20]    [4]

0x40d734 SUB	X13, X12, X11

0x40d738 CMP	X19, X13,ASR #3

0x40d73c B.CC	40d6f0

/usr/include/c++/11/bits/refwrap.h: 338 - 338

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

338:       { return *_M_data; }

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/WaveFunction.cpp: 204 - 207

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

204:   for (size_t i = 0; i < Jastrows.size(); i++)

205:   {

206:     ScopedTimer local_timer(jastrow_timers[i]);

207:     ratio *= Jastrows[i]->ratioGrad(P, iat, grad);

/usr/include/c++/11/bits/stl_vector.h: 919 - 919

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

919:       { return size_type(this->_M_impl._M_finish - this->_M_impl._M_start); }

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Utilities/NewTimer.h: 242 - 249

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

242:   ScopeGuard(TIMER& t) : timer(t) { timer.start(); }

[...]

249:   ~ScopeGuard() { timer.stop(); }

Coverage (%)	Name	Source Location	Module
►50.00+	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
►50.00+	main._omp_fn.1	stl_vector.h:1043	exec
○	GOMP_parallel		libomp.so
○	main	miniqmc.cpp:404	exec
○	__libc_start_main		libc-2.31.so
○	_start	miniqmc.cpp:546	exec

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	7.62
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.09
Bottlenecks
Function	qmcplusplus::WaveFunction::ratioGrad(qmcplusplus::ParticleSet&, int, qmcplusplus::TinyVector&)
Source	refwrap.h:338-338,WaveFunction.cpp:204-207,stl_vector.h:919-919,NewTimer.h:242-242,NewTimer.h:249-249
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	3.00
CQA cycles if no scalar integer	3.00
CQA cycles if FP arith vectorized	3.00
CQA cycles if fully vectorized	0.39
Front-end cycles	2.50
DIV/SQRT cycles	2.00
P0 cycles	2.00
P1 cycles	2.75
P2 cycles	2.75
P3 cycles	2.75
P4 cycles	2.75
P5 cycles	0.25
P6 cycles	0.25
P7 cycles	0.25
P8 cycles	0.25
P9 cycles	2.33
P10 cycles	2.33
P11 cycles	2.33
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	0.00
Inter-iter dependencies cycles	3
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	20.00
Nb uops	20.00
Nb loads	NA
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	0.33
Nb FLOP add-sub	0.00
Nb FLOP mul	1.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	21.33
Bytes prefetched	0.00
Bytes loaded	64.00
Bytes stored	0.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	NA
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	18.75
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	18.75

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	7.62
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.09
Bottlenecks
Function	qmcplusplus::WaveFunction::ratioGrad(qmcplusplus::ParticleSet&, int, qmcplusplus::TinyVector&)
Source	refwrap.h:338-338,WaveFunction.cpp:204-207,stl_vector.h:919-919,NewTimer.h:242-242,NewTimer.h:249-249
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	3.00
CQA cycles if no scalar integer	3.00
CQA cycles if FP arith vectorized	3.00
CQA cycles if fully vectorized	0.39
Front-end cycles	2.50
DIV/SQRT cycles	2.00
P0 cycles	2.00
P1 cycles	2.75
P2 cycles	2.75
P3 cycles	2.75
P4 cycles	2.75
P5 cycles	0.25
P6 cycles	0.25
P7 cycles	0.25
P8 cycles	0.25
P9 cycles	2.33
P10 cycles	2.33
P11 cycles	2.33
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	0.00
Inter-iter dependencies cycles	3
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	20.00
Nb uops	20.00
Nb loads	NA
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	0.33
Nb FLOP add-sub	0.00
Nb FLOP mul	1.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	21.33
Bytes prefetched	0.00
Bytes loaded	64.00
Bytes stored	0.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	NA
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	18.75
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	18.75

Path /

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

Function	qmcplusplus::WaveFunction::ratioGrad(qmcplusplus::ParticleSet&, int, qmcplusplus::TinyVector&)
Source file and lines	WaveFunction.cpp:204-207
Module	exec

The loop is defined in:

/usr/include/c++/11/bits/refwrap.h:338
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/WaveFunction.cpp:204-207
/usr/include/c++/11/bits/stl_vector.h:919
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Utilities/NewTimer.h:242,249

The related source loop is not unrolled or unrolled with no peel/tail loop.

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. Only 18% of vector register length is used (average across all VPU instructions). By vectorizing your loop, you can lower the cost of an iteration from 3.00 to 0.39 cycles (7.62x 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 loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with ftree-vectorize (included in O3) to enable loop vectorization (including FP reductions).
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop 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 loop (in the source code) and the binary loop

The binary loop is composed of 1 FP arithmetical operations:

1: multiply

The binary loop is loading 64 bytes.

Arithmetic intensity

Arithmetic intensity is 0.02 FP operations per loaded or stored byte.

General properties

nb instructions	20
loop length	80
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

2.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	2.00	2.00	2.75	2.75	2.75	2.75	0.25	0.25	0.25	0.25	2.33	2.33	2.33	0.00	0.00
cycles	2.00	2.00	2.75	2.75	2.75	2.75	0.25	0.25	0.25	0.25	2.33	2.33	2.33	0.00	0.00

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	3.00

Cycles summary

Front-end	2.50
Data deps.	3.00
Overall L1	3.00

Vectorization ratios

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
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 iteration of the binary loop takes 3.00 cycles. At this rate:

22% of peak load performance is reached (21.33 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 40d6f0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput
LDR X3, [X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
LDR X21, [X3, X19,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
ORR X0, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
BL 479f20 <_ZN11qmcplusplus9TimerTypeINS_8CPUClockEE5startEv>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
ORR X3, XZR, X23	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
ORR W2, WZR, W24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
ORR X1, XZR, X25	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
LDR X0, [X8, X19,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
LDR X9, [X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
LDR X10, [X9, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
BLR X10	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
FMUL D8, D8, D0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25
ORR X0, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
ADD X19, X19, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
BL 47a180 <_ZN11qmcplusplus9TimerTypeINS_8CPUClockEE4stopEv>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
LDP X11, X12, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	1
SUB X13, X12, X11	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
CMP X19, X13,ASR #3	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B.CC 40d6f0 <_ZN11qmcplusplus12WaveFunction9ratioGradERNS_11ParticleSetEiRNS_10TinyVectorIdLj3EEE+0x70>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50

Function	qmcplusplus::WaveFunction::ratioGrad(qmcplusplus::ParticleSet&, int, qmcplusplus::TinyVector&)
Source file and lines	WaveFunction.cpp:204-207
Module	exec

The loop is defined in:

/usr/include/c++/11/bits/refwrap.h:338
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/WaveFunction.cpp:204-207
/usr/include/c++/11/bits/stl_vector.h:919
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Utilities/NewTimer.h:242,249

The related source loop is not unrolled or unrolled with no peel/tail loop.

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. Only 18% of vector register length is used (average across all VPU instructions). By vectorizing your loop, you can lower the cost of an iteration from 3.00 to 0.39 cycles (7.62x 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 loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with ftree-vectorize (included in O3) to enable loop vectorization (including FP reductions).
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop 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 loop (in the source code) and the binary loop

The binary loop is composed of 1 FP arithmetical operations:

1: multiply

The binary loop is loading 64 bytes.

Arithmetic intensity

Arithmetic intensity is 0.02 FP operations per loaded or stored byte.

General properties

nb instructions	20
loop length	80
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

2.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	2.00	2.00	2.75	2.75	2.75	2.75	0.25	0.25	0.25	0.25	2.33	2.33	2.33	0.00	0.00
cycles	2.00	2.00	2.75	2.75	2.75	2.75	0.25	0.25	0.25	0.25	2.33	2.33	2.33	0.00	0.00

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	3.00

Cycles summary

Front-end	2.50
Data deps.	3.00
Overall L1	3.00

Vectorization ratios

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
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 iteration of the binary loop takes 3.00 cycles. At this rate:

22% of peak load performance is reached (21.33 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 40d6f0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput
LDR X3, [X22]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
LDR X21, [X3, X19,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
ORR X0, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
BL 479f20 <_ZN11qmcplusplus9TimerTypeINS_8CPUClockEE5startEv>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
LDR X8, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
ORR X3, XZR, X23	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
ORR W2, WZR, W24	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
ORR X1, XZR, X25	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
LDR X0, [X8, X19,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
LDR X9, [X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
LDR X10, [X9, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33
BLR X10	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
FMUL D8, D8, D0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25
ORR X0, XZR, X21	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
ADD X19, X19, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
BL 47a180 <_ZN11qmcplusplus9TimerTypeINS_8CPUClockEE4stopEv>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
LDP X11, X12, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	1
SUB X13, X12, X11	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25
CMP X19, X13,ASR #3	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50
B.CC 40d6f0 <_ZN11qmcplusplus12WaveFunction9ratioGradERNS_11ParticleSetEiRNS_10TinyVectorIdLj3EEE+0x70>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50

Report Configuration

Vectorization

Details

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

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