OV - exec - Loop 711

	Loop Id: 711	Module: exec	Source: MultiBsplineRef.hpp:276-286	Coverage: 0.08%

0x255130 VMULPD	(%R14,%RCX,8),%XMM6,%XMM15    [6]

0x255136 VMOVUPD	%XMM15,(%R14,%RCX,8)    [6]

0x25513c VMULPD	(%RDX,%RCX,8),%XMM7,%XMM15    [1]

0x255141 VMOVUPD	%XMM15,(%RDX,%RCX,8)    [1]

0x255146 VMULPD	(%R13,%RCX,8),%XMM8,%XMM15    [5]

0x25514d VMOVUPD	%XMM15,(%R13,%RCX,8)    [5]

0x255154 VMULPD	(%R15,%RCX,8),%XMM9,%XMM15    [7]

0x25515a VMOVUPD	%XMM15,(%R15,%RCX,8)    [7]

0x255160 VMULPD	(%R9,%RCX,8),%XMM10,%XMM15    [4]

0x255166 VMOVUPD	%XMM15,(%R9,%RCX,8)    [4]

0x25516c VMULPD	(%R11,%RCX,8),%XMM11,%XMM15    [9]

0x255172 VMOVUPD	%XMM15,(%R11,%RCX,8)    [9]

0x255178 VMULPD	(%RDI,%RCX,8),%XMM12,%XMM15    [2]

0x25517d VMOVUPD	%XMM15,(%RDI,%RCX,8)    [2]

0x255182 VMULPD	(%R8,%RCX,8),%XMM13,%XMM15    [3]

0x255188 VMOVUPD	%XMM15,(%R8,%RCX,8)    [3]

0x25518e VMULPD	(%R10,%RCX,8),%XMM14,%XMM15    [8]

0x255194 VMOVUPD	%XMM15,(%R10,%RCX,8)    [8]

0x25519a ADD	$0x2,%RCX

0x25519e CMP	%RCX,%RAX

0x2551a1 JNE	255130

/home/eoseret/qaas_runs_ZEN5/173-938-0042/intel/miniqmc/build/miniqmc/src/Numerics/Spline2/MultiBsplineRef.hpp: 276 - 286

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

276:   for (int n = 0; n < num_splines; n++)

277:   {

278:     gx[n] *= dxInv;

279:     gy[n] *= dyInv;

280:     gz[n] *= dzInv;

281:     hxx[n] *= dxx;

282:     hyy[n] *= dyy;

283:     hzz[n] *= dzz;

284:     hxy[n] *= dxy;

285:     hxz[n] *= dxz;

286:     hyz[n] *= dyz;

Coverage (%)	Name	Source Location	Module
►79.53+	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:206	exec
○	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:219	exec
○	miniqmcreference::DiracDetermi[...]	DiracDeterminantRef.cpp:100	exec
○	qmcplusplus::WaveFunction::rat[...]	stl_vector.h:919	exec
○	main.omp_outlined.64	stl_vector.h:1046	exec
○	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so
○	__kmp_launch_thread		libomp.so
○	__kmp_launch_worker(void*)		libomp.so
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►8.67+	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:206	exec
○	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:219	exec
○	qmcplusplus::SPOSet::evaluate_[...]	OhmmsVector.h:144	exec
○	miniqmcreference::DiracDetermi[...]	DiracDeterminantRef.cpp:263	exec
○	miniqmcreference::DiracDetermi[...]	DiracDeterminantRef.cpp:238	exec
○	qmcplusplus::WaveFunction::eva[...]	WaveFunction.cpp:170	exec
○	main.omp_outlined	miniqmc.cpp:379	exec
○	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so
○	__kmp_launch_thread		libomp.so
○	__kmp_launch_worker(void*)		libomp.so
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►7.80+	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:206	exec
○	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:219	exec
○	qmcplusplus::SPOSet::evaluate_[...]	OhmmsVector.h:144	exec
○	miniqmcreference::DiracDetermi[...]	DiracDeterminantRef.cpp:263	exec
○	miniqmcreference::DiracDetermi[...]	DiracDeterminantRef.cpp:238	exec
○	qmcplusplus::WaveFunction::eva[...]	WaveFunction.cpp:171	exec
○	main.omp_outlined	miniqmc.cpp:379	exec
○	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so
○	__kmp_launch_thread		libomp.so
○	__kmp_launch_worker(void*)		libomp.so
○	start_thread		libc.so.6
○	__GI___clone3		libc.so.6
►2.54+	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:206	exec
○	miniqmcreference::einspline_sp[...]	einspline_spo_ref.hpp:219	exec
○	miniqmcreference::DiracDetermi[...]	DiracDeterminantRef.cpp:100	exec
○	qmcplusplus::WaveFunction::rat[...]	stl_vector.h:919	exec
○	main.omp_outlined.64	stl_vector.h:1046	exec
○	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so
○	__kmp_fork_call		libomp.so
○	__kmpc_fork_call		libomp.so
○	main	miniqmc.cpp:411	exec
○	__libc_start_call_main		libc.so.6

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.80
Bottlenecks	P6, P7, P8, P9, P10, P11, P14, P15,
Function	void miniqmcreference::MultiBsplineEvalRef::evaluate_vgh(qmcplusplus::bspline_traits::SplineType const, double, double, double, double, double, double, unsigned long)
Source	MultiBsplineRef.hpp:276-286
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	max
Unroll/vectorization loop type	main
Unroll factor	2
CQA cycles	4.50
CQA cycles if no scalar integer	4.50
CQA cycles if FP arith vectorized	4.50
CQA cycles if fully vectorized	1.13
Front-end cycles	2.50
P0 cycles	0.33
P1 cycles	0.33
P2 cycles	0.33
P3 cycles	0.33
P4 cycles	0.33
P5 cycles	0.33
P6 cycles	4.50
P7 cycles	4.50
P8 cycles	4.50
P9 cycles	4.50
P10 cycles	4.50
P11 cycles	4.50
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	4.50
P15 cycles	4.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	21.00
Nb uops	20.00
Nb loads	9.00
Nb stores	9.00
Nb stack references	0.00
FLOP/cycle	4.00
Nb FLOP add-sub	0.00
Nb FLOP mul	18.00
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	64.00
Bytes prefetched	0.00
Bytes loaded	144.00
Bytes stored	144.00
Stride 0	0.00
Stride 1	9.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	25.00
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
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	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.80
Bottlenecks	P6, P7, P8, P9, P10, P11, P14, P15,
Function	void miniqmcreference::MultiBsplineEvalRef::evaluate_vgh(qmcplusplus::bspline_traits::SplineType const, double, double, double, double, double, double, unsigned long)
Source	MultiBsplineRef.hpp:276-286
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	max
Unroll/vectorization loop type	main
Unroll factor	2
CQA cycles	4.50
CQA cycles if no scalar integer	4.50
CQA cycles if FP arith vectorized	4.50
CQA cycles if fully vectorized	1.13
Front-end cycles	2.50
P0 cycles	0.33
P1 cycles	0.33
P2 cycles	0.33
P3 cycles	0.33
P4 cycles	0.33
P5 cycles	0.33
P6 cycles	4.50
P7 cycles	4.50
P8 cycles	4.50
P9 cycles	4.50
P10 cycles	4.50
P11 cycles	4.50
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	4.50
P15 cycles	4.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	21.00
Nb uops	20.00
Nb loads	9.00
Nb stores	9.00
Nb stack references	0.00
FLOP/cycle	4.00
Nb FLOP add-sub	0.00
Nb FLOP mul	18.00
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	64.00
Bytes prefetched	0.00
Bytes loaded	144.00
Bytes stored	144.00
Stride 0	0.00
Stride 1	9.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	25.00
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
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	void miniqmcreference::MultiBsplineEvalRef::evaluate_vgh(qmcplusplus::bspline_traits::SplineType const, double, double, double, double, double, double, unsigned long)
Source file and lines	MultiBsplineRef.hpp:276-286
Module	exec

The loop is defined in /home/eoseret/qaas_runs_ZEN5/173-938-0042/intel/miniqmc/build/miniqmc/src/Numerics/Spline2/MultiBsplineRef.hpp:276-286.

It is main loop of related source loop which is unrolled by 2 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is vectorized, but using only 128 out of 512 bits (SSE/AVX-128 instructions on AVX-512 processors).

By fully vectorizing your loop, you can lower the cost of an iteration from 4.50 to 1.12 cycles (4.00x speedup).

Details

All SSE/AVX instructions are used in vector version (process two or more data elements in vector registers). Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Execution units bottlenecks

Performance is limited by:

execution of FP multiply or FMA (fused multiply-add) operations (the FP multiply/FMA unit is a bottleneck)
reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

Workaround

Reduce the number of FP multiply/FMA instructions
Read less array elements
Write less array elements
Provide more information to your compiler:
- hardcode the bounds of the corresponding 'for' loop

No data for this section

Type of elements and instruction set

9 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in vector mode (two at a time).

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

The binary loop is composed of 18 FP arithmetical operations:

18: multiply

The binary loop is loading 144 bytes (18 double precision FP elements). The binary loop is storing 144 bytes (18 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.

General properties

nb instructions	21
nb uops	20
loop length	115
used x86 registers	11
used mmx registers	0
used xmm registers	10
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	2.50 cycles
front end	2.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.33	0.33	0.33	0.33	0.33	0.33	4.50	4.50	4.50	4.50	4.50	4.50	0.00	0.00	4.50	4.50
cycles	0.33	0.33	0.33	0.33	0.33	0.33	4.50	4.50	4.50	4.50	4.50	4.50	0.00	0.00	4.50	4.50

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	2.50
Dispatch	4.50
Data deps.	1.00
Overall L1	4.50

Vectorization ratios

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

Vector efficiency ratios

all	25%
load	25%
store	25%
mul	25%
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	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 4.50 cycles. At this rate:

12% of peak load performance is reached (32.00 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
25% of peak store performance is reached (32.00 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 255130

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P14	P15	Latency	Recip. throughput	Vectorization
VMULPD (%R14,%RCX,8),%XMM6,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R14,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%RDX,%RCX,8),%XMM7,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%RDX,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R13,%RCX,8),%XMM8,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R13,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R15,%RCX,8),%XMM9,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R15,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R9,%RCX,8),%XMM10,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R9,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R11,%RCX,8),%XMM11,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R11,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%RDI,%RCX,8),%XMM12,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%RDI,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R8,%RCX,8),%XMM13,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R8,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R10,%RCX,8),%XMM14,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R10,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
ADD $0x2,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %RCX,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
JNE 255130 <_ZN16miniqmcreference19MultiBsplineEvalRef12evaluate_vghIdEEvPKN11qmcplusplus14bspline_traitsIT_Lj3EE10SplineTypeES4_S4_S4_PS4_S9_S9_m+0x1400>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Function	void miniqmcreference::MultiBsplineEvalRef::evaluate_vgh(qmcplusplus::bspline_traits::SplineType const, double, double, double, double, double, double, unsigned long)
Source file and lines	MultiBsplineRef.hpp:276-286
Module	exec

gain
potential
hint
expert

Vectorization

Your loop is vectorized, but using only 128 out of 512 bits (SSE/AVX-128 instructions on AVX-512 processors).

By fully vectorizing your loop, you can lower the cost of an iteration from 4.50 to 1.12 cycles (4.00x speedup).

Details

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Execution units bottlenecks

Performance is limited by:

execution of FP multiply or FMA (fused multiply-add) operations (the FP multiply/FMA unit is a bottleneck)
reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

Workaround

Reduce the number of FP multiply/FMA instructions
Read less array elements
Write less array elements
Provide more information to your compiler:
- hardcode the bounds of the corresponding 'for' loop

No data for this section

Type of elements and instruction set

9 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in vector mode (two at a time).

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

The binary loop is composed of 18 FP arithmetical operations:

18: multiply

The binary loop is loading 144 bytes (18 double precision FP elements). The binary loop is storing 144 bytes (18 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.

General properties

nb instructions	21
nb uops	20
loop length	115
used x86 registers	11
used mmx registers	0
used xmm registers	10
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

ASSUMED MACRO FUSION FIT IN UOP CACHE

micro-operation queue	2.50 cycles
front end	2.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	0.33	0.33	0.33	0.33	0.33	0.33	4.50	4.50	4.50	4.50	4.50	4.50	0.00	0.00	4.50	4.50
cycles	0.33	0.33	0.33	0.33	0.33	0.33	4.50	4.50	4.50	4.50	4.50	4.50	0.00	0.00	4.50	4.50

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	2.50
Dispatch	4.50
Data deps.	1.00
Overall L1	4.50

Vectorization ratios

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

Vector efficiency ratios

all	25%
load	25%
store	25%
mul	25%
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	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 4.50 cycles. At this rate:

12% of peak load performance is reached (32.00 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
25% of peak store performance is reached (32.00 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 255130

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P14	P15	Latency	Recip. throughput	Vectorization
VMULPD (%R14,%RCX,8),%XMM6,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R14,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%RDX,%RCX,8),%XMM7,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%RDX,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R13,%RCX,8),%XMM8,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R13,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R15,%RCX,8),%XMM9,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R15,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R9,%RCX,8),%XMM10,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R9,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R11,%RCX,8),%XMM11,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R11,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%RDI,%RCX,8),%XMM12,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%RDI,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R8,%RCX,8),%XMM13,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R8,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
VMULPD (%R10,%RCX,8),%XMM14,%XMM15	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	0	3	0.50	vect (25.0%)
VMOVUPD %XMM15,(%R10,%RCX,8)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	4	0.50	vect (25.0%)
ADD $0x2,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP %RCX,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	1	0.17	N/A
JNE 255130 <_ZN16miniqmcreference19MultiBsplineEvalRef12evaluate_vghIdEEvPKN11qmcplusplus14bspline_traitsIT_Lj3EE10SplineTypeES4_S4_S4_PS4_S9_S9_m+0x1400>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	1	0.50-1	N/A

Report Configuration

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Type of elements and instruction set

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

Type of elements and instruction set

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code