OV - exec - Function miniqmcreference::TwoBodyJastrowRef<qmcplusplus::BsplineFunctor<double> >::ratio(qmcplusplus::ParticleSet&, int)

	Function: miniqmcreference::TwoBodyJastrowRef<qmcplusplus::BsplineFunctor<double> >::ratio(qmcpluspl ...	Module: exec	Source: TwoBodyJastrowRef.h:124-284 [...]	Coverage: 0.04%

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h: 1046 - 1169

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

1046: 	return *(this->_M_impl._M_start + __n);

[...]

1169:       { return _M_data_ptr(this->_M_impl._M_start); }

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h: 223 - 229

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

223:     return X[i];

[...]

229:     return X[i];

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h: 124 - 284

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

124:     const int igt = P.GroupID[iat] * NumGroups;

125:     for (int jg = 0; jg < NumGroups; ++jg)

126:     {

127:       const FuncType& f2(*F[igt + jg]);

128:       int iStart = P.first(jg);

129:       int iEnd   = P.last(jg);

130:       curUat += f2.evaluateV(iat, iStart, iEnd, dist, DistCompressed.data());

[...]

280: {

281:   // only ratio, ready to compute it again

282:   UpdateMode = ORB_PBYP_RATIO;

283:   cur_Uat    = computeU(P, iat, P.DistTables[0]->Temp_r.data());

284:   return std::exp(Uat[iat] - cur_Uat);

/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h: 302 - 305

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

302:   inline int first(int igroup) const { return SubPtcl[igroup]; }

303:

304:   /// return the last index of a group i

305:   inline int last(int igroup) const { return SubPtcl[igroup + 1]; }

0x419480 STR	D8, [SP, #432]!

0x419484 STP	X29, X30, [SP, #8]

0x419488 STR	X25, [SP, #24]

0x41948c STP	X24, X23, [SP, #32]

0x419490 STP	X22, X21, [SP, #48]

0x419494 STP	X20, X19, [SP, #64]

0x419498 ADD	X29, SP, #8

0x41949c LDR	X8, [X0, #160]

0x4194a0 ORR	W20, WZR, W2

0x4194a4 ORR	X19, XZR, X0

0x4194a8 STR	WZR, [X0, #12]

0x4194ac CBZ	X8, 419520

0x4194b0 LDR	X9, [X1, #2576]

0x4194b4 ORR	W10, WZR, W20

0x4194b8 ORR	X21, XZR, X1

0x4194bc MOVI	D8, #0

0x4194c0 SBFM	X23, X10, #0, #31

0x4194c4 ORR	X24, XZR, XZR

0x4194c8 LDR	X9, [X9]

0x4194cc LDR	X22, [X9, #104]

0x4194d0 LDR	X9, [X1, #1472]

0x4194d4 LDR	W9, [X9, X20,SXTW #2]

0x4194d8 MADD	W8, W9, W8, WZR

0x4194dc SBFM	X25, X8, #61, #31

(378) 0x4194e0 LDR	X8, [X19, #512]

(378) 0x4194e4 LDR	X5, [X19, #464]

(378) 0x4194e8 ORR	W1, WZR, W20

(378) 0x4194ec ORR	X4, XZR, X22

(378) 0x4194f0 ADD	X8, X8, X25

(378) 0x4194f4 LDR	X0, [X8, X24,LSL #3]

(378) 0x4194f8 LDR	X8, [X21, #2712]

(378) 0x4194fc ADD	X8, X8, X24,LSL #2

(378) 0x419500 ADD	X24, X24, #1

(378) 0x419504 LDP	W2, W3, [X8]

(378) 0x419508 BL	416f40

(378) 0x41950c LDR	X8, [X19, #160]

(378) 0x419510 FADD	D8, D0, D8

(378) 0x419514 CMP	X8, X24

(378) 0x419518 B.HI	4194e0

0x41951c B	419528

0x419520 SBFM	X23, X20, #0, #31

0x419524 MOVI	D8, #0

0x419528 LDR	X8, [X19, #216]

0x41952c STR	D8, [X19, #312]

0x419530 LDR	D0, [X8, X23,LSL #3]

0x419534 FSUB	D0, D0, S8

0x419538 LDP	X20, X19, [SP, #64]

0x41953c LDP	X22, X21, [SP, #48]

0x419540 LDP	X24, X23, [SP, #32]

0x419544 LDR	X25, [SP, #24]

0x419548 LDP	X29, X30, [SP, #8]

0x41954c LDR	D8, [SP], #80

0x419550 B	403d80

0x419554 HINT	#0

0x419558 HINT	#0

0x41955c HINT	#0

Coverage (%)	Name	Source Location	Module
►100.00+	qmcplusplus::WaveFunction::rat[...]	NewTimer.h:249	exec
○	.omp_outlined..62	miniqmc.cpp:485	exec
○	__kmp_invoke_microtask		libomp.so

Path /

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

Source file and lines	TwoBodyJastrowRef.h:124-284
Module	exec

The function is defined in:

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:1046,1169
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,229
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:124-130,280-284
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:302-305

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 7.33 to 1.33 cycles (5.50x 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 23% of vector register length is used (average across all VPU instructions). By vectorizing your function, you can lower the cost of an iteration from 7.33 to 1.08 cycles (6.77x 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 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 is composed of 1 FP arithmetical operations:

1: addition or subtraction

The binary function is loading 140 bytes. The binary function is storing 92 bytes.

Arithmetic intensity

Arithmetic intensity is 0.00 FP operations per loaded or stored byte.

General properties

nb instructions	41
loop length	164
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

4.75 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	1.50	1.50	2.75	2.75	2.75	2.75	1.25	1.25	1.25	1.25	7.33	7.33	7.33	3.00	3.00
cycles	1.50	1.50	2.75	2.75	2.75	2.75	1.25	1.25	1.25	1.25	7.33	7.33	7.33	3.00	3.00

Cycles executing div or sqrt instructions

Cycles summary

Front-end	4.75
Overall L1	7.33

Vectorization ratios

all	0%
load	0%
store	0%
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 7.33 cycles. At this rate:

19% of peak load performance is reached (19.09 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
26% of peak store performance is reached (12.55 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 419480

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
STR D8, [SP, #432]!	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
STP X29, X30, [SP, #8]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STR X25, [SP, #24]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
ADD X29, SP, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X8, [X0, #160]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR W20, WZR, W2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X19, XZR, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR WZR, [X0, #12]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
CBZ X8, 419520 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE5ratioERNS1_11ParticleSetEi+0xa0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X9, [X1, #2576]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR W10, WZR, W20	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X21, XZR, X1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVI D8, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
SBFM X23, X10, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X24, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X9, [X9]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X22, [X9, #104]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X9, [X1, #1472]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR W9, [X9, X20,SXTW #2]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
MADD W8, W9, W8, WZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
SBFM X25, X8, #61, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
B 419528 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE5ratioERNS1_11ParticleSetEi+0xa8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
SBFM X23, X20, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVI D8, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
LDR X8, [X19, #216]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
STR D8, [X19, #312]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D0, [X8, X23,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
FSUB D0, D0, S8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
LDP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDR X25, [SP, #24]	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, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDR D8, [SP], #80	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
B 403d80 <@plt_start@+0x450>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0
HINT #0
HINT #0

Source file and lines	TwoBodyJastrowRef.h:124-284
Module	exec

The function is defined in:

/usr/lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:1046,1169
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,229
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:124-130,280-284
/home/hbollore/qaas/qaas-runs/169-816-8620/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:302-305

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 7.33 to 1.33 cycles (5.50x speedup).

Workaround

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

Vectorization

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 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 is composed of 1 FP arithmetical operations:

1: addition or subtraction

The binary function is loading 140 bytes. The binary function is storing 92 bytes.

Arithmetic intensity

Arithmetic intensity is 0.00 FP operations per loaded or stored byte.

General properties

nb instructions	41
loop length	164
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

4.75 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	1.50	1.50	2.75	2.75	2.75	2.75	1.25	1.25	1.25	1.25	7.33	7.33	7.33	3.00	3.00
cycles	1.50	1.50	2.75	2.75	2.75	2.75	1.25	1.25	1.25	1.25	7.33	7.33	7.33	3.00	3.00

Cycles executing div or sqrt instructions

Cycles summary

Front-end	4.75
Overall L1	7.33

Vectorization ratios

all	0%
load	0%
store	0%
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 7.33 cycles. At this rate:

19% of peak load performance is reached (19.09 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
26% of peak store performance is reached (12.55 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the function is: 419480

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
STR D8, [SP, #432]!	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
STP X29, X30, [SP, #8]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STR X25, [SP, #24]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
STP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
ADD X29, SP, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X8, [X0, #160]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR W20, WZR, W2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X19, XZR, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
STR WZR, [X0, #12]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
CBZ X8, 419520 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE5ratioERNS1_11ParticleSetEi+0xa0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR X9, [X1, #2576]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ORR W10, WZR, W20	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X21, XZR, X1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVI D8, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
SBFM X23, X10, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X24, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
LDR X9, [X9]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X22, [X9, #104]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR X9, [X1, #1472]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
LDR W9, [X9, X20,SXTW #2]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
MADD W8, W9, W8, WZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
SBFM X25, X8, #61, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
B 419528 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE5ratioERNS1_11ParticleSetEi+0xa8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
SBFM X23, X20, #0, #31	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MOVI D8, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
LDR X8, [X19, #216]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
STR D8, [X19, #312]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D0, [X8, X23,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
FSUB D0, D0, S8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
LDP X20, X19, [SP, #64]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X22, X21, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDP X24, X23, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDR X25, [SP, #24]	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, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
LDR D8, [SP], #80	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
B 403d80 <@plt_start@+0x450>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
HINT #0
HINT #0
HINT #0

Name	Coverage (%)	Time (s)
▼miniqmcreference::TwoBodyJastrowRef >::ratio(qmcplusplus::ParticleSet&, int)–	0.04	0.02
○Loop 378 - TwoBodyJastrowRef.h:125-130 - exec	0.03	0.01

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

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

Arithmetic intensity

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code