OV - exec - Loop 371

	Loop Id: 371	Module: exec	Source: TwoBodyJastrowRef.h:107-132 [...]	Coverage: 0.02%

0x427000 XORPD	%XMM1,%XMM1

0x427004 MOVSD	-0x48(%RBP),%XMM0

0x427009 SUBSD	%XMM1,%XMM0

0x42700d CALL	4f1fd0 <exp>

0x427012 MOV	-0x40(%RBP),%RCX

0x427016 MOV	(%RCX),%RAX

0x427019 MOV	-0x50(%RBP),%RBX

0x42701d MOVSD	%XMM0,(%RAX,%RBX,8)

0x427022 INC	%RBX

0x427025 MOV	0x8(%RCX),%RCX

0x427029 SUB	%RAX,%RCX

0x42702c SAR	$0x3,%RCX

0x427030 CMP	%RBX,%RCX

0x427033 MOV	-0x38(%RBP),%RCX

0x427037 JBE	42711b

0x42703d CMPB	$0,0x298(%RCX)

0x427044 JE	42712a

0x42704a MOVSXD	0x2a0(%RCX),%R12

0x427051 MOV	0xd8(%R15),%RAX

0x427058 MOVSD	(%RAX,%R12,8),%XMM0

0x42705e MOVSD	%XMM0,-0x48(%RBP)

0x427063 MOV	0x290(%RCX),%RAX

0x42706a MOV	%RAX,-0x30(%RBP)

0x42706e MOV	0x248(%R15),%ESI

0x427075 MOV	%RCX,%RDI

0x427078 CALL	462dd0 <_ZNK11qmcplusplus11ParticleSet14getDistTableABEi>

0x42707d MOV	0xa0(%R15),%RCX

0x427084 TEST	%RCX,%RCX

0x427087 MOV	%RBX,-0x50(%RBP)

0x42708b JE	427000

0x427091 MOV	0x48(%RAX),%RAX

0x427095 LEA	(%RBX,%RBX,4),%RDX

0x427099 MOV	0x18(%RAX,%RDX,8),%R13

0x42709e MOV	-0x30(%RBP),%RAX

0x4270a2 MOV	0x18(%RAX),%RAX

0x4270a6 IMUL	(%RAX,%R12,4),%ECX

0x4270ab MOVSXD	%ECX,%RBX

0x4270ae SAL	$0x3,%RBX

0x4270b2 XORPS	%XMM1,%XMM1

0x4270b5 MOV	$0x1,%R14D

0x4270bb NOPL	(%RAX,%RAX,1)

  (372) 0x4270c0 MOVSD	%XMM1,-0x58(%RBP)

  (372) 0x4270c5 MOV	0x200(%R15),%RAX

  (372) 0x4270cc ADD	%RBX,%RAX

  (372) 0x4270cf MOV	-0x8(%RAX,%R14,8),%RDI

  (372) 0x4270d4 MOV	-0x30(%RBP),%RAX

  (372) 0x4270d8 MOV	0x268(%RAX),%RAX

  (372) 0x4270df MOV	0x18(%RAX),%RAX

  (372) 0x4270e3 MOV	-0x4(%RAX,%R14,4),%EDX

  (372) 0x4270e8 MOV	(%RAX,%R14,4),%ECX

  (372) 0x4270ec MOV	0x1d0(%R15),%R9

  (372) 0x4270f3 MOV	%R12D,%ESI

  (372) 0x4270f6 MOV	%R13,%R8

  (372) 0x4270f9 CALL	424900 <_ZNK11qmcplusplus14BsplineFunctorIdE9evaluateVEiiiPKdPd>

  (372) 0x4270fe MOVSD	-0x58(%RBP),%XMM1

  (372) 0x427103 ADDSD	%XMM0,%XMM1

  (372) 0x427107 MOV	%R14D,%EAX

  (372) 0x42710a INC	%R14

  (372) 0x42710d CMP	%RAX,0xa0(%R15)

  (372) 0x427114 JA	4270c0

0x427116 JMP	427004

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/optional: 433 - 950

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

433:       { return static_cast<const _Dp*>(this)->_M_payload._M_engaged; }

[...]

950: 	if (this->_M_is_engaged())

/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h: 313 - 316

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

313:   inline int first(int igroup) const { return (*group_offsets_)[igroup]; }

314:

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

316:   inline int last(int igroup) const { return (*group_offsets_)[igroup + 1]; }

/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h: 223 - 249

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

223:     return X[i];

[...]

229:     return X[i];

[...]

249:   inline const_pointer data() const { return X; }

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/shared_ptr_base.h: 1296 - 1296

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

1296:       { return _M_ptr; }

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h: 919 - 1169

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

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

[...]

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

[...]

1064: 	return *(this->_M_impl._M_start + __n);

[...]

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

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/refwrap.h: 338 - 338

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

338:       { return *_M_data; }

/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h: 107 - 132

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

107:     for (int k = 0; k < ratios.size(); ++k)

108:       ratios[k] = std::exp(Uat[VP.refPtcl] - computeU(VP.getRefPS(), VP.refPtcl, VP.getDistTableAB(myTableID).getDistRow(k).data()));

[...]

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

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

128:     {

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

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

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

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

Path /

Metric	Value
CQA speedup if no scalar integer	2.39
CQA speedup if FP arith vectorized	1.88
CQA speedup if fully vectorized	5.38
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.16
Bottlenecks
Function	_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE
Source	optional:433-433,optional:950-950,OhmmsVector.h:223-223,OhmmsVector.h:229-229,OhmmsVector.h:249-249,stl_vector.h:919-919,stl_vector.h:1046-1046,stl_vector.h:1064-1064,refwrap.h:338-338,TwoBodyJastrowRef.h:107-108,TwoBodyJastrowRef.h:126-127
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	7.17
CQA cycles if no scalar integer	3.00
CQA cycles if FP arith vectorized	3.81
CQA cycles if fully vectorized	1.33
Front-end cycles	6.17
DIV/SQRT cycles	3.13
P0 cycles	3.13
P1 cycles	3.00
P2 cycles	3.00
P3 cycles	2.75
P4 cycles	7.17
P5 cycles	7.17
P6 cycles	7.17
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.50
P10 cycles	0.50
P11 cycles	1.00
P12 cycles	1.00
P13 cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	35.50
Nb uops	37.00
Nb loads	15.50
Nb stores	4.00
Nb stack references	5.00
FLOP/cycle	0.14
Nb FLOP add-sub	1.00
Nb FLOP mul	0.00
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	19.37
Bytes prefetched	0.00
Bytes loaded	107.00
Bytes stored	32.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	5.50
Stride indirect	0.00
Vectorization ratio all	7.74
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	22.50
Vector-efficiency ratio all	11.91
Vector-efficiency ratio load	8.63
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	12.23

Metric	Value
CQA speedup if no scalar integer	2.67
CQA speedup if FP arith vectorized	1.94
CQA speedup if fully vectorized	5.26
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.14
Bottlenecks	P5, P6, P7,
Function	_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE
Source	optional:433-433,optional:950-950,OhmmsVector.h:223-223,OhmmsVector.h:229-229,OhmmsVector.h:249-249,stl_vector.h:919-919,stl_vector.h:1046-1046,stl_vector.h:1064-1064,refwrap.h:338-338,TwoBodyJastrowRef.h:107-108,TwoBodyJastrowRef.h:126-127
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	8.00
CQA cycles if no scalar integer	3.00
CQA cycles if FP arith vectorized	4.13
CQA cycles if fully vectorized	1.52
Front-end cycles	7.00
DIV/SQRT cycles	3.75
P0 cycles	3.75
P1 cycles	3.75
P2 cycles	3.75
P3 cycles	3.00
P4 cycles	8.00
P5 cycles	8.00
P6 cycles	8.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.50
P10 cycles	0.50
P11 cycles	1.00
P12 cycles	1.00
P13 cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	41.00
Nb uops	42.00
Nb loads	18.00
Nb stores	4.00
Nb stack references	5.00
FLOP/cycle	0.13
Nb FLOP add-sub	1.00
Nb FLOP mul	0.00
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	19.63
Bytes prefetched	0.00
Bytes loaded	125.00
Bytes stored	32.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	8.00
Stride indirect	0.00
Vectorization ratio all	7.14
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	20.00
Vector-efficiency ratio all	11.72
Vector-efficiency ratio load	9.06
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	11.56

Metric	Value
CQA speedup if no scalar integer	2.11
CQA speedup if FP arith vectorized	1.81
CQA speedup if fully vectorized	5.53
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.19
Bottlenecks	P5, P6, P7,
Function	_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE
Source	optional:433-433,optional:950-950,OhmmsVector.h:223-223,OhmmsVector.h:229-229,OhmmsVector.h:249-249,stl_vector.h:919-919,stl_vector.h:1046-1046,stl_vector.h:1064-1064,refwrap.h:338-338,TwoBodyJastrowRef.h:107-108,TwoBodyJastrowRef.h:126-127
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	6.33
CQA cycles if no scalar integer	3.00
CQA cycles if FP arith vectorized	3.50
CQA cycles if fully vectorized	1.15
Front-end cycles	5.33
DIV/SQRT cycles	2.50
P0 cycles	2.50
P1 cycles	2.25
P2 cycles	2.25
P3 cycles	2.50
P4 cycles	6.33
P5 cycles	6.33
P6 cycles	6.33
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.50
P10 cycles	0.50
P11 cycles	1.00
P12 cycles	1.00
P13 cycles	0.00
Inter-iter dependencies cycles	0
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	30.00
Nb uops	32.00
Nb loads	13.00
Nb stores	4.00
Nb stack references	5.00
FLOP/cycle	0.16
Nb FLOP add-sub	1.00
Nb FLOP mul	0.00
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	19.11
Bytes prefetched	0.00
Bytes loaded	89.00
Bytes stored	32.00
Stride 0	2.00
Stride 1	0.00
Stride n	0.00
Stride unknown	3.00
Stride indirect	0.00
Vectorization ratio all	8.33
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	25.00
Vector-efficiency ratio all	12.11
Vector-efficiency ratio load	8.20
Vector-efficiency ratio store	12.50
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	12.50
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	12.89

Path /

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

Function	_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE
Source file and lines	TwoBodyJastrowRef.h:107-132
Module	exec

The loop is defined in:

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/optional:433,950
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:313-316
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,229,249
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/shared_ptr_base.h:1296
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h:919,1046,1064,1169
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/refwrap.h:338
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:107-108,126-132

Warnings:
Non-innermost loop: analyzing only self part (ignoring child loops).
The structure of this loop is probably <if then [else] end>.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Vectorization

Your loop is probably not vectorized. Only 11% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your loop, you can lower the cost of an iteration from 7.17 to 1.33 cycles (5.37x speedup).

Details

Store and arithmetical SSE/AVX 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 ffast-math (included in Ofast) to extend loop vectorization to 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

CALL instructions

Detected function call instructions.

Details

Calling (and then returning from) a function prevents many compiler optimizations (like vectorization), breaks control flow (which reduces pipeline performance) and executes extra instructions to save/restore the registers used inside it, which is very expensive (dozens of cycles). Consider to inline small functions.

_ZNK11qmcplusplus11ParticleSet14getDistTableABEi: 1 occurrences
exp: 1 occurrences

Type of elements and instruction set

1 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 1 FP arithmetical operations:

1: addition or subtraction

The binary loop is loading 107 bytes (13 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.01 FP operations per loaded or stored byte.

General properties

nb instructions	35.50
nb uops	37
loop length	169
used x86 registers	9.50
used mmx registers	0
used xmm registers	2
used ymm registers	0
used zmm registers	0
nb stack references	5

Front-end

MACRO FUSION NOT POSSIBLE

micro-operation queue	6.17 cycles
front end	6.17 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	3.13	3.13	3.00	3.00	2.75	7.17	7.17	7.17	0.00	0.00	0.50	0.50	1.00	1.00
cycles	3.13	3.13	3.00	3.00	2.75	7.17	7.17	7.17	0.00	0.00	0.50	0.50	1.00	1.00

Execution ports to units layout:

ALU0/BRU0: ALU
ALU1: ALU
ALU2: ALU
ALU3: ALU
BRU1:
AGU0 (256 bits): store address, load
AGU1 (256 bits): store address, load
AGU2 (256 bits): store address, load
FP0 (256 bits): VPU, DIV/SQRT
FP1 (256 bits): VPU, DIV/SQRT
FP2 (256 bits): VPU
FP3 (256 bits): VPU
FP4 (256 bits): FP store data
FP5 (256 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	6.17
Dispatch	7.17
Data deps.	0.00
Overall L1	7.17

Vectorization ratios

INT

all	0%
load	0%
store	0%
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	16%
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	100%

INT+FP

all	7%
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	22%

Vector efficiency ratios

INT

all	9%
load	5%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	8%

all	14%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	25%

INT+FP

all	11%
load	8%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 7.17 cycles. At this rate:

15% of peak load performance is reached (14.84 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
7% of peak store performance is reached (4.53 out of 64.00 bytes stored per cycle (GB/s @ 1GHz))

Function	_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE
Source file and lines	TwoBodyJastrowRef.h:107-132
Module	exec

The loop is defined in:

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/optional:433,950
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:313-316
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,229,249
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/shared_ptr_base.h:1296
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h:919,1046,1064,1169
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/refwrap.h:338
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:107-108,126-132

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

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 8.00 to 3.00 cycles (2.67x speedup).

Workaround

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

Vectorization

Details

Workaround

Try another compiler or update/tune your current one:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to 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)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 8.00 to 7.00 cycles (1.14x speedup).

Workaround

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

CALL instructions

Detected function call instructions.

Details

_ZNK11qmcplusplus11ParticleSet14getDistTableABEi: 1 occurrences
exp: 1 occurrences

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Constant unknown stride: 8 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

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

Type of elements and instruction set

1 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 1 FP arithmetical operations:

1: addition or subtraction

The binary loop is loading 125 bytes (15 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.01 FP operations per loaded or stored byte.

General properties

nb instructions	41
nb uops	42
loop length	193
used x86 registers	11
used mmx registers	0
used xmm registers	2
used ymm registers	0
used zmm registers	0
nb stack references	5

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.00 cycles
front end	7.00 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	3.75	3.75	3.75	3.75	3.00	8.00	8.00	8.00	0.00	0.00	0.50	0.50	1.00	1.00
cycles	3.75	3.75	3.75	3.75	3.00	8.00	8.00	8.00	0.00	0.00	0.50	0.50	1.00	1.00

Execution ports to units layout:

ALU0/BRU0: ALU
ALU1: ALU
ALU2: ALU
ALU3: ALU
BRU1:
AGU0 (256 bits): store address, load
AGU1 (256 bits): store address, load
AGU2 (256 bits): store address, load
FP0 (256 bits): VPU, DIV/SQRT
FP1 (256 bits): VPU, DIV/SQRT
FP2 (256 bits): VPU
FP3 (256 bits): VPU
FP4 (256 bits): FP store data
FP5 (256 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	7.00
Dispatch	8.00
Data deps.	0.00
Overall L1	8.00

Vectorization ratios

INT

all	0%
load	0%
store	0%
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	16%
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	100%

INT+FP

all	7%
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	20%

Vector efficiency ratios

INT

all	9%
load	6%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	8%

all	14%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	25%

INT+FP

all	11%
load	9%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	11%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.00 cycles. At this rate:

16% of peak load performance is reached (15.63 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
6% of peak store performance is reached (4.00 out of 64.00 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: 42703d

Instruction	Nb FU	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP2	FP3	FP4	FP5	Latency	Recip. throughput
MOVSD -0x48(%RBP),%XMM0	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50
SUBSD %XMM1,%XMM0	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50
CALL 4f1fd0 <exp>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0	0	0.50
MOV -0x40(%RBP),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV (%RCX),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV -0x50(%RBP),%RBX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOVSD %XMM0,(%RAX,%RBX,8)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0.50	0.50	1	1
INC %RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV 0x8(%RCX),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
SUB %RAX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
SAR $0x3,%RCX	1	0	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
CMP %RBX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV -0x38(%RBP),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
JBE 42711b <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x14b>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	1	0.50-1
CMPB $0,0x298(%RCX)	1	0.25	0.25	0.25	0.25	0	0.33	0.33	0.33	0	0	0	0	1	0.33
JE 42712a <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x15a>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	1	0.50-1
MOVSXD 0x2a0(%RCX),%R12	1	0.25	0.25	0.25	0.25	0	0.33	0.33	0.33	0	0	0	0	1	0.33
MOV 0xd8(%R15),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOVSD (%RAX,%R12,8),%XMM0	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50
MOVSD %XMM0,-0x48(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0.50	0.50	1	1
MOV 0x290(%RCX),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV %RAX,-0x30(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	4	0.50
MOV 0x248(%R15),%ESI	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV %RCX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17
CALL 462dd0 <_ZNK11qmcplusplus11ParticleSet14getDistTableABEi>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0	0	0.50
MOV 0xa0(%R15),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
TEST %RCX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV %RBX,-0x50(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	4	0.50
JE 427000 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x30>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	1	0.50-1
MOV 0x48(%RAX),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
LEA (%RBX,%RBX,4),%RDX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV 0x18(%RAX,%RDX,8),%R13	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV -0x30(%RBP),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV 0x18(%RAX),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
IMUL (%RAX,%R12,4),%ECX	1	0	1	0	0	0	0.33	0.33	0.33	0	0	0	0	3	1
MOVSXD %ECX,%RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
SAL $0x3,%RBX	1	0	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
XORPS %XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17
MOV $0x1,%R14D	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09
JMP 427004 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x34>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	0	1

Function	_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE
Source file and lines	TwoBodyJastrowRef.h:107-132
Module	exec

The loop is defined in:

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/optional:433,950
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Particle/ParticleSet.h:313-316
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Numerics/OhmmsPETE/OhmmsVector.h:223,229,249
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/shared_ptr_base.h:1296
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h:919,1046,1064,1169
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/refwrap.h:338
/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/QMCWaveFunctions/Jastrow/TwoBodyJastrowRef.h:107-108,126-132

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

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 6.33 to 3.00 cycles (2.11x speedup).

Workaround

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

Vectorization

Your loop is probably not vectorized. Only 12% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your loop, you can lower the cost of an iteration from 6.33 to 1.15 cycles (5.53x speedup).

Details

Workaround

Try another compiler or update/tune your current one:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to 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)

Execution units bottlenecks

Performance is limited by:

reading data from caches/RAM (load units are a bottleneck)
writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 6.33 to 5.33 cycles (1.19x speedup).

Workaround

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

CALL instructions

Detected function call instructions.

Details

_ZNK11qmcplusplus11ParticleSet14getDistTableABEi: 1 occurrences
exp: 1 occurrences

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Constant unknown stride: 3 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

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

Type of elements and instruction set

1 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).

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

The binary loop is composed of 1 FP arithmetical operations:

1: addition or subtraction

The binary loop is loading 89 bytes (11 double precision FP elements). The binary loop is storing 32 bytes (4 double precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.01 FP operations per loaded or stored byte.

General properties

nb instructions	30
nb uops	32
loop length	145
used x86 registers	8
used mmx registers	0
used xmm registers	2
used ymm registers	0
used zmm registers	0
nb stack references	5

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	5.33 cycles
front end	5.33 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	2.50	2.50	2.25	2.25	2.50	6.33	6.33	6.33	0.00	0.00	0.50	0.50	1.00	1.00
cycles	2.50	2.50	2.25	2.25	2.50	6.33	6.33	6.33	0.00	0.00	0.50	0.50	1.00	1.00

Execution ports to units layout:

ALU0/BRU0: ALU
ALU1: ALU
ALU2: ALU
ALU3: ALU
BRU1:
AGU0 (256 bits): store address, load
AGU1 (256 bits): store address, load
AGU2 (256 bits): store address, load
FP0 (256 bits): VPU, DIV/SQRT
FP1 (256 bits): VPU, DIV/SQRT
FP2 (256 bits): VPU
FP3 (256 bits): VPU
FP4 (256 bits): FP store data
FP5 (256 bits): FP store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	0.00

Cycles summary

Front-end	5.33
Dispatch	6.33
Data deps.	0.00
Overall L1	6.33

Vectorization ratios

INT

all	0%
load	0%
store	0%
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	16%
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	100%

INT+FP

all	8%
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	25%

Vector efficiency ratios

INT

all	9%
load	3%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
other	8%

all	14%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	25%

INT+FP

all	12%
load	8%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	12%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 6.33 cycles. At this rate:

14% of peak load performance is reached (14.05 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
7% of peak store performance is reached (5.05 out of 64.00 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: 42703d

Instruction	Nb FU	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP2	FP3	FP4	FP5	Latency	Recip. throughput
XORPD %XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17
MOVSD -0x48(%RBP),%XMM0	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50
SUBSD %XMM1,%XMM0	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	3	0.50
CALL 4f1fd0 <exp>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0	0	0.50
MOV -0x40(%RBP),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV (%RCX),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV -0x50(%RBP),%RBX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOVSD %XMM0,(%RAX,%RBX,8)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0.50	0.50	1	1
INC %RBX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV 0x8(%RCX),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
SUB %RAX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
SAR $0x3,%RCX	1	0	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50
CMP %RBX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV -0x38(%RBP),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
JBE 42711b <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x14b>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	1	0.50-1
CMPB $0,0x298(%RCX)	1	0.25	0.25	0.25	0.25	0	0.33	0.33	0.33	0	0	0	0	1	0.33
JE 42712a <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x15a>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	1	0.50-1
MOVSXD 0x2a0(%RCX),%R12	1	0.25	0.25	0.25	0.25	0	0.33	0.33	0.33	0	0	0	0	1	0.33
MOV 0xd8(%R15),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOVSD (%RAX,%R12,8),%XMM0	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	1	0.50
MOVSD %XMM0,-0x48(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0.50	0.50	1	1
MOV 0x290(%RCX),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV %RAX,-0x30(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	4	0.50
MOV 0x248(%R15),%ESI	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
MOV %RCX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17
CALL 462dd0 <_ZNK11qmcplusplus11ParticleSet14getDistTableABEi>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0	0	0	0	0.50
MOV 0xa0(%R15),%RCX	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	3	0.33
TEST %RCX,%RCX	1	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	1	0.25
MOV %RBX,-0x50(%RBP)	1	0	0	0	0	0	0.33	0.33	0.33	0	0	0	0	4	0.50
JE 427000 <_ZN16miniqmcreference17TwoBodyJastrowRefIN11qmcplusplus14BsplineFunctorIdEEE14evaluateRatiosERNS1_18VirtualParticleSetERSt6vectorIdSaIdEE+0x30>	1	0.50	0	0	0	0.50	0	0	0	0	0	0	0	1	0.50-1

Report Configuration

Vectorization

Details

Workaround

CALL instructions

Details

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

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

CALL instructions

Details

Slow data structures access

Details

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

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

Workaround

CALL instructions

Details

Slow data structures access

Details

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