OV - exec - Function _ZN11qmcplusplus11ParticleSet6updateEb.A

	Function: _ZN11qmcplusplus11ParticleSet6updateEb.A	Module: exec	Source: ParticleSet.cpp:238-245 [...]	Coverage: 0.02%

/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Utilities/NewTimer.h: 242 - 249

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

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

[...]

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

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h: 919 - 1046

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

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

[...]

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

/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/Particle/ParticleSet.cpp: 238 - 245

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

238: {

239:   ScopedTimer update_scope(myTimers[PS_update]);

240:

241:   coordinates_->setAllParticlePos(R);

242:   for (int i = 0; i < DistTables.size(); i++)

243:     DistTables[i]->evaluate(*this);

244:   active_ptcl_ = -1;

245: }

/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/unique_ptr.h: 173 - 173

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

173:       pointer    _M_ptr() const { return std::get<0>(_M_t); }

0x462e00 PUSH	%RBP

0x462e01 MOV	%RSP,%RBP

0x462e04 PUSH	%R15

0x462e06 PUSH	%R14

0x462e08 PUSH	%RBX

0x462e09 PUSH	%RAX

0x462e0a MOV	%RDI,%R14

0x462e0d MOV	0x210(%RDI),%RAX

0x462e14 MOV	0x20(%RAX),%RBX

0x462e18 MOV	%RBX,%RDI

0x462e1b CALL	497e00 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE5startEv>

0x462e20 MOV	0x278(%R14),%RDI

0x462e27 LEA	0x28(%R14),%RSI

0x462e2b MOV	(%RDI),%RAX

0x462e2e CALLQ	0x28(%RAX)

0x462e31 MOV	0x1e0(%R14),%RAX

0x462e38 CMP	%RAX,0x1e8(%R14)

0x462e3f JE	462e7a

0x462e41 XOR	%R15D,%R15D

0x462e44 NOPW	%CS:(%RAX,%RAX,1)

(1146) 0x462e50 MOV	(%RAX,%R15,8),%RDI

(1146) 0x462e54 MOV	(%RDI),%RAX

(1146) 0x462e57 MOV	%R14,%RSI

(1146) 0x462e5a CALLQ	0x10(%RAX)

(1146) 0x462e5d INC	%R15

(1146) 0x462e60 MOV	0x1e0(%R14),%RAX

(1146) 0x462e67 MOV	0x1e8(%R14),%RCX

(1146) 0x462e6e SUB	%RAX,%RCX

(1146) 0x462e71 SAR	$0x3,%RCX

(1146) 0x462e75 CMP	%R15,%RCX

(1146) 0x462e78 JA	462e50

0x462e7a MOVL	$-0x1,0x124(%R14)

0x462e85 MOV	%RBX,%RDI

0x462e88 CALL	497fd0 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv>

0x462e8d ADD	$0x8,%RSP

0x462e91 POP	%RBX

0x462e92 POP	%R14

0x462e94 POP	%R15

0x462e96 POP	%RBP

0x462e97 RET

0x462e98 MOV	%RAX,%RDI

0x462e9b CALL	410010 <__clang_call_terminate>

0x462ea0 JMP	462ea2

0x462ea2 MOV	%RAX,%R14

0x462ea5 MOV	%RBX,%RDI

0x462ea8 CALL	497fd0 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv>

0x462ead MOV	%R14,%RDI

0x462eb0 CALL	4053b0 <_Unwind_Resume@plt>

0x462eb5 MOV	%RAX,%RDI

0x462eb8 CALL	410010 <__clang_call_terminate>

0x462ebd NOPL	(%RAX)

Path /

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

Source file and lines	ParticleSet.cpp:238-245
Module	exec

The function is defined in:

/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Utilities/NewTimer.h:242,249
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h:919,1046
/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/Particle/ParticleSet.cpp:238-245
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/unique_ptr.h:173

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 10% of vector register length is used (average across all SSE/AVX instructions). By vectorizing your function, you can lower the cost of an iteration from 7.83 to 0.81 cycles (9.64x speedup).

Details

All 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 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)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

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.

_Unwind_Resume@plt: 1 occurrences
_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv: 2 occurrences
_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE5startEv: 1 occurrences
__clang_call_terminate: 2 occurrences
unknown: 1 occurrences

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This function is probably writing/copying data or processing integer elements.

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 56 bytes. The binary function is storing 4 bytes.

General properties

nb instructions	40
nb uops	47
loop length	150
used x86 registers	8
used mmx registers	0
used xmm registers	0
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.83 cycles
front end	7.83 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	5.00	1.00	1.00	1.00	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	0.00	0.00
cycles	5.00	1.00	1.00	1.00	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	0.00	0.00

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

Cycles summary

Front-end	7.83
Dispatch	5.00
Overall L1	7.83

Vectorization ratios

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)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

all	10%
load	12%
store	6%
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	10%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 7.83 cycles. At this rate:

7% of peak load performance is reached (7.15 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (0.51 out of 64.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Performance is limited by instruction throughput (loading/decoding program instructions to execution core) (front-end is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 7.83 to 5.00 cycles (1.57x speedup).

ASM code

In the binary file, the address of the function is: 462e00

Instruction	Nb FU	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	Latency	Recip. throughput
PUSH %RBP	1	0	0	0	0	0	0	0	0	1	0.50
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0.17
PUSH %R15	1	0	0	0	0	0	0	0	0	1	0.50
PUSH %R14	1	0	0	0	0	0	0	0	0	1	0.50
PUSH %RBX	1	0	0	0	0	0	0	0	0	1	0.50
PUSH %RAX	1	0	0	0	0	0	0	0	0	1	0.50
MOV %RDI,%R14	1	0	0	0	0	0	0	0	0	0	0.17
MOV 0x210(%RDI),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
MOV 0x20(%RAX),%RBX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 497e00 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE5startEv>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
MOV 0x278(%R14),%RDI	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
LEA 0x28(%R14),%RSI	1	0.25	0.25	0.25	0.25	0	0	0	0	1	0.25
MOV (%RDI),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
CALLQ 0x28(%RAX)	4	0.50	0	0	0	0.50	0.67	0.67	0.67	0	2.20
MOV 0x1e0(%R14),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
CMP %RAX,0x1e8(%R14)	1	0.25	0.25	0.25	0.25	0	0.33	0.33	0.33	1	0.33
JE 462e7a <_ZN11qmcplusplus11ParticleSet6updateEb.A+0x7a>	1	0.50	0	0	0	0.50	0	0	0	1	0.50-1
XOR %R15D,%R15D	1	0	0	0	0	0	0	0	0	0	0.25
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0.09
MOVL $-0x1,0x124(%R14)	1	0	0	0	0	0	0.33	0.33	0.33	4	0.50
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 497fd0 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
ADD $0x8,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	1	0.25
POP %RBX	1	0	0	0	0	0	0	0	0	1	0.33
POP %R14	1	0	0	0	0	0	0	0	0	1	0.33
POP %R15	1	0	0	0	0	0	0	0	0	1	0.33
POP %RBP	1	0	0	0	0	0	0	0	0	1	0.33
RET	1	0.50	0	0	0	0.50	0	0	0	1	0.50
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 410010 <__clang_call_terminate>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
JMP 462ea2 <_ZN11qmcplusplus11ParticleSet6updateEb.A+0xa2>	1	0.50	0	0	0	0.50	0	0	0	0	1
MOV %RAX,%R14	1	0	0	0	0	0	0	0	0	0	0.17
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 497fd0 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
MOV %R14,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 4053b0 <_Unwind_Resume@plt>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 410010 <__clang_call_terminate>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0.09

Source file and lines	ParticleSet.cpp:238-245
Module	exec

The function is defined in:

/beegfs/hackathon/users/eoseret/qaas_runs/170-855-3059/intel/miniqmc/build/miniqmc/src/Utilities/NewTimer.h:242,249
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_vector.h:919,1046
/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/Particle/ParticleSet.cpp:238-245
/usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/unique_ptr.h:173

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Details

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)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

No data for this section

CALL instructions

Detected function call instructions.

Details

_Unwind_Resume@plt: 1 occurrences
_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv: 2 occurrences
_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE5startEv: 1 occurrences
__clang_call_terminate: 2 occurrences
unknown: 1 occurrences

Type of elements and instruction set

No instructions are processing arithmetic or math operations on FP elements. This function is probably writing/copying data or processing integer elements.

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

The binary function does not contain any FP arithmetical operations. The binary function is loading 56 bytes. The binary function is storing 4 bytes.

General properties

nb instructions	40
nb uops	47
loop length	150
used x86 registers	8
used mmx registers	0
used xmm registers	0
used ymm registers	0
used zmm registers	0
nb stack references	0

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.83 cycles
front end	7.83 cycles

Back-end

	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	FP0	FP1	FP2	FP3	FP4	FP5
uops	5.00	1.00	1.00	1.00	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	0.00	0.00
cycles	5.00	1.00	1.00	1.00	5.00	5.00	5.00	5.00	0.00	0.00	0.00	0.00	0.00	0.00

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

Cycles summary

Front-end	7.83
Dispatch	5.00
Overall L1	7.83

Vectorization ratios

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)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

all	10%
load	12%
store	6%
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	10%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 7.83 cycles. At this rate:

7% of peak load performance is reached (7.15 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
0% of peak store performance is reached (0.51 out of 64.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the function is: 462e00

Instruction	Nb FU	ALU0/BRU0	ALU1	ALU2	ALU3	BRU1	AGU0	AGU1	AGU2	Latency	Recip. throughput
PUSH %RBP	1	0	0	0	0	0	0	0	0	1	0.50
MOV %RSP,%RBP	1	0	0	0	0	0	0	0	0	0	0.17
PUSH %R15	1	0	0	0	0	0	0	0	0	1	0.50
PUSH %R14	1	0	0	0	0	0	0	0	0	1	0.50
PUSH %RBX	1	0	0	0	0	0	0	0	0	1	0.50
PUSH %RAX	1	0	0	0	0	0	0	0	0	1	0.50
MOV %RDI,%R14	1	0	0	0	0	0	0	0	0	0	0.17
MOV 0x210(%RDI),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
MOV 0x20(%RAX),%RBX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 497e00 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE5startEv>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
MOV 0x278(%R14),%RDI	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
LEA 0x28(%R14),%RSI	1	0.25	0.25	0.25	0.25	0	0	0	0	1	0.25
MOV (%RDI),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
CALLQ 0x28(%RAX)	4	0.50	0	0	0	0.50	0.67	0.67	0.67	0	2.20
MOV 0x1e0(%R14),%RAX	1	0	0	0	0	0	0.33	0.33	0.33	3	0.33
CMP %RAX,0x1e8(%R14)	1	0.25	0.25	0.25	0.25	0	0.33	0.33	0.33	1	0.33
JE 462e7a <_ZN11qmcplusplus11ParticleSet6updateEb.A+0x7a>	1	0.50	0	0	0	0.50	0	0	0	1	0.50-1
XOR %R15D,%R15D	1	0	0	0	0	0	0	0	0	0	0.25
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0.09
MOVL $-0x1,0x124(%R14)	1	0	0	0	0	0	0.33	0.33	0.33	4	0.50
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 497fd0 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
ADD $0x8,%RSP	1	0.25	0.25	0.25	0.25	0	0	0	0	1	0.25
POP %RBX	1	0	0	0	0	0	0	0	0	1	0.33
POP %R14	1	0	0	0	0	0	0	0	0	1	0.33
POP %R15	1	0	0	0	0	0	0	0	0	1	0.33
POP %RBP	1	0	0	0	0	0	0	0	0	1	0.33
RET	1	0.50	0	0	0	0.50	0	0	0	1	0.50
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 410010 <__clang_call_terminate>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
JMP 462ea2 <_ZN11qmcplusplus11ParticleSet6updateEb.A+0xa2>	1	0.50	0	0	0	0.50	0	0	0	0	1
MOV %RAX,%R14	1	0	0	0	0	0	0	0	0	0	0.17
MOV %RBX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 497fd0 <_ZN11qmcplusplus9TimerTypeINSt6chrono3_V212system_clockEE4stopEv>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
MOV %R14,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 4053b0 <_Unwind_Resume@plt>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0.17
CALL 410010 <__clang_call_terminate>	2	0.50	0	0	0	0.50	0.33	0.33	0.33	0	0.50
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0.09

Name	Coverage (%)	Time (s)
▼_ZN11qmcplusplus11ParticleSet6updateEb.A–	0.02	0.03
○Loop 1146 - ParticleSet.cpp:242-243 - exec	0.01	0.02

Report Configuration

Vectorization

Details

Workaround

Execution units bottlenecks

CALL instructions

Details

Type of elements and instruction set

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

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

CALL instructions

Details

Type of elements and instruction set

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code