OV - exec - Function kineticEnergy._omp

	Function: kineticEnergy._omp_fn.0	Module: exec	Source: timestep.c:107-116	Coverage (incl. loops): 0.04% \| (excl. loops): 0.00%

/home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c: 107 - 116

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

107:    #pragma omp parallel for reduction(+:kenergy)

108:    for (int iBox=0; iBox<s->boxes->nLocalBoxes; iBox++)

109:    {

110:       for (int iOff=MAXATOMS*iBox,ii=0; ii<s->boxes->nAtoms[iBox]; ii++,iOff++)

111:       {

112:          int iSpecies = s->atoms->iSpecies[iOff];

113:          real_t invMass = 0.5/s->species[iSpecies].mass;

114:          kenergy += ( s->atoms->p[iOff][0] * s->atoms->p[iOff][0] +

115:          s->atoms->p[iOff][1] * s->atoms->p[iOff][1] +

116:          s->atoms->p[iOff][2] * s->atoms->p[iOff][2] )*invMass;

0x41aae0 STP	X29, X30, [SP, #976]!

0x41aae4 ADD	X29, SP, #0

0x41aae8 STP	X19, X20, [SP, #16]

0x41aaec LDR	X19, [X0]

0x41aaf0 STR	X21, [SP, #32]

0x41aaf4 ORR	X21, XZR, X0

0x41aaf8 BL	410050

0x41aafc ORR	W20, WZR, W0

0x41ab00 BL	410150

0x41ab04 LDR	X2, [X19, #24]

0x41ab08 ORR	W9, WZR, W0

0x41ab0c LDR	W0, [X2, #12]

0x41ab10 SDIV	W10, W0, W20

0x41ab14 MSUB	W1, W10, W20, W0

0x41ab18 CMP	W9, W1

0x41ab1c B.LT	41ac0c

0x41ab20 MADD	W3, W10, W9, W1

0x41ab24 MOVI	D5, #0

0x41ab28 ADD	W12, W10, W3

0x41ab2c CMP	W3, W12

0x41ab30 B.GE	41abd8

0x41ab34 MOVZ	W4, #1536

0x41ab38 LDR	X11, [X2, #120]

0x41ab3c SBFM	X8, X3, #0, #31

0x41ab40 CNTD	X7, ALL

0x41ab44 CNTB	X6, ALL, MUL #3

0x41ab48 PTRUE	P0.B, ALL

0x41ab4c SMULL	X9, W3, W4

0x41ab50 FDUP	Z6.D, #96

0x41ab54 HINT	#0

0x41ab58 HINT	#0

0x41ab5c HINT	#0

(108) 0x41ab60 LDR	W13, [X11, X8,LSL #2]

(108) 0x41ab64 CMP	W13, #0

(108) 0x41ab68 B.LE	41abc8

(108) 0x41ab6c LDP	X14, X5, [X19, #32]

(108) 0x41ab70 MOVZ	X30, #0

(108) 0x41ab74 WHILELO	P1.D, WZR, W13

(108) 0x41ab78 LDR	X16, [X14, #16]

(108) 0x41ab7c ADD	X15, X5, #8

(108) 0x41ab80 LDR	X17, [X14, #32]

(108) 0x41ab84 ADD	X18, X16, X8,LSL #8

(108) 0x41ab88 ADD	X20, X17, X9

  (109) 0x41ab8c LD3D	{Z1.D, Z2.D, Z3.D}, P1/Z, [X20, MUL VL]

  (109) 0x41ab90 LD1SW	{Z0.D}, P1/Z, [X18, X30,LSL #2]

  (109) 0x41ab94 FMUL	Z2.D, Z2.D, Z2.D

  (109) 0x41ab98 ADD	X30, X30, X7

  (109) 0x41ab9c LSL	Z4.D, Z0.D, #4

  (109) 0x41aba0 FMLA	Z2.D, P0/M, Z1.D, Z1.D

  (109) 0x41aba4 LD1D	{Z16.D}, P1/Z, [X15, Z4.D]

  (109) 0x41aba8 MOVPRFX	Z1, Z2

  (109) 0x41abac FMLA	Z1.D, P0/M, Z3.D, Z3.D

  (109) 0x41abb0 FDIVR	Z16.D, P0/M, Z16.D, Z6.D

  (109) 0x41abb4 ADD	X20, X20, X6

  (109) 0x41abb8 FMUL	Z7.D, Z16.D, Z1.D

  (109) 0x41abbc FADDA	D5, P1, D5, Z7.D

  (109) 0x41abc0 WHILELO	P1.D, W30, W13

  (109) 0x41abc4 B.NE	41ab8c

(108) 0x41abc8 ADD	X8, X8, #1

(108) 0x41abcc ADD	X9, X9, #1536

(108) 0x41abd0 CMP	W12, W8

(108) 0x41abd4 B.GT	41ab60

0x41abd8 ADD	X2, X21, #8

0x41abdc LDR	X0, [X2]

(107) 0x41abe0 FMOV	D17, X0

(107) 0x41abe4 ORR	X21, XZR, X0

(107) 0x41abe8 FADD	D18, D5, D17

(107) 0x41abec FMOV	X19, D18

(107) 0x41abf0 CASAL	X21, X19, [X2]

(107) 0x41abf4 CMP	X0, X21

(107) 0x41abf8 B.NE	41ac18

0x41abfc LDP	X19, X20, [SP, #16]

0x41ac00 LDR	X21, [SP, #32]

0x41ac04 LDP	X29, X30, [SP], #48

0x41ac08 RET

0x41ac0c ADD	W10, W10, #1

0x41ac10 MOVZ	W1, #0

0x41ac14 B	41ab20

(107) 0x41ac18 ORR	X0, XZR, X21

(107) 0x41ac1c B	41abe0

Coverage (%)	Name	Source Location	Module
►77.36+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►7.55+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:102	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►7.55+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►7.55+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	setTemperature	initAtoms.c:171	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►50.00+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►38.46+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►3.85+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►3.85+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:102	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►3.85+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	setTemperature	initAtoms.c:171	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►74.07+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►18.52+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►3.70+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:102	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►1.85+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	setTemperature	initAtoms.c:171	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►1.85+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►94.64+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►3.57+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►1.79+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:102	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►93.24+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►4.05+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►1.35+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►1.35+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:102	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►97.10+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►2.90+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►98.57+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.43+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►96.70+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►3.30+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Source Location	Module
►97.33+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.33+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►1.33+	GOMP_parallel		libgomp.so.1.0.0
○	kineticEnergy	timestep.c:107	exec
○	main	CoMD.c:200	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

Coverage (%)	Name	Module
►99.14+	omp_fulfill_event	libgomp.so.1.0.0
○	start_thread	libc.so.6
○	thread_start	libc.so.6

Coverage (%)	Name	Source Location	Module
►98.53+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.47+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:107	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec

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 /

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

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run 1x1

Source file and lines	timestep.c:107-116
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:107-116.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 28% of vector register length is used (average across all VPU instructions).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers).

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 12.50 to 5.25 cycles (2.38x speedup).

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.

unknown: 2 occurrences►
- /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:107
- /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:107

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 3 bytes.

General properties

nb instructions	41
nb uops	38
loop length	164
used w registers	9
used x registers	14
used b registers	0
used h registers	0
used s registers	1
used d registers	1
used q registers	0
used v registers	0
used z registers	1
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	4.75 cycles
front end	4.75 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.00	3.00	5.50	5.25	6.00	5.25	0.50	0.50	0.50	0.50	3.83	3.50	3.67	1.50	1.50
cycles	3.00	3.00	5.50	5.25	6.00	5.25	0.50	0.50	0.50	0.50	3.83	3.50	3.67	1.50	1.50

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P7 (128 bits): VPU, ALU, FP store data
P8 (128 bits): VPU, ALU, DIV/SQRT
P9 (128 bits): ALU, VPU
P10 (256 bits): store address, load
P11 (256 bits): store address, load
P12 (256 bits): load
P13 (64 bits): store data
P14 (64 bits): store data

Cycles executing div or sqrt instructions

5.00-12.50

Cycles summary

Front-end	4.75
Dispatch	6.00
DIV/SQRT	5.00-12.50
Overall L1	6.00-12.50

Vectorization ratios

INT

all	0%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
other	0%

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

INT+FP

all	0%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	29%
load	33%
store	41%
mul	12%
add-sub	18%
fma	12%
other	31%

all	3%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	3%

INT+FP

all	28%
load	33%
store	41%
mul	12%
add-sub	18%
fma	12%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	29%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 12.50 cycles. At this rate:

0% of peak load performance is reached (0.24 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 41aae0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
STP X29, X30, [SP, #976]!	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ADD X29, SP, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (25.0%)
STP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDR X19, [X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
STR X21, [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	scal (25.0%)
ORR X21, 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	scal (25.0%)
BL 410050 <@plt_start@+0x30>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR W20, WZR, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
BL 410150 <@plt_start@+0x130>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X2, [X19, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ORR W9, WZR, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
LDR W0, [X2, #12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
SDIV W10, W0, W20	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
MSUB W1, W10, W20, W0	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
CMP W9, W1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.LT 41ac0c <kineticEnergy._omp_fn.0+0x12c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD W3, W10, W9, W1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
MOVI D5, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25	scal (25.0%)
ADD W12, W10, W3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMP W3, W12	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.GE 41abd8 <kineticEnergy._omp_fn.0+0xf8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOVZ W4, #1536	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
LDR X11, [X2, #120]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
SBFM X8, X3, #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	scal (100.0%)
CNTD X7, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
CNTB X6, ALL, MUL #3	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
PTRUE P0.B, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (100.0%)
SMULL X9, W3, W4	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50	scal (12.5%)
FDUP Z6.D, #96	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50	scal (3.1%)
HINT #0																			N/A
HINT #0																			N/A
HINT #0																			N/A
ADD X2, X21, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
LDR X0, [X2]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	N/A
LDR X21, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
LDP X29, X30, [SP], #48	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W10, W10, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOVZ W1, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
B 41ab20 <kineticEnergy._omp_fn.0+0x40>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run 1x1

Source file and lines	timestep.c:107-116
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:107-116.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 28% of vector register length is used (average across all VPU instructions).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers).

Expensive FP math instructions/calls

Detected performance impact from expensive FP math instructions/calls. By removing/reexpressing them, you can lower the cost of an iteration from 12.50 to 5.25 cycles (2.38x speedup).

CALL instructions

Detected function call instructions.

Details

unknown: 2 occurrences►
- /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:107
- /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:107

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 3 bytes.

General properties

nb instructions	41
nb uops	38
loop length	164
used w registers	9
used x registers	14
used b registers	0
used h registers	0
used s registers	1
used d registers	1
used q registers	0
used v registers	0
used z registers	1
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	4.75 cycles
front end	4.75 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.00	3.00	5.50	5.25	6.00	5.25	0.50	0.50	0.50	0.50	3.83	3.50	3.67	1.50	1.50
cycles	3.00	3.00	5.50	5.25	6.00	5.25	0.50	0.50	0.50	0.50	3.83	3.50	3.67	1.50	1.50

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P7 (128 bits): VPU, ALU, FP store data
P8 (128 bits): VPU, ALU, DIV/SQRT
P9 (128 bits): ALU, VPU
P10 (256 bits): store address, load
P11 (256 bits): store address, load
P12 (256 bits): load
P13 (64 bits): store data
P14 (64 bits): store data

Cycles executing div or sqrt instructions

5.00-12.50

Cycles summary

Front-end	4.75
Dispatch	6.00
DIV/SQRT	5.00-12.50
Overall L1	6.00-12.50

Vectorization ratios

INT

all	0%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
other	0%

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

INT+FP

all	0%
load	0%
store	0%
mul	0%
add-sub	0%
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	29%
load	33%
store	41%
mul	12%
add-sub	18%
fma	12%
other	31%

all	3%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	NA (no add-sub vectorizable/vectorized instructions)
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	3%

INT+FP

all	28%
load	33%
store	41%
mul	12%
add-sub	18%
fma	12%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	29%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 12.50 cycles. At this rate:

0% of peak load performance is reached (0.24 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 41aae0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
STP X29, X30, [SP, #976]!	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ADD X29, SP, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (25.0%)
STP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDR X19, [X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
STR X21, [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	scal (25.0%)
ORR X21, 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	scal (25.0%)
BL 410050 <@plt_start@+0x30>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR W20, WZR, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
BL 410150 <@plt_start@+0x130>	1	0.50	0.50	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X2, [X19, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ORR W9, WZR, W0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
LDR W0, [X2, #12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
SDIV W10, W0, W20	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
MSUB W1, W10, W20, W0	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
CMP W9, W1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.LT 41ac0c <kineticEnergy._omp_fn.0+0x12c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD W3, W10, W9, W1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
MOVI D5, #0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25	scal (25.0%)
ADD W12, W10, W3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
CMP W3, W12	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (12.5%)
B.GE 41abd8 <kineticEnergy._omp_fn.0+0xf8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MOVZ W4, #1536	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
LDR X11, [X2, #120]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
SBFM X8, X3, #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	scal (100.0%)
CNTD X7, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
CNTB X6, ALL, MUL #3	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
PTRUE P0.B, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (100.0%)
SMULL X9, W3, W4	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50	scal (12.5%)
FDUP Z6.D, #96	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50	scal (3.1%)
HINT #0																			N/A
HINT #0																			N/A
HINT #0																			N/A
ADD X2, X21, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
LDR X0, [X2]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDP X19, X20, [SP, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	N/A
LDR X21, [SP, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (25.0%)
LDP X29, X30, [SP], #48	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
RET	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD W10, W10, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	N/A
MOVZ W1, #0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25	scal (12.5%)
B 41ab20 <kineticEnergy._omp_fn.0+0x40>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

▶Scalability Runs Description

Run 1x1	Number processes: 1Number nodes: NARun Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_NUM_THREADS: 1OMP_PLACES: threads
Run 1x2	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 2OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x4	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 4OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x8	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 8OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x16	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 16OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x24	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 24OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x32	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 32OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x40	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 40OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x48	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 48OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x56	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 56OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x64	Number processes: 1Run Command: <executable> -x 100 -y 100 -z 100MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=64 --rank-by fill --report-bindings Dataset: Run Directory: /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/run/oneview_runs/multicore/gcc/oneview_run_1781767694OMP_NUM_THREADS: 64OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads

(1x1) Efficiency	(1x1) Potential Speed-Up (%)	(1x2) Efficiency	(1x2) Potential Speed-Up (%)	(1x4) Efficiency	(1x4) Potential Speed-Up (%)	(1x8) Efficiency	(1x8) Potential Speed-Up (%)	(1x16) Efficiency	(1x16) Potential Speed-Up (%)	(1x24) Efficiency	(1x24) Potential Speed-Up (%)	(1x32) Efficiency	(1x32) Potential Speed-Up (%)	(1x40) Efficiency	(1x40) Potential Speed-Up (%)	(1x48) Efficiency	(1x48) Potential Speed-Up (%)	(1x56) Efficiency	(1x56) Potential Speed-Up (%)	(1x64) Efficiency	(1x64) Potential Speed-Up (%)
1	0	1.01	0	0.96	0	0.89	0	0.64	0.02	0.66	0.02	0.65	0.02	0.48	0.03	0.57	0.02	0.35	0.05	0.29	0.07

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency

Run	Number of threads	Efficiency (ideal is 1)	Speedup	Ideal Speedup	Time (s)	Coverage (%)
1x1	1	1	1	1	0.26500001549721	0.040639601647854
1x2	2	1.01	2.02	2	0.12999999523163	0.03980628028512
1x4	4	0.96	3.84	4	0.070000000298023	0.041286718100309
1x8	8	0.89	7.16	8	0.050000000745058	0.042443331331015
1x16	15	0.64	10.2	16	0.035000000149012	0.055664967745543
1x24	24	0.66	15.78	24	0.025000000372529	0.051440037786961
1x32	31	0.65	20.81	32	0.025000000372529	0.051882304251194
1x40	39	0.48	19.27	40	0.025000000372529	0.066505409777164
1x48	43	0.57	27.27	48	0.024999998509884	0.054324556142092
1x56	51	0.35	19.79	56	0.025000002235174	0.083710916340351
1x64	63	0.29	18.55	64	0.024999998509884	0.097435064613819

Name	Coverage (%)	Time (s)
▼kineticEnergy._omp_fn.0–	0.04	0.26
▼Loop 108 - timestep.c:110-116 - exec–	0.00	0.00
○Loop 109 - timestep.c:110-116 - exec	0.04	0.27
○Loop 107 - timestep.c:107-107 - exec	0.00	0.00

Report Configuration

Vectorization

Details

Expensive FP math instructions/calls

CALL instructions

Details

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

Expensive FP math instructions/calls

CALL instructions

Details

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

▶Scalability Runs Description

▼Columns Filter

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency