OV - exec - Function advancePosition._omp

	Function: advancePosition._omp_fn.0	Module: exec	Source: timestep.c:85-94	Coverage (incl. loops): 0.42% \| (excl. loops): 0.00%

/home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c: 85 - 94

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

85:    #pragma omp parallel for

86:    for (int iBox=0; iBox<nBoxes; iBox++)

87:    {

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

89:       {

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

91:          real_t invMass = 1.0/s->species[iSpecies].mass;

92:          s->atoms->r[iOff][0] += dt*s->atoms->p[iOff][0]*invMass;

93:          s->atoms->r[iOff][1] += dt*s->atoms->p[iOff][1]*invMass;

94:          s->atoms->r[iOff][2] += dt*s->atoms->p[iOff][2]*invMass;

0x41a460 STP	X29, X30, [SP, #976]!

0x41a464 ADD	X29, SP, #0

0x41a468 STR	X21, [SP, #32]

0x41a46c LDR	W21, [X0, #16]

0x41a470 STP	X19, X20, [SP, #16]

0x41a474 ORR	X20, XZR, X0

0x41a478 BL	410050

0x41a47c ORR	W19, WZR, W0

0x41a480 BL	410150

0x41a484 SDIV	W1, W21, W19

0x41a488 ORR	W14, WZR, W0

0x41a48c MSUB	W2, W1, W19, W21

0x41a490 CMP	W0, W2

0x41a494 B.LT	41a650

0x41a498 MADD	W3, W1, W14, W2

0x41a49c ADD	W16, W1, W3

0x41a4a0 CMP	W3, W16

0x41a4a4 B.GE	41a640

0x41a4a8 LDR	X17, [X20]

0x41a4ac UBFM	W15, W3, #26, #25

0x41a4b0 MOVZ	W0, #24

0x41a4b4 FMOV	D4, #1.0000000

0x41a4b8 SBFM	X14, X3, #0, #31

0x41a4bc LDR	D0, [X20, #8]

0x41a4c0 SMULL	X15, W15, W0

0x41a4c4 LDR	X4, [X17, #24]

0x41a4c8 LDR	X18, [X4, #120]

(102) 0x41a4cc LDR	W5, [X18, X14,LSL #2]

(102) 0x41a4d0 CMP	W5, #0

(102) 0x41a4d4 B.LE	41a630

(102) 0x41a4d8 LDP	X10, X13, [X17, #32]

(102) 0x41a4dc SUB	W9, W5, #1

(102) 0x41a4e0 UBFM	X7, X14, #56, #55

(102) 0x41a4e4 ADD	X8, X9, X14,LSL #6

(102) 0x41a4e8 ORR	X0, XZR, X15

(102) 0x41a4ec LDP	X6, X30, [X10, #16]

(102) 0x41a4f0 LDR	X20, [X10, #32]

(102) 0x41a4f4 ADD	X11, X6, #4

(102) 0x41a4f8 ADD	X3, X6, X7

(102) 0x41a4fc ADD	X21, X11, X8,LSL #2

(102) 0x41a500 ADD	X19, X30, #8

(102) 0x41a504 ADD	X2, X30, #16

(102) 0x41a508 SUB	X1, X21, X3

(102) 0x41a50c ADD	X12, X20, #8

(102) 0x41a510 ADD	X4, X20, #16

(102) 0x41a514 TBZ	W1, #2, 41a580

(102) 0x41a518 LDRSW	X5, [X6, X7]

(102) 0x41a51c ADD	X3, X3, #4

(102) 0x41a520 ADD	X0, X15, #24

(102) 0x41a524 LDR	D3, [X20, X15]

(102) 0x41a528 LDR	D2, [X30, X15]

(102) 0x41a52c ADD	X9, X13, X5,LSL #4

(102) 0x41a530 LDR	D1, [X9, #8]

(102) 0x41a534 FMUL	D5, D0, D3

(102) 0x41a538 FDIV	D6, D4, D1

(102) 0x41a53c FMADD	D7, D6, D5, D2

(102) 0x41a540 STR	D7, [X30, X15]

(102) 0x41a544 LDR	D16, [X12, X15]

(102) 0x41a548 LDR	D17, [X19, X15]

(102) 0x41a54c FMUL	D18, D0, D16

(102) 0x41a550 FMADD	D19, D6, D18, D17

(102) 0x41a554 STR	D19, [X19, X15]

(102) 0x41a558 LDR	D20, [X4, X15]

(102) 0x41a55c LDR	D21, [X2, X15]

(102) 0x41a560 FMUL	D22, D0, D20

(102) 0x41a564 FMADD	D23, D6, D22, D21

(102) 0x41a568 STR	D23, [X2, X15]

(102) 0x41a56c CMP	X3, X21

(102) 0x41a570 B.EQ	41a630

(102) 0x41a574 HINT	#0

(102) 0x41a578 HINT	#0

(102) 0x41a57c HINT	#0

  (103) 0x41a580 ORR	X8, XZR, X3

  (103) 0x41a584 LDR	D24, [X20, X0]

  (103) 0x41a588 ADD	X10, X0, #24

  (103) 0x41a58c LDRSW	X7, [X8], #4

  (103) 0x41a590 LDR	D25, [X30, X0]

  (103) 0x41a594 FMUL	D26, D0, D24

  (103) 0x41a598 LDRSW	X6, [X3, #4]

  (103) 0x41a59c ADD	X11, X13, X7,LSL #4

  (103) 0x41a5a0 ADD	X3, X8, #4

  (103) 0x41a5a4 LDR	D27, [X11, #8]

  (103) 0x41a5a8 ADD	X1, X13, X6,LSL #4

  (103) 0x41a5ac FDIV	D28, D4, D27

  (103) 0x41a5b0 FMADD	D29, D28, D26, D25

  (103) 0x41a5b4 STR	D29, [X30, X0]

  (103) 0x41a5b8 LDR	D30, [X12, X0]

  (103) 0x41a5bc LDR	D31, [X19, X0]

  (103) 0x41a5c0 FMUL	D3, D0, D30

  (103) 0x41a5c4 FMADD	D2, D28, D3, D31

  (103) 0x41a5c8 STR	D2, [X19, X0]

  (103) 0x41a5cc LDR	D5, [X4, X0]

  (103) 0x41a5d0 LDR	D1, [X2, X0]

  (103) 0x41a5d4 FMUL	D6, D0, D5

  (103) 0x41a5d8 FMADD	D7, D28, D6, D1

  (103) 0x41a5dc STR	D7, [X2, X0]

  (103) 0x41a5e0 ADD	X0, X0, #48

  (103) 0x41a5e4 LDR	D19, [X1, #8]

  (103) 0x41a5e8 LDR	D16, [X20, X10]

  (103) 0x41a5ec LDR	D17, [X30, X10]

  (103) 0x41a5f0 FDIV	D20, D4, D19

  (103) 0x41a5f4 FMUL	D18, D0, D16

  (103) 0x41a5f8 FMADD	D21, D20, D18, D17

  (103) 0x41a5fc STR	D21, [X30, X10]

  (103) 0x41a600 LDR	D22, [X12, X10]

  (103) 0x41a604 LDR	D23, [X19, X10]

  (103) 0x41a608 FMUL	D24, D0, D22

  (103) 0x41a60c FMADD	D25, D20, D24, D23

  (103) 0x41a610 STR	D25, [X19, X10]

  (103) 0x41a614 LDR	D26, [X4, X10]

  (103) 0x41a618 LDR	D27, [X2, X10]

  (103) 0x41a61c FMUL	D28, D0, D26

  (103) 0x41a620 FMADD	D29, D20, D28, D27

  (103) 0x41a624 STR	D29, [X2, X10]

  (103) 0x41a628 CMP	X3, X21

  (103) 0x41a62c B.NE	41a580

(102) 0x41a630 ADD	X14, X14, #1

(102) 0x41a634 ADD	X15, X15, #1536

(102) 0x41a638 CMP	W16, W14

(102) 0x41a63c B.GT	41a4cc

0x41a640 LDP	X19, X20, [SP, #16]

0x41a644 LDR	X21, [SP, #32]

0x41a648 LDP	X29, X30, [SP], #48

0x41a64c RET

0x41a650 ADD	W1, W1, #1

0x41a654 MOVZ	W2, #0

0x41a658 B	41a498

0x41a65c HINT	#0

Coverage (%)	Name	Source Location	Module
►100.00+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►50.61+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	exec
○	main	CoMD.c:125	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	CoMD.c:266	exec
►49.39+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6

Coverage (%)	Name	Source Location	Module
►73.26+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►26.74+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►86.42+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►13.58+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►93.02+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►6.98+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►95.46+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►4.54+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►96.50+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►3.50+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►98.26+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.74+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►97.29+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►2.71+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►98.45+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.55+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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
►98.47+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.53+	GOMP_parallel		libgomp.so.1.0.0
○	timestep	timestep.c:43	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:85-94
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:85-94.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 25% 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 4.50 cycles (2.78x 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:86
- /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:86

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	35
nb uops	34
loop length	140
used w registers	10
used x registers	13
used b registers	0
used h registers	0
used s registers	1
used d registers	2
used q registers	0
used v registers	0
used z registers	0
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	4.25 cycles
front end	4.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.00	3.00	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	3.83	3.50	3.67	1.50	1.50
cycles	3.00	3.00	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	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	4.25
Dispatch	4.75
DIV/SQRT	5.00-12.50
Overall L1	5.00-12.50

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	25%
load	27%
store	41%
mul	12%
add-sub	18%
fma	12%
other	25%

all	25%
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	25%

INT+FP

all	25%
load	27%
store	41%
mul	12%
add-sub	18%
fma	12%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	25%

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: 41a460

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%)
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%)
LDR W21, [X0, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (12.5%)
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%)
ORR X20, 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	N/A
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 W19, 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
SDIV W1, W21, W19	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
ORR W14, 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%)
MSUB W2, W1, W19, W21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
CMP W0, W2	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 41a650 <advancePosition._omp_fn.0+0x1f0>	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, W1, W14, W2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
ADD W16, W1, 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, W16	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 41a640 <advancePosition._omp_fn.0+0x1e0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X17, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
UBFM W15, W3, #26, #25	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%)
MOVZ W0, #24	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
FMOV D4, #1.0000000	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%)
SBFM X14, 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%)
LDR D0, [X20, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (25.0%)
SMULL X15, W15, W0	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50	scal (12.5%)
LDR X4, [X17, #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
LDR X18, [X4, #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%)
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 W1, W1, #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 W2, #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 41a498 <advancePosition._omp_fn.0+0x38>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
HINT #0																			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:85-94
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:85-94.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 25% 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 4.50 cycles (2.78x 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:86
- /home/eoseret/qaas/qaas_runs/178-176-0594/intel/CoMD/build/CoMD/CoMD/src-openmp/timestep.c:86

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	35
nb uops	34
loop length	140
used w registers	10
used x registers	13
used b registers	0
used h registers	0
used s registers	1
used d registers	2
used q registers	0
used v registers	0
used z registers	0
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	4.25 cycles
front end	4.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.00	3.00	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	3.83	3.50	3.67	1.50	1.50
cycles	3.00	3.00	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	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	4.25
Dispatch	4.75
DIV/SQRT	5.00-12.50
Overall L1	5.00-12.50

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	25%
load	27%
store	41%
mul	12%
add-sub	18%
fma	12%
other	25%

all	25%
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	25%

INT+FP

all	25%
load	27%
store	41%
mul	12%
add-sub	18%
fma	12%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	25%

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: 41a460

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%)
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%)
LDR W21, [X0, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (12.5%)
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%)
ORR X20, 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	N/A
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 W19, 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
SDIV W1, W21, W19	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-12	5-12.50	N/A
ORR W14, 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%)
MSUB W2, W1, W19, W21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
CMP W0, W2	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 41a650 <advancePosition._omp_fn.0+0x1f0>	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, W1, W14, W2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (12.5%)
ADD W16, W1, 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, W16	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 41a640 <advancePosition._omp_fn.0+0x1e0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X17, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
UBFM W15, W3, #26, #25	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%)
MOVZ W0, #24	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
FMOV D4, #1.0000000	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%)
SBFM X14, 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%)
LDR D0, [X20, #8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33	scal (25.0%)
SMULL X15, W15, W0	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	0	0	2	0.50	scal (12.5%)
LDR X4, [X17, #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
LDR X18, [X4, #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%)
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 W1, W1, #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 W2, #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 41a498 <advancePosition._omp_fn.0+0x38>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
HINT #0																			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	0.96	0.02	0.9	0.05	0.81	0.09	0.58	0.26	0.47	0.4	0.41	0.51	0.36	0.59	0.31	0.72	0.26	0.86	0.22	1.02

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	2.7699997425079	0.42479884624481
1x2	2	0.96	1.92	2	1.450000166893	0.43863460421562
1x4	4	0.9	3.6	4	0.80500000715256	0.460270434618
1x8	8	0.81	6.47	8	0.44000002741814	0.4911299943924
1x16	16	0.58	9.33	16	0.29499998688698	0.63563376665115
1x24	24	0.47	11.23	24	0.25999999046326	0.7551993727684
1x32	32	0.41	13.01	32	0.27500000596046	0.86717563867569
1x40	40	0.36	14.5	40	0.18999999761581	0.92376744747162
1x48	48	0.31	14.84	48	0.1950000077486	1.0437558889389
1x56	56	0.26	14.83	56	0.21500000357628	1.167622923851
1x64	64	0.22	14.39	64	0.21500000357628	1.3132240772247

Name	Coverage (%)	Time (s)
▼advancePosition._omp_fn.0–	0.42	2.77
▼Loop 102 - timestep.c:88-94 - exec–	0.02	0.14
○Loop 103 - timestep.c:88-94 - exec	0.40	2.62

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