OV - exec - Function hypre_BoomerAMGCoarsenPMIS._omp

	Function: hypre_BoomerAMGCoarsenPMIS._omp_fn.7	Module: exec	Source: par_coarsen.c:2435-2477 [...]	Coverage (incl. loops): 1.15% \| (excl. loops): 0.00%

/home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c: 2435 - 2477

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

2435: #pragma omp parallel for private(ig, i, jS, j) HYPRE_SMP_SCHEDULE

[...]

2446:        if(measure_array[i]<1.) CF_marker[i]= F_PT;

[...]

2453:        if (CF_marker[i] > 0) CF_marker[i] = C_PT;

[...]

2465: 	 for (jS = S_diag_i[i]; jS < S_diag_i[i+1]; jS++)

2466:          {

2467: 	   /* j is the column number, or the local number of the point influencing i */

2468: 	   j = S_diag_j[jS];

2469: 	   if (CF_marker[j] > 0) /* j is a C-point */

2470: 	     CF_marker[i] = F_PT;

2471: 	 }

2472: 	 /* now the external part */

2473: 	 for (jS = S_offd_i[i]; jS < S_offd_i[i+1]; jS++)

2474: 	 {

2475: 	   j = S_offd_j[jS];

2476: 	   if (CF_marker_offd[j] > 0) /* j is a C-point */

2477: 	     CF_marker[i] = F_PT;

0x4275a4 STP	X29, X30, [SP, #976]!

0x4275a8 ADD	X29, SP, #0

0x4275ac STP	X19, X20, [SP, #16]

0x4275b0 ORR	X20, XZR, X0

0x4275b4 LDR	X0, [X0, #64]

0x4275b8 STR	X21, [SP, #32]

0x4275bc LDR	X21, [X0]

0x4275c0 BL	4101b0

0x4275c4 SBFM	X19, X0, #0, #31

0x4275c8 BL	4101c0

0x4275cc SBFM	X1, X0, #0, #31

0x4275d0 SDIV	X7, X21, X19

0x4275d4 MSUB	X2, X7, X19, X21

0x4275d8 CMP	X1, X2

0x4275dc B.LT	42770c

0x4275e0 MADD	X4, X7, X1, X2

0x4275e4 ADD	X5, X7, X4

0x4275e8 CMP	X4, X5

0x4275ec B.GE	4276ec

0x4275f0 LDP	X13, X6, [X20, #48]

0x4275f4 FMOV	D31, #1.0000000

0x4275f8 MOVN	X8, #0

0x4275fc MOVZ	X15, #1

0x427600 LDP	X17, X11, [X20]

0x427604 ADD	X1, X6, X4,LSL #3

0x427608 LDP	X16, X10, [X20, #16]

0x42760c ADD	X7, X6, X5,LSL #3

0x427610 LDP	X3, X9, [X20, #32]

0x427614 B	427634

(115) 0x427618 LDR	X14, [X3, X18]

(115) 0x42761c CMP	X14, #0

(115) 0x427620 B.LE	427650

(115) 0x427624 STR	X15, [X3, X18]

(115) 0x427628 ADD	X1, X1, #8

(115) 0x42762c CMP	X7, X1

(115) 0x427630 B.EQ	4276ec

(115) 0x427634 LDR	X12, [X1]

(115) 0x427638 UBFM	X18, X12, #61, #60

(115) 0x42763c ADD	X30, X3, X12,LSL #3

(115) 0x427640 LDR	D30, [X13, X18]

(115) 0x427644 FCMPE	D30, D31

(115) 0x427648 B.GE	427618

(115) 0x42764c STR	X8, [X3, X18]

(115) 0x427650 ADD	X20, X18, #8

(115) 0x427654 LDR	X0, [X17, X18]

(115) 0x427658 ADD	X21, X17, X20

(115) 0x42765c LDR	X4, [X17, X20]

(115) 0x427660 CMP	X0, X4

(115) 0x427664 B.GE	42768c

  (118) 0x427668 LDR	X19, [X11, X0,LSL #3]

  (118) 0x42766c LDR	X2, [X3, X19,LSL #3]

  (118) 0x427670 CMP	X2, #0

  (118) 0x427674 B.LE	4276fc

  (118) 0x427678 STR	X8, [X30]

  (118) 0x42767c ADD	X0, X0, #1

  (118) 0x427680 LDR	X4, [X21]

  (118) 0x427684 CMP	X4, X0

  (118) 0x427688 B.GT	427668

(115) 0x42768c LDR	X14, [X16, X18]

(115) 0x427690 ADD	X6, X16, X20

(115) 0x427694 LDR	X5, [X16, X20]

(115) 0x427698 CMP	X14, X5

(115) 0x42769c B.GE	427628

  (116) 0x4276a0 LDR	X12, [X10, X14,LSL #3]

  (116) 0x4276a4 LDR	X18, [X9, X12,LSL #3]

  (116) 0x4276a8 CMP	X18, #0

  (116) 0x4276ac B.LE	4276d4

  (117) 0x4276b0 STR	X8, [X30]

  (117) 0x4276b4 ADD	X14, X14, #1

  (117) 0x4276b8 LDR	X5, [X6]

  (117) 0x4276bc CMP	X14, X5

  (117) 0x4276c0 B.GE	427628

  (117) 0x4276c4 LDR	X12, [X10, X14,LSL #3]

  (117) 0x4276c8 LDR	X18, [X9, X12,LSL #3]

  (117) 0x4276cc CMP	X18, #0

  (117) 0x4276d0 B.GT	4276b0

  (116) 0x4276d4 ADD	X14, X14, #1

  (116) 0x4276d8 CMP	X14, X5

  (116) 0x4276dc B.LT	4276a0

(115) 0x4276e0 ADD	X1, X1, #8

(115) 0x4276e4 CMP	X7, X1

(115) 0x4276e8 B.NE	427634

0x4276ec LDR	X21, [SP, #32]

0x4276f0 LDP	X19, X20, [SP, #16]

0x4276f4 LDP	X29, X30, [SP], #48

0x4276f8 RET

  (118) 0x4276fc ADD	X0, X0, #1

  (118) 0x427700 CMP	X0, X4

  (118) 0x427704 B.LT	427668

(115) 0x427708 B	42768c

0x42770c ADD	X7, X7, #1

0x427710 MOVZ	X2, #0

0x427714 B	4275e0

0x427718 HINT	#0

0x42771c HINT	#0

Coverage (%)	Name	Source Location	Module
►92.18+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec
►5.80+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:638	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec
►2.02+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2491	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:638	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►49.87+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►46.47+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec
►2.77+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:638	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►75.00+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►22.77+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec
►1.73+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:638	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►87.42+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►11.49+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►93.88+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►5.77+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►96.18+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►3.60+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►96.92+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►2.81+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►97.93+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.99+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►98.08+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.84+	GOMP_parallel		libgomp.so.1.0.0
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►98.44+	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
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	exec

Coverage (%)	Name	Source Location	Module
►98.55+	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
○	hypre_BoomerAMGCoarsenPMIS	par_coarsen.c:2492	exec
○	hypre_BoomerAMGSetup	par_amg_setup.c:612	exec
○	hypre_PCGSetup	pcg.c:234	exec
○	main	amg.c:398	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	amg.c:253	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	par_coarsen.c:2435-2477
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2435,2446,2453,2465-2477.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 39% 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 20.00 to 4.50 cycles (4.44x 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/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2435
- /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2435

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 does not load or store any data.

General properties

nb instructions	38
nb uops	36
loop length	152
used w registers	0
used x registers	22
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	0
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	4.50 cycles
front end	4.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.50	3.50	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	4.17	3.83	4.00	1.50	1.50
cycles	3.50	3.50	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	4.17	3.83	4.00	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

5.00-20.00

Cycles summary

Front-end	4.50
Dispatch	4.75
DIV/SQRT	5.00-20.00
Overall L1	5.00-20.00

Vectorization ratios

INT

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

Vector efficiency ratios

INT

all	39%
load	42%
store	41%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	25%
other	46%

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	39%
load	42%
store	41%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	25%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	43%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 20.00 cycles.

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 4275a4

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%)
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
LDR X0, [X0, #64]	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%)
LDR X21, [X0]	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%)
BL 4101b0 <@plt_start@+0x190>	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
SBFM X19, X0, #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%)
BL 4101c0 <@plt_start@+0x1a0>	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
SBFM X1, X0, #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%)
SDIV X7, X21, X19	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-20	5-20	N/A
MSUB X2, X7, X19, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
CMP X1, X2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.LT 42770c <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x168>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD X4, X7, X1, X2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
ADD X5, X7, X4	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%)
CMP X4, X5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.GE 4276ec <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x148>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X13, X6, [X20, #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%)
FMOV D31, #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%)
MOVN X8, #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%)
MOVZ X15, #1	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%)
LDP X17, X11, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
ADD X1, X6, X4,LSL #3	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%)
LDP X16, X10, [X20, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
ADD X7, X6, X5,LSL #3	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
LDP X3, X9, [X20, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
B 427634 <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x90>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	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 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
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 X7, X7, #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 X2, #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%)
B 4275e0 <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x3c>	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
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	par_coarsen.c:2435-2477
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2435,2446,2453,2465-2477.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 39% 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 20.00 to 4.50 cycles (4.44x speedup).

CALL instructions

Detected function call instructions.

Details

unknown: 2 occurrences►
- /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2435
- /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2435

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 does not load or store any data.

General properties

nb instructions	38
nb uops	36
loop length	152
used w registers	0
used x registers	22
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	0
nb stack references	6

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	4.50 cycles
front end	4.50 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	3.50	3.50	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	4.17	3.83	4.00	1.50	1.50
cycles	3.50	3.50	4.75	4.75	4.75	4.75	0.25	0.25	0.25	0.25	4.17	3.83	4.00	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

5.00-20.00

Cycles summary

Front-end	4.50
Dispatch	4.75
DIV/SQRT	5.00-20.00
Overall L1	5.00-20.00

Vectorization ratios

INT

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

Vector efficiency ratios

INT

all	39%
load	42%
store	41%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	25%
other	46%

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	39%
load	42%
store	41%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	25%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	43%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each call to the function takes 20.00 cycles.

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the function is: 4275a4

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%)
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
LDR X0, [X0, #64]	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%)
LDR X21, [X0]	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%)
BL 4101b0 <@plt_start@+0x190>	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
SBFM X19, X0, #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%)
BL 4101c0 <@plt_start@+0x1a0>	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
SBFM X1, X0, #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%)
SDIV X7, X21, X19	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-20	5-20	N/A
MSUB X2, X7, X19, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
CMP X1, X2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.LT 42770c <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x168>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD X4, X7, X1, X2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
ADD X5, X7, X4	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%)
CMP X4, X5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33	scal (25.0%)
B.GE 4276ec <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x148>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X13, X6, [X20, #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%)
FMOV D31, #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%)
MOVN X8, #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%)
MOVZ X15, #1	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%)
LDP X17, X11, [X20]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
ADD X1, X6, X4,LSL #3	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%)
LDP X16, X10, [X20, #16]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
ADD X7, X6, X5,LSL #3	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
LDP X3, X9, [X20, #32]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
B 427634 <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x90>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	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 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
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 X7, X7, #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 X2, #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%)
B 4275e0 <hypre_BoomerAMGCoarsenPMIS._omp_fn.7+0x3c>	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
HINT #0																			N/A

▶Scalability Runs Description

Run 1x1	Number processes: 1Number nodes: NARun Command: <executable> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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> -n 400 400 400MPI 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/178-188-3659/intel/AMG/run/oneview_runs/multicore/gcc_1/oneview_run_1781892409OMP_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	1.92	0	3.53	0	6.23	0	11.45	0	15.26	0	16.47	0	18.84	0	19.57	0	20.36	0	21.24	0

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	3.9650001525879	1.1456248760223
1x2	2	1.92	1.92	2	1.9899997711182	1.0995657444
1x4	4	3.53	3.53	4	1.0249999761581	1.0945166349411
1x8	8	6.23	6.23	8	0.52999997138977	1.0297570228577
1x16	16	11.45	11.45	16	0.28000000119209	0.65911000967026
1x24	24	15.26	15.26	24	0.21000000834465	0.52039098739624
1x32	32	16.47	16.47	32	0.19000001251698	0.5152872800827
1x40	40	18.84	18.84	40	0.17000000178814	0.4764911532402
1x48	48	19.57	19.57	48	0.15500001609325	0.45723041892052
1x56	56	20.36	20.36	56	0.15500000119209	0.44673562049866
1x64	64	21.24	21.24	64	0.1499999910593	0.420344799757

Name	Coverage (%)	Time (s)
▼hypre_BoomerAMGCoarsenPMIS._omp_fn.7–	1.15	3.97
▼Loop 115 - par_coarsen.c:2446-2477 - exec–	0.17	0.60
○Loop 118 - par_coarsen.c:2465-2470 - exec	0.97	3.37
○Loop 117 - par_coarsen.c:2473-2477 - exec	0.00	0.00
○Loop 116 - par_coarsen.c:2473-2476 - 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