OV - exec - Function hypre_SeqVectorInnerProd._omp

	Function: hypre_SeqVectorInnerProd._omp_fn.0	Module: exec	Source: vector.c:483-486	Coverage (incl. loops): 0.67% \| (excl. loops): 0.00%

/home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/seq_mv/vector.c: 483 - 486

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

483: #pragma omp parallel for private(i) reduction(+:result) HYPRE_SMP_SCHEDULE

484: #endif

485:    for (i = 0; i < size; i++)

486:       result += hypre_conj(y_data[i]) * x_data[i];

0x51d668 STP	X29, X30, [SP, #976]!

0x51d66c ADD	X29, SP, #0

0x51d670 STP	X19, X20, [SP, #16]

0x51d674 ORR	X19, XZR, X0

0x51d678 STR	X21, [SP, #32]

0x51d67c LDR	X21, [X0, #16]

0x51d680 BL	4100b0

0x51d684 SBFM	X20, X0, #0, #31

0x51d688 BL	410200

0x51d68c SBFM	X3, X0, #0, #31

0x51d690 SDIV	X2, X21, X20

0x51d694 MSUB	X0, X2, X20, X21

0x51d698 CMP	X3, X0

0x51d69c B.LT	51d71c

0x51d6a0 MADD	X6, X2, X3, X0

0x51d6a4 MOVI	D30, #0

0x51d6a8 ADD	X1, X2, X6

0x51d6ac CMP	X6, X1

0x51d6b0 B.GE	51d6e8

0x51d6b4 LDP	X7, X5, [X19]

0x51d6b8 MOVZ	X10, #0

0x51d6bc CNTD	X4, ALL

0x51d6c0 WHILELO	P7.D, XZR, X2

0x51d6c4 ADD	X8, X7, X6,LSL #3

0x51d6c8 ADD	X9, X5, X6,LSL #3

(3291) 0x51d6cc LD1D	{Z31.D}, P7/Z, [X9, X10,LSL #3]

(3291) 0x51d6d0 LD1D	{Z29.D}, P7/Z, [X8, X10,LSL #3]

(3291) 0x51d6d4 ADD	X10, X10, X4

(3291) 0x51d6d8 FMUL	Z31.D, P7/M, Z31.D, Z29.D

(3291) 0x51d6dc FADDA	D30, P7, D30, Z31.D

(3291) 0x51d6e0 WHILELO	P7.D, X10, X2

(3291) 0x51d6e4 B.NE	51d6cc

0x51d6e8 ADD	X13, X19, #24

0x51d6ec LDR	X1, [X13]

(3290) 0x51d6f0 FMOV	D0, X1

(3290) 0x51d6f4 ORR	X11, XZR, X1

(3290) 0x51d6f8 FADD	D1, D30, D0

(3290) 0x51d6fc FMOV	X12, D1

(3290) 0x51d700 CAS	X11, X12, [X13]

(3290) 0x51d704 CMP	X1, X11

(3290) 0x51d708 B.NE	51d728

0x51d70c LDR	X21, [SP, #32]

0x51d710 LDP	X19, X20, [SP, #16]

0x51d714 LDP	X29, X30, [SP], #48

0x51d718 RET

0x51d71c ADD	X2, X2, #1

0x51d720 MOVZ	X0, #0

0x51d724 B	51d6a0

(3290) 0x51d728 ORR	X1, XZR, X11

(3290) 0x51d72c B	51d6f0

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

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 gcc_0

Source file and lines	vector.c:483-486
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/seq_mv/vector.c:483-486.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 37% 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 5.00 cycles (4.00x 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/seq_mv/vector.c:483
- /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/seq_mv/vector.c:483

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	34
nb uops	34
loop length	136
used w registers	0
used x registers	18
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	4.25 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	5.25	5.25	5.25	5.25	0.25	0.25	0.25	0.25	3.17	2.83	3.00	1.50	1.50
cycles	3.00	3.00	5.25	5.25	5.25	5.25	0.25	0.25	0.25	0.25	3.17	2.83	3.00	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	4.25
Dispatch	5.25
DIV/SQRT	5.00-20.00
Overall L1	5.25-20.00

Vectorization ratios

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

all	37%
load	35%
store	41%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	25%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	46%

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: 51d668

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 X19, 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
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, #16]	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 4100b0 <@plt_start@+0x90>	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 X20, 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 410200 <@plt_start@+0x1e0>	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 X3, 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 X2, X21, X20	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-20	5-20	N/A
MSUB X0, X2, X20, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	N/A
CMP X3, X0	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 51d71c <hypre_SeqVectorInnerProd._omp_fn.0+0xb4>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD X6, X2, X3, X0	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
MOVI D30, #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 X1, X2, X6	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 X6, X1	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 51d6e8 <hypre_SeqVectorInnerProd._omp_fn.0+0x80>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X7, X5, [X19]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
MOVZ X10, #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%)
CNTD X4, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
WHILELO P7.D, XZR, X2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2	N/A
ADD X8, X7, X6,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%)
ADD X9, X5, X6,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%)
ADD X13, X19, #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
LDR X1, [X13]	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%)
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 X2, X2, #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 X0, #0	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
B 51d6a0 <hypre_SeqVectorInnerProd._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

The code analyzed by CQA in that panel excludes loops and represents 0.00% of application time for run gcc_0

Source file and lines	vector.c:483-486
Module	exec

The function is defined in /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/seq_mv/vector.c:483-486.

Warnings:
Ignoring paths for analysis

gain
potential
hint
expert

Vectorization

Your function is not vectorized. Only 37% 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 5.00 cycles (4.00x speedup).

CALL instructions

Detected function call instructions.

Details

unknown: 2 occurrences►
- /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/seq_mv/vector.c:483
- /home/eoseret/qaas/qaas_runs/178-188-3659/intel/AMG/build/AMG/AMG/seq_mv/vector.c:483

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	34
nb uops	34
loop length	136
used w registers	0
used x registers	18
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	4.25 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	5.25	5.25	5.25	5.25	0.25	0.25	0.25	0.25	3.17	2.83	3.00	1.50	1.50
cycles	3.00	3.00	5.25	5.25	5.25	5.25	0.25	0.25	0.25	0.25	3.17	2.83	3.00	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	4.25
Dispatch	5.25
DIV/SQRT	5.00-20.00
Overall L1	5.25-20.00

Vectorization ratios

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

all	37%
load	35%
store	41%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	25%
fma	25%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	46%

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: 51d668

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 X19, 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
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, #16]	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 4100b0 <@plt_start@+0x90>	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 X20, 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 410200 <@plt_start@+0x1e0>	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 X3, 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 X2, X21, X20	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	5-20	5-20	N/A
MSUB X0, X2, X20, X21	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	N/A
CMP X3, X0	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 51d71c <hypre_SeqVectorInnerProd._omp_fn.0+0xb4>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
MADD X6, X2, X3, X0	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
MOVI D30, #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 X1, X2, X6	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 X6, X1	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 51d6e8 <hypre_SeqVectorInnerProd._omp_fn.0+0x80>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDP X7, X5, [X19]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1	scal (50.0%)
MOVZ X10, #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%)
CNTD X4, ALL	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
WHILELO P7.D, XZR, X2	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	3	2	N/A
ADD X8, X7, X6,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%)
ADD X9, X5, X6,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%)
ADD X13, X19, #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
LDR X1, [X13]	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%)
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 X2, X2, #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 X0, #0	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
B 51d6a0 <hypre_SeqVectorInnerProd._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

Name	Coverage (%)	Time (s)
▼hypre_SeqVectorInnerProd._omp_fn.0–	0.67	0.28
○Loop 3291 - vector.c:486-486 - exec	0.67	0.21
○Loop 3290 - vector.c:483-483 - 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