OV - exec - Loop 3650

0x4a4820 LD1D	{Z1.D}, P0/Z, [X17, X15,LSL #3]    [3]

0x4a4824 LD1D	{Z3.D}, P0/Z, [X11, X15,LSL #3]    [1]

0x4a4828 LD1D	{Z2.D}, P0/Z, [X16, X15,LSL #3]    [2]

0x4a482c FMUL	Z1.D, Z1.D, Z3.D

0x4a4830 LD1D	{Z4.D}, P0/Z, [X18, X15,LSL #3]    [4]

0x4a4834 FMUL	Z2.D, Z2.D, Z4.D

0x4a4838 INCW	X15, ALL

0x4a483c FADDA	D0, P0, D0, Z1.D

0x4a4840 FADDA	D0, P0, D0, Z2.D

0x4a4844 CMP	X13, X15

0x4a4848 B.NE	4a4820

/home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/seq_mv/vector.c: 486 - 486

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

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

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

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 /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	5.82
Bottlenecks
Function	hypre_SeqVectorInnerProd.omp_outlined
Source	vector.c:486-486
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	2
CQA cycles	8.00
CQA cycles if no scalar integer	8.00
CQA cycles if FP arith vectorized	8.00
CQA cycles if fully vectorized	8.00
Front-end cycles	1.38
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.00
P5 cycles	0.00
P6 cycles	0.50
P7 cycles	0.50
P8 cycles	1.00
P9 cycles	1.00
P10 cycles	1.00
P11 cycles	1.00
P12 cycles	1.33
P13 cycles	1.33
P14 cycles	1.33
P15 cycles	0.00
P16 cycles	0.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	8
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	11.00
Nb uops	11.00
Nb loads	NA
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	1.00
Nb FLOP add-sub	4.00
Nb FLOP mul	4.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	8.00
Bytes prefetched	0.00
Bytes loaded	64.00
Bytes stored	0.00
Stride 0	0.00
Stride 1	0.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	NA
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	100.00
Vector-efficiency ratio load	100.00
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	NA

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	5.82
Bottlenecks
Function	hypre_SeqVectorInnerProd.omp_outlined
Source	vector.c:486-486
Source loop unroll info	unrolled by 2
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	2
CQA cycles	8.00
CQA cycles if no scalar integer	8.00
CQA cycles if FP arith vectorized	8.00
CQA cycles if fully vectorized	8.00
Front-end cycles	1.38
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.00
P5 cycles	0.00
P6 cycles	0.50
P7 cycles	0.50
P8 cycles	1.00
P9 cycles	1.00
P10 cycles	1.00
P11 cycles	1.00
P12 cycles	1.33
P13 cycles	1.33
P14 cycles	1.33
P15 cycles	0.00
P16 cycles	0.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	8
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	11.00
Nb uops	11.00
Nb loads	NA
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	1.00
Nb FLOP add-sub	4.00
Nb FLOP mul	4.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	8.00
Bytes prefetched	0.00
Bytes loaded	64.00
Bytes stored	0.00
Stride 0	0.00
Stride 1	0.00
Stride n	0.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	NA
Vectorization ratio mul	100.00
Vectorization ratio add_sub	100.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	100.00
Vector-efficiency ratio load	100.00
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	100.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	NA

Path /

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

Function	hypre_SeqVectorInnerProd.omp_outlined
Source file and lines	vector.c:486-486
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/seq_mv/vector.c:486.

The related source loop is unrolled by 2 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is fully vectorized, using full register length.

Details

All VPU instructions are used in vector version (process two or more data elements in vector registers).

FMA

Presence of both ADD/SUB and MUL operations.

Workaround

Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible. For instance a + b*c is a valid FMA (MUL then ADD). However (a+b)* c cannot be translated into an FMA (ADD then MUL).

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

The binary loop is composed of 8 FP arithmetical operations:

4: addition or subtraction
4: multiply

The binary loop is loading 64 bytes.

Arithmetic intensity

Arithmetic intensity is 0.12 FP operations per loaded or stored byte.

General properties

nb instructions	11
loop length	44
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

1.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	0.50	0.50	0.50	0.50	0.00	0.00	0.50	0.50	1.00	1.00	1.00	1.00	1.33	1.33	1.33	0.00	0.00
cycles	0.50	0.50	0.50	0.50	0.00	0.00	0.50	0.50	1.00	1.00	1.00	1.00	1.33	1.33	1.33	0.00	0.00

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6: ALU
P7: ALU
P8 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P9 (128 bits): VPU, ALU, FP store data
P10 (128 bits): VPU, ALU, DIV/SQRT
P11 (128 bits): ALU, VPU
P12 (256 bits): store address, load
P13 (256 bits): store address, load
P14 (256 bits): load
P15 (64 bits): store data
P16 (64 bits): store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	8.00

Cycles summary

Front-end	1.38
Data deps.	8.00
Overall L1	8.00

Vectorization ratios

INT

all	100%
load	100%
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)
other	NA (no other vectorizable/vectorized instructions)

FP

all	100%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	100%
load	100%
store	NA (no store vectorizable/vectorized instructions)
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.00 cycles. At this rate:

8% of peak load performance is reached (8.00 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 4a4820

Instruction	Nb FU	P0	P1	P2	P3	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
LD1D {Z1.D}, P0/Z, [X17, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
LD1D {Z3.D}, P0/Z, [X11, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
LD1D {Z2.D}, P0/Z, [X16, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
FMUL Z1.D, Z1.D, Z3.D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (100.0%)
LD1D {Z4.D}, P0/Z, [X18, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
FMUL Z2.D, Z2.D, Z4.D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (100.0%)
INCW X15, ALL	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50	N/A
FADDA D0, P0, D0, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	4	0.25	vect (100.0%)
FADDA D0, P0, D0, Z2.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	4	0.25	vect (100.0%)
CMP X13, X15	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.33	N/A
B.NE 4a4820 <hypre_SeqVectorInnerProd.omp_outlined+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	hypre_SeqVectorInnerProd.omp_outlined
Source file and lines	vector.c:486-486
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/seq_mv/vector.c:486.

The related source loop is unrolled by 2 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is fully vectorized, using full register length.

Details

All VPU instructions are used in vector version (process two or more data elements in vector registers).

FMA

Presence of both ADD/SUB and MUL operations.

Workaround

Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible. For instance a + b*c is a valid FMA (MUL then ADD). However (a+b)* c cannot be translated into an FMA (ADD then MUL).

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

The binary loop is composed of 8 FP arithmetical operations:

4: addition or subtraction
4: multiply

The binary loop is loading 64 bytes.

Arithmetic intensity

Arithmetic intensity is 0.12 FP operations per loaded or stored byte.

General properties

nb instructions	11
loop length	44
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

1.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	0.50	0.50	0.50	0.50	0.00	0.00	0.50	0.50	1.00	1.00	1.00	1.00	1.33	1.33	1.33	0.00	0.00
cycles	0.50	0.50	0.50	0.50	0.00	0.00	0.50	0.50	1.00	1.00	1.00	1.00	1.33	1.33	1.33	0.00	0.00

Execution ports to units layout:

P0: BRU
P1: BRU
P2: ALU
P3: ALU
P4: ALU
P5: ALU
P6: ALU
P7: ALU
P8 (128 bits): VPU, FP store data, ALU, DIV/SQRT
P9 (128 bits): VPU, ALU, FP store data
P10 (128 bits): VPU, ALU, DIV/SQRT
P11 (128 bits): ALU, VPU
P12 (256 bits): store address, load
P13 (256 bits): store address, load
P14 (256 bits): load
P15 (64 bits): store data
P16 (64 bits): store data

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	8.00

Cycles summary

Front-end	1.38
Data deps.	8.00
Overall L1	8.00

Vectorization ratios

INT

all	100%
load	100%
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)
other	NA (no other vectorizable/vectorized instructions)

FP

all	100%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	100%
load	100%
store	NA (no store vectorizable/vectorized instructions)
mul	100%
add-sub	100%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.00 cycles. At this rate:

8% of peak load performance is reached (8.00 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 4a4820

Instruction	Nb FU	P0	P1	P2	P3	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
LD1D {Z1.D}, P0/Z, [X17, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
LD1D {Z3.D}, P0/Z, [X11, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
LD1D {Z2.D}, P0/Z, [X16, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
FMUL Z1.D, Z1.D, Z3.D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (100.0%)
LD1D {Z4.D}, P0/Z, [X18, X15,LSL #3]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (100.0%)
FMUL Z2.D, Z2.D, Z4.D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (100.0%)
INCW X15, ALL	1	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50	N/A
FADDA D0, P0, D0, Z1.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	4	0.25	vect (100.0%)
FADDA D0, P0, D0, Z2.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	4	0.25	vect (100.0%)
CMP X13, X15	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.33	N/A
B.NE 4a4820 <hypre_SeqVectorInnerProd.omp_outlined+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Report Configuration

Vectorization

Details

FMA

Workaround

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

Vectorization

Details

FMA

Workaround

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

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code