OV - exec - Loop 12

	Loop Id: 12	Module: exec	Source: stream.c:336-336	Coverage: 22.18%

0x401a34 LDR	Q5, [X30, X0]    [1]

0x401a38 ADD	X6, X0, #32

0x401a3c ADD	X8, X0, #64

0x401a40 ADD	X7, X0, #96

0x401a44 ADD	X10, X0, #128

0x401a48 LDR	Q6, [X19, X0]    [4]

0x401a4c ADD	X13, X0, #160

0x401a50 ADD	X12, X0, #192

0x401a54 ADD	X11, X0, #224

0x401a58 FMUL	V26.2D, V5.2D, V30.2D

0x401a5c FMUL	V27.2D, V6.2D, V30.2D

0x401a60 STR	Q26, [X18, X0]    [3]

0x401a64 STR	Q27, [X2, X0]    [2]

0x401a68 ADD	X0, X0, #256

0x401a6c LDR	Q7, [X30, X6]    [1]

0x401a70 LDR	Q16, [X19, X6]    [4]

0x401a74 FMUL	V24.2D, V7.2D, V30.2D

0x401a78 FMUL	V25.2D, V16.2D, V30.2D

0x401a7c STR	Q24, [X18, X6]    [3]

0x401a80 STR	Q25, [X2, X6]    [2]

0x401a84 LDR	Q17, [X30, X8]    [1]

0x401a88 LDR	Q18, [X19, X8]    [4]

0x401a8c FMUL	V23.2D, V17.2D, V30.2D

0x401a90 FMUL	V29.2D, V18.2D, V30.2D

0x401a94 STR	Q23, [X18, X8]    [3]

0x401a98 STR	Q29, [X2, X8]    [2]

0x401a9c LDR	Q19, [X30, X7]    [1]

0x401aa0 LDR	Q20, [X19, X7]    [4]

0x401aa4 FMUL	V21.2D, V19.2D, V30.2D

0x401aa8 FMUL	V22.2D, V20.2D, V30.2D

0x401aac STR	Q21, [X18, X7]    [3]

0x401ab0 STR	Q22, [X2, X7]    [2]

0x401ab4 LDR	Q1, [X30, X10]    [1]

0x401ab8 LDR	Q0, [X19, X10]    [4]

0x401abc FMUL	V3.2D, V1.2D, V30.2D

0x401ac0 FMUL	V2.2D, V0.2D, V30.2D

0x401ac4 STR	Q3, [X18, X10]    [3]

0x401ac8 STR	Q2, [X2, X10]    [2]

0x401acc LDR	Q28, [X30, X13]    [1]

0x401ad0 LDR	Q4, [X19, X13]    [4]

0x401ad4 FMUL	V5.2D, V28.2D, V30.2D

0x401ad8 FMUL	V6.2D, V4.2D, V30.2D

0x401adc STR	Q5, [X18, X13]    [3]

0x401ae0 STR	Q6, [X2, X13]    [2]

0x401ae4 LDR	Q26, [X30, X12]    [1]

0x401ae8 LDR	Q27, [X19, X12]    [4]

0x401aec FMUL	V7.2D, V26.2D, V30.2D

0x401af0 FMUL	V16.2D, V27.2D, V30.2D

0x401af4 STR	Q7, [X18, X12]    [3]

0x401af8 STR	Q16, [X2, X12]    [2]

0x401afc LDR	Q24, [X30, X11]    [1]

0x401b00 LDR	Q25, [X19, X11]    [4]

0x401b04 FMUL	V17.2D, V24.2D, V30.2D

0x401b08 FMUL	V18.2D, V25.2D, V30.2D

0x401b0c STR	Q17, [X18, X11]    [3]

0x401b10 STR	Q18, [X2, X11]    [2]

0x401b14 CMP	X20, X0

0x401b18 B.NE	401a34

/home/eoseret/Tools/QaaS/qaas_runs/175-491-9774/intel/stream/build/stream/src/stream.c: 336 - 336

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

336: 	    b[j] = scalar*c[j];

Coverage (%)	Name	Source Location	Module
►98.43+	gomp_thread_start	team.c:130	libgomp.so.1.0.0
○	pthread_condattr_setpshared		libc.so.6
○	__clone		libc.so.6
►1.57+	GOMP_parallel	libgomp.h:980	libgomp.so.1.0.0
○	main	stream.c:448	exec
○	__libc_init_first		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	stream.c:290	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 /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	2.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.33
Bottlenecks	P10, P11, P12,
Function	main._omp_fn.5
Source	stream.c:336-336
Source loop unroll info	unrolled by 4
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	4
CQA cycles	10.67
CQA cycles if no scalar integer	10.67
CQA cycles if FP arith vectorized	10.67
CQA cycles if fully vectorized	5.33
Front-end cycles	7.25
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	2.25
P3 cycles	2.25
P4 cycles	2.25
P5 cycles	2.25
P6 cycles	8.00
P7 cycles	8.00
P8 cycles	8.00
P9 cycles	8.00
P10 cycles	10.67
P11 cycles	10.67
P12 cycles	10.67
P13 cycles	0.00
P14 cycles	0.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	58.00
Nb uops	58.00
Nb loads	NA
Nb stores	16.00
Nb stack references	0.00
FLOP/cycle	3.00
Nb FLOP add-sub	0.00
Nb FLOP mul	32.00
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	48.00
Bytes prefetched	0.00
Bytes loaded	256.00
Bytes stored	256.00
Stride 0	0.00
Stride 1	0.00
Stride n	4.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	50.00
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	50.00
Vector-efficiency ratio add_sub	NA
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	2.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.33
Bottlenecks	P10, P11, P12,
Function	main._omp_fn.5
Source	stream.c:336-336
Source loop unroll info	unrolled by 4
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	4
CQA cycles	10.67
CQA cycles if no scalar integer	10.67
CQA cycles if FP arith vectorized	10.67
CQA cycles if fully vectorized	5.33
Front-end cycles	7.25
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	2.25
P3 cycles	2.25
P4 cycles	2.25
P5 cycles	2.25
P6 cycles	8.00
P7 cycles	8.00
P8 cycles	8.00
P9 cycles	8.00
P10 cycles	10.67
P11 cycles	10.67
P12 cycles	10.67
P13 cycles	0.00
P14 cycles	0.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	58.00
Nb uops	58.00
Nb loads	NA
Nb stores	16.00
Nb stack references	0.00
FLOP/cycle	3.00
Nb FLOP add-sub	0.00
Nb FLOP mul	32.00
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	48.00
Bytes prefetched	0.00
Bytes loaded	256.00
Bytes stored	256.00
Stride 0	0.00
Stride 1	0.00
Stride n	4.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	100.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	50.00
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	50.00
Vector-efficiency ratio add_sub	NA
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	main._omp_fn.5
Source file and lines	stream.c:336-336
Module	exec

The loop is defined in /home/eoseret/Tools/QaaS/qaas_runs/175-491-9774/intel/stream/build/stream/src/stream.c:336.

The related source loop is unrolled by 4 (including vectorization).

gain
potential
hint
expert

Vectorization

Your loop is vectorized, but using only 128 out of 256 bits (SSE/AVX-128 instructions on AVX/AVX2 processors). By fully vectorizing your loop, you can lower the cost of an iteration from 10.67 to 5.33 cycles (2.00x speedup).

Details

All VPU instructions are used in vector version (process two or more data elements in vector registers). Since your execution units are vector units, only a fully vectorized loop can use their full power.

No data for this section

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Constant non-unit stride: 4 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

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

The binary loop is composed of 32 FP arithmetical operations:

32: multiply

The binary loop is loading 256 bytes. The binary loop is storing 256 bytes.

Arithmetic intensity

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.

General properties

nb instructions	58
nb uops	58
loop length	232
used w registers	0
used x registers	13
used b registers	0
used h registers	0
used s registers	0
used d registers	0
used q registers	22
used v registers	23
used z registers	0
nb stack references	0

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.25 cycles
front end	7.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	2.25	2.25	2.25	2.25	8.00	8.00	8.00	8.00	10.67	10.67	10.67	0.00	0.00
cycles	0.50	0.50	2.25	2.25	2.25	2.25	8.00	8.00	8.00	8.00	10.67	10.67	10.67	0.00	0.00

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	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	7.25
Dispatch	10.67
Data deps.	1.00
Overall L1	10.67

Vectorization ratios

INT

all	100%
load	100%
store	100%
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)

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

INT+FP

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

Vector efficiency ratios

INT

all	50%
load	50%
store	50%
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)

all	50%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	50%
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	NA (no other vectorizable/vectorized instructions)

INT+FP

all	50%
load	50%
store	50%
mul	50%
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	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 10.67 cycles. At this rate:

24% of peak load performance is reached (24.00 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
49% of peak store performance is reached (24.00 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 401a34

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput	Vectorization
LDR Q5, [X30, X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
ADD X6, X0, #32	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X8, X0, #64	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X7, X0, #96	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X10, X0, #128	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
LDR Q6, [X19, X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
ADD X13, X0, #160	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X12, X0, #192	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X11, X0, #224	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
FMUL V26.2D, V5.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V27.2D, V6.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q26, [X18, X0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q27, [X2, X0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
ADD X0, X0, #256	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
LDR Q7, [X30, X6]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q16, [X19, X6]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V24.2D, V7.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V25.2D, V16.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q24, [X18, X6]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q25, [X2, X6]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q17, [X30, X8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q18, [X19, X8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V23.2D, V17.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V29.2D, V18.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q23, [X18, X8]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q29, [X2, X8]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q19, [X30, X7]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q20, [X19, X7]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V21.2D, V19.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V22.2D, V20.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q21, [X18, X7]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q22, [X2, X7]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q1, [X30, X10]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q0, [X19, X10]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V3.2D, V1.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V2.2D, V0.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q3, [X18, X10]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q2, [X2, X10]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q28, [X30, X13]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q4, [X19, X13]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V5.2D, V28.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V6.2D, V4.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q5, [X18, X13]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q6, [X2, X13]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q26, [X30, X12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q27, [X19, X12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V7.2D, V26.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V16.2D, V27.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q7, [X18, X12]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q16, [X2, X12]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q24, [X30, X11]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q25, [X19, X11]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V17.2D, V24.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V18.2D, V25.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q17, [X18, X11]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q18, [X2, X11]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
CMP X20, X0	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 401a34 <main._omp_fn.5+0x18c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	main._omp_fn.5
Source file and lines	stream.c:336-336
Module	exec

The loop is defined in /home/eoseret/Tools/QaaS/qaas_runs/175-491-9774/intel/stream/build/stream/src/stream.c:336.

The related source loop is unrolled by 4 (including vectorization).

gain
potential
hint
expert

Vectorization

Details

No data for this section

Slow data structures access

Detected data structures (typically arrays) that cannot be efficiently read/written

Details

Constant non-unit stride: 4 occurrence(s)

Non-unit stride (uncontiguous) accesses are not efficiently using data caches

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

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

The binary loop is composed of 32 FP arithmetical operations:

32: multiply

The binary loop is loading 256 bytes. The binary loop is storing 256 bytes.

Arithmetic intensity

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.

General properties

nb instructions	58
nb uops	58
loop length	232
used w registers	0
used x registers	13
used b registers	0
used h registers	0
used s registers	0
used d registers	0
used q registers	22
used v registers	23
used z registers	0
nb stack references	0

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	7.25 cycles
front end	7.25 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	2.25	2.25	2.25	2.25	8.00	8.00	8.00	8.00	10.67	10.67	10.67	0.00	0.00
cycles	0.50	0.50	2.25	2.25	2.25	2.25	8.00	8.00	8.00	8.00	10.67	10.67	10.67	0.00	0.00

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	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	7.25
Dispatch	10.67
Data deps.	1.00
Overall L1	10.67

Vectorization ratios

INT

all	100%
load	100%
store	100%
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)

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

INT+FP

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

Vector efficiency ratios

INT

all	50%
load	50%
store	50%
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)

all	50%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	50%
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	NA (no other vectorizable/vectorized instructions)

INT+FP

all	50%
load	50%
store	50%
mul	50%
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	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 10.67 cycles. At this rate:

24% of peak load performance is reached (24.00 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
49% of peak store performance is reached (24.00 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 401a34

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput	Vectorization
LDR Q5, [X30, X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
ADD X6, X0, #32	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X8, X0, #64	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X7, X0, #96	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X10, X0, #128	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
LDR Q6, [X19, X0]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
ADD X13, X0, #160	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X12, X0, #192	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
ADD X11, X0, #224	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
FMUL V26.2D, V5.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V27.2D, V6.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q26, [X18, X0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q27, [X2, X0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
ADD X0, X0, #256	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
LDR Q7, [X30, X6]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q16, [X19, X6]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V24.2D, V7.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V25.2D, V16.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q24, [X18, X6]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q25, [X2, X6]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q17, [X30, X8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q18, [X19, X8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V23.2D, V17.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V29.2D, V18.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q23, [X18, X8]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q29, [X2, X8]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q19, [X30, X7]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q20, [X19, X7]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V21.2D, V19.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V22.2D, V20.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q21, [X18, X7]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q22, [X2, X7]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q1, [X30, X10]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q0, [X19, X10]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V3.2D, V1.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V2.2D, V0.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q3, [X18, X10]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q2, [X2, X10]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q28, [X30, X13]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q4, [X19, X13]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V5.2D, V28.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V6.2D, V4.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q5, [X18, X13]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q6, [X2, X13]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q26, [X30, X12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q27, [X19, X12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V7.2D, V26.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V16.2D, V27.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q7, [X18, X12]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q16, [X2, X12]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
LDR Q24, [X30, X11]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q25, [X19, X11]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMUL V17.2D, V24.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
FMUL V18.2D, V25.2D, V30.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	3	0.25	vect (50.0%)
STR Q17, [X18, X11]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
STR Q18, [X2, X11]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	vect (50.0%)
CMP X20, X0	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 401a34 <main._omp_fn.5+0x18c>	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

Slow data structures access

Details

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

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Vectorization

Details

Slow data structures access

Details

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

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code