OV - exec - Loop 26

0x411760 LDR	Q29, [X5, X1]    [1]

0x411764 LDR	Q31, [X3, X1]    [4]

0x411768 FMLA	V31.2D, V29.2D, V28.2D

0x41176c STR	Q31, [X3, X1]    [4]

0x411770 LDR	Q29, [X6, X1]    [2]

0x411774 LDR	Q31, [X2, X1]    [3]

0x411778 FMLS	V31.2D, V28.2D, V29.2D

0x41177c FMUL	V29.2D, V31.2D, V31.2D

0x411780 STR	Q31, [X2, X1]    [3]

0x411784 ADD	X1, X1, #16

0x411788 FADD	D30, D30, D29

0x41178c MOV	D31, V29.D[1]

0x411790 FADD	D30, D31, D30

0x411794 CMP	X7, X1

0x411798 B.NE	411760

/home/eoseret/qaas/qaas_runs/178-237-4322/intel/TeaLeaf/build/TeaLeaf/src/omp/cg.cpp: 108 - 113

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

108:     for (int kk = halo_depth; kk < x - halo_depth; ++kk) {

109:       const int index = kk + jj * x;

110:

111:       u[index] += alpha * p[index];

112:       r[index] -= alpha * w[index];

113:       rrn_temp += r[index] * r[index];

Coverage (%)	Name	Source Location	Module
►98.42+	omp_fulfill_event		libgomp.so.1.0.0
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.58+	GOMP_parallel		libgomp.so.1.0.0
○	run_cg_calc_ur(Chunk*, Setting[...]	cg.cpp:117	exec
○	cg_main_step_driver(Chunk*, Se[...]	cg_driver.cpp:69	exec
○	cg_driver(Chunk*, Settings&, d[...]	cg_driver.cpp:18	exec
○	solve(Chunk*, Settings&, int, [...]	diffuse.cpp:51	exec
○	diffuse(Chunk*, Settings&)	diffuse.cpp:12	exec
○	main	main.cpp:179	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	new_allocator.h:104	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	1.67
CQA speedup if fully vectorized	2.50
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	2.00
Bottlenecks
Function	cg_calc_ur(int, int, int, double, double, double, double const, double, double const*) [clone ._omp_fn.0]
Source	cg.cpp:108-113
Source loop unroll info	multi-versionned
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	4.00
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	2.40
CQA cycles if fully vectorized	1.60
Front-end cycles	1.88
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.50
P5 cycles	0.50
P6 cycles	2.00
P7 cycles	2.00
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	2.00
P11 cycles	2.00
P12 cycles	2.00
P13 cycles	0.00
P14 cycles	0.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	4
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	15.00
Nb uops	15.00
Nb loads	NA
Nb stores	2.00
Nb stack references	0.00
FLOP/cycle	3.00
Nb FLOP add-sub	2.00
Nb FLOP mul	2.00
Nb FLOP fma	4.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	0.00
Bytes prefetched	0.00
Bytes loaded	0.00
Bytes stored	0.00
Stride 0	0.00
Stride 1	0.00
Stride n	4.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	75.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	0.00
Vectorization ratio fma	100.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	43.75
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	50.00
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	50.00
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	25.00

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.67
CQA speedup if fully vectorized	2.50
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	2.00
Bottlenecks
Function	cg_calc_ur(int, int, int, double, double, double, double const, double, double const*) [clone ._omp_fn.0]
Source	cg.cpp:108-113
Source loop unroll info	multi-versionned
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	4.00
CQA cycles if no scalar integer	4.00
CQA cycles if FP arith vectorized	2.40
CQA cycles if fully vectorized	1.60
Front-end cycles	1.88
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.50
P5 cycles	0.50
P6 cycles	2.00
P7 cycles	2.00
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	2.00
P11 cycles	2.00
P12 cycles	2.00
P13 cycles	0.00
P14 cycles	0.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	4
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	15.00
Nb uops	15.00
Nb loads	NA
Nb stores	2.00
Nb stack references	0.00
FLOP/cycle	3.00
Nb FLOP add-sub	2.00
Nb FLOP mul	2.00
Nb FLOP fma	4.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	0.00
Bytes prefetched	0.00
Bytes loaded	0.00
Bytes stored	0.00
Stride 0	0.00
Stride 1	0.00
Stride n	4.00
Stride unknown	0.00
Stride indirect	0.00
Vectorization ratio all	75.00
Vectorization ratio load	100.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	0.00
Vectorization ratio fma	100.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	43.75
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	50.00
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	50.00
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	25.00

Path /

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

Function	cg_calc_ur(int, int, int, double, double, double, double const, double, double const*) [clone ._omp_fn.0]
Source file and lines	cg.cpp:108-113
Module	exec

The loop is defined in /home/eoseret/qaas/qaas_runs/178-237-4322/intel/TeaLeaf/build/TeaLeaf/src/omp/cg.cpp:108-113.

The related source loop is multi-versionned.

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 43% of vector register length is used (average across all VPU instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 4.00 to 1.60 cycles (2.50x speedup).

Details

75% of VPU instructions are used in vector version (process two or more data elements in vector registers):

0% of VPU addition or subtraction instructions are used in vector version.
0% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with fassociative-math (included in Ofast or ffast-math) to extend loop vectorization to FP reductions.
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

FMA

Detected 4 FMA (fused multiply-add) operations. 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).

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 12 FP arithmetical operations:

6: addition or subtraction (4 inside FMA instructions)
6: multiply (4 inside FMA instructions)

The binary loop does not load or store any data.

General properties

nb instructions	15
nb uops	15
loop length	60
used w registers	0
used x registers	6
used b registers	0
used h registers	0
used s registers	0
used d registers	3
used q registers	2
used v registers	3
used z registers	0
nb stack references	0
ADD-SUB / MUL ratio	2.00

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	1.88 cycles
front end	1.88 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	0.50	0.50	0.50	0.50	2.00	2.00	2.00	2.00	2.00	2.00	2.00	0.00	0.00
cycles	0.50	0.50	0.50	0.50	0.50	0.50	2.00	2.00	2.00	2.00	2.00	2.00	2.00	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)	4.00

Cycles summary

Front-end	1.88
Dispatch	2.00
Data deps.	4.00
Overall L1	4.00

Vectorization ratios

INT

all	85%
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	0%

FP

all	60%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	100%
add-sub	0%
fma	100%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	75%
load	100%
store	100%
mul	100%
add-sub	0%
fma	100%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	46%
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	25%

FP

all	40%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	50%
add-sub	25%
fma	50%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	43%
load	50%
store	50%
mul	50%
add-sub	25%
fma	50%
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 iteration of the binary loop takes 4.00 cycles.

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 411760

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput	Vectorization
LDR Q29, [X5, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q31, [X3, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMLA V31.2D, V29.2D, V28.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	4	0.25	vect (50.0%)
STR Q31, [X3, X1]	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 Q29, [X6, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q31, [X2, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMLS V31.2D, V28.2D, V29.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	4	0.25	vect (50.0%)
FMUL V29.2D, V31.2D, V31.2D	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 Q31, [X2, X1]	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 X1, X1, #16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
FADD D30, D30, D29	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	scal (25.0%)
MOV D31, V29.D[1]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	scal (25.0%)
FADD D30, D31, D30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	scal (25.0%)
CMP X7, X1	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 411760 <_Z10cg_calc_uriiidPdS_PKdS_S1_._omp_fn.0+0x134>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	cg_calc_ur(int, int, int, double, double, double, double const, double, double const*) [clone ._omp_fn.0]
Source file and lines	cg.cpp:108-113
Module	exec

The loop is defined in /home/eoseret/qaas/qaas_runs/178-237-4322/intel/TeaLeaf/build/TeaLeaf/src/omp/cg.cpp:108-113.

The related source loop is multi-versionned.

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 43% of vector register length is used (average across all VPU instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 4.00 to 1.60 cycles (2.50x speedup).

Details

75% of VPU instructions are used in vector version (process two or more data elements in vector registers):

0% of VPU addition or subtraction instructions are used in vector version.
0% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with fassociative-math (included in Ofast or ffast-math) to extend loop vectorization to FP reductions.
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

FMA

Detected 4 FMA (fused multiply-add) operations. 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).

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 12 FP arithmetical operations:

6: addition or subtraction (4 inside FMA instructions)
6: multiply (4 inside FMA instructions)

The binary loop does not load or store any data.

General properties

nb instructions	15
nb uops	15
loop length	60
used w registers	0
used x registers	6
used b registers	0
used h registers	0
used s registers	0
used d registers	3
used q registers	2
used v registers	3
used z registers	0
nb stack references	0
ADD-SUB / MUL ratio	2.00

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	1.88 cycles
front end	1.88 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	0.50	0.50	0.50	0.50	2.00	2.00	2.00	2.00	2.00	2.00	2.00	0.00	0.00
cycles	0.50	0.50	0.50	0.50	0.50	0.50	2.00	2.00	2.00	2.00	2.00	2.00	2.00	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)	4.00

Cycles summary

Front-end	1.88
Dispatch	2.00
Data deps.	4.00
Overall L1	4.00

Vectorization ratios

INT

all	85%
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	0%

FP

all	60%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	100%
add-sub	0%
fma	100%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	75%
load	100%
store	100%
mul	100%
add-sub	0%
fma	100%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Vector efficiency ratios

INT

all	46%
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	25%

FP

all	40%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	50%
add-sub	25%
fma	50%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	NA (no other vectorizable/vectorized instructions)

INT+FP

all	43%
load	50%
store	50%
mul	50%
add-sub	25%
fma	50%
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 iteration of the binary loop takes 4.00 cycles.

Front-end bottlenecks

Found no such bottlenecks.

ASM code

In the binary file, the address of the loop is: 411760

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput	Vectorization
LDR Q29, [X5, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q31, [X3, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMLA V31.2D, V29.2D, V28.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	4	0.25	vect (50.0%)
STR Q31, [X3, X1]	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 Q29, [X6, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
LDR Q31, [X2, X1]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	vect (50.0%)
FMLS V31.2D, V28.2D, V29.2D	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	4	0.25	vect (50.0%)
FMUL V29.2D, V31.2D, V31.2D	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 Q31, [X2, X1]	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 X1, X1, #16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
FADD D30, D30, D29	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	scal (25.0%)
MOV D31, V29.D[1]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	scal (25.0%)
FADD D30, D31, D30	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	scal (25.0%)
CMP X7, X1	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 411760 <_Z10cg_calc_uriiidPdS_PKdS_S1_._omp_fn.0+0x134>	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

Workaround

FMA

Workaround

Slow data structures access

Details

Workaround

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

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

Workaround

FMA

Workaround

Slow data structures access

Details

Workaround

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code