OV - exec - Loop 205

	Loop Id: 205	Module: exec	Source: context.h:69-69 [...]	Coverage: 0.04%

0x42ef80 SEL	Z20.D, P1, Z4.D, Z6.D

0x42ef84 LDR	D22, [X8]    [3]

0x42ef88 ADD	X14, X14, X9

0x42ef8c SDIVR	Z20.D, P0/M, Z20.D, Z19.D

0x42ef90 DUP	Z22.D, Z22.D[0]

0x42ef94 MOVPRFX	Z21, Z19

0x42ef98 MLS	Z21.D, P0/M, Z20.D, Z4.D

0x42ef9c SXTW	Z20.D, P0/M, Z20.D

0x42efa0 ADD	Z19.D, Z19.D, Z5.D

0x42efa4 MOVPRFX	Z23, Z0

0x42efa8 MUL	Z23.D, P0/M, Z23.D, Z20.D

0x42efac ADR	Z24.D, [Z7, Z21.D,SXTW #3]    [2]

0x42efb0 ADR	Z23.D, [Z24, Z23.D,LSL #3]    [10]

0x42efb4 ADR	Z24.D, [Z16, Z21.D,SXTW #3]    [11]

0x42efb8 ST1D	{Z22.D}, P1, [V23.D]    [7]

0x42efbc MOVPRFX	Z23, Z1

0x42efc0 MUL	Z23.D, P0/M, Z23.D, Z20.D

0x42efc4 LDR	D22, [X10]    [12]

0x42efc8 ADR	Z23.D, [Z24, Z23.D,LSL #3]    [9]

0x42efcc ADR	Z24.D, [Z17, Z21.D,SXTW #3]    [13]

0x42efd0 ADR	Z21.D, [Z18, Z21.D,SXTW #3]    [15]

0x42efd4 DUP	Z22.D, Z22.D[0]

0x42efd8 ST1D	{Z22.D}, P1, [V23.D]    [6]

0x42efdc MOVPRFX	Z23, Z2

0x42efe0 MUL	Z23.D, P0/M, Z23.D, Z20.D

0x42efe4 LDR	D22, [X11]    [14]

0x42efe8 MUL	Z20.D, P0/M, Z20.D, Z3.D

0x42efec ADR	Z23.D, [Z24, Z23.D,LSL #3]    [8]

0x42eff0 ADR	Z20.D, [Z21, Z20.D,LSL #3]    [4]

0x42eff4 DUP	Z22.D, Z22.D[0]

0x42eff8 ST1D	{Z22.D}, P1, [V23.D]    [5]

0x42effc LDR	D22, [X12]    [16]

0x42f000 DUP	Z22.D, Z22.D[0]

0x42f004 ST1D	{Z22.D}, P1, [V20.D]    [1]

0x42f008 WHILELO	P1.D, X14, X13

0x42f00c B.MI	42ef80

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h: 69 - 69

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

69:   T &operator()(size_t i, size_t j) const { return data[i + j * sizeX]; }

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/generate_chunk.cpp: 75 - 80

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

75:   for (int j = (0); j < (yrange); j++) {

76:     for (int i = (0); i < (xrange); i++) {

77:       field.energy0(i, j) = state_energy[0];

78:       field.density0(i, j) = state_density[0];

79:       field.xvel0(i, j) = state_xvel[0];

80:       field.yvel0(i, j) = state_yvel[0];

Coverage (%)	Name	Source Location	Module
►98.44+	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so
○	__kmp_launch_thread		libomp.so
○	__kmp_launch_worker(void*)		libomp.so
○	start_thread		libc.so.6
○	thread_start		libc.so.6
►1.56+	__kmp_invoke_microtask		libomp.so
○	__kmp_invoke_task_func		libomp.so
○	__kmp_fork_call		libomp.so
○	__kmpc_fork_call		libomp.so
○	generate_chunk(int, global_var[...]	generate_chunk.cpp:74	exec
○	start(parallel_&, global_confi[...]	start.cpp:81	exec
○	initialise(parallel_&, std::ve[...]	clover_leaf.cpp:192	exec
○	main	stl_vector.h:735	exec
○	__libc_start_call_main		libc.so.6
○	__libc_start_main		libc.so.6
○	_start	ostream:93	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	1.32 - 1.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	2.78 - 1.60
Bottlenecks	P6, P8,
Function	generate_chunk(int, global_variables&) [clone .omp_outlined]
Source	context.h:69-69,generate_chunk.cpp:75-80
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	12.50 - 20.00
CQA cycles if no scalar integer	12.50 - 20.00
CQA cycles if FP arith vectorized	12.50 - 20.00
CQA cycles if fully vectorized	9.50 - 20.00
Front-end cycles	4.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.25
P4 cycles	1.00
P5 cycles	0.25
P6 cycles	12.50
P7 cycles	12.50
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	2.67
P11 cycles	2.67
P12 cycles	2.67
P13 cycles	0.00
P14 cycles	0.00
DIV/SQRT cycles	0.87 - 20.00
Inter-iter dependencies cycles	2
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	36.00
Nb uops	36.00
Nb loads	NA
Nb stores	4.00
Nb stack references	0.00
FLOP/cycle	0.00 - 0.00
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
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	6.45 - 10.32
Bytes prefetched	0.00
Bytes loaded	1.00
Bytes stored	128.00
Stride 0	4.00
Stride 1	0.00
Stride n	0.00
Stride unknown	8.00
Stride indirect	4.00
Vectorization ratio all	51.52
Vectorization ratio load	0.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	20.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	100.00
Vectorization ratio other	60.00
Vector-efficiency ratio all	72.73
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	100.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	40.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	100.00
Vector-efficiency ratio other	100.00

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.32 - 1.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	2.78 - 1.60
Bottlenecks	P6, P8,
Function	generate_chunk(int, global_variables&) [clone .omp_outlined]
Source	context.h:69-69,generate_chunk.cpp:75-80
Source loop unroll info	not unrolled or unrolled with no peel/tail loop
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	12.50 - 20.00
CQA cycles if no scalar integer	12.50 - 20.00
CQA cycles if FP arith vectorized	12.50 - 20.00
CQA cycles if fully vectorized	9.50 - 20.00
Front-end cycles	4.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.25
P4 cycles	1.00
P5 cycles	0.25
P6 cycles	12.50
P7 cycles	12.50
P8 cycles	2.00
P9 cycles	2.00
P10 cycles	2.67
P11 cycles	2.67
P12 cycles	2.67
P13 cycles	0.00
P14 cycles	0.00
DIV/SQRT cycles	0.87 - 20.00
Inter-iter dependencies cycles	2
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	36.00
Nb uops	36.00
Nb loads	NA
Nb stores	4.00
Nb stack references	0.00
FLOP/cycle	0.00 - 0.00
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
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	6.45 - 10.32
Bytes prefetched	0.00
Bytes loaded	1.00
Bytes stored	128.00
Stride 0	4.00
Stride 1	0.00
Stride n	0.00
Stride unknown	8.00
Stride indirect	4.00
Vectorization ratio all	51.52
Vectorization ratio load	0.00
Vectorization ratio store	100.00
Vectorization ratio mul	100.00
Vectorization ratio add_sub	20.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	100.00
Vectorization ratio other	60.00
Vector-efficiency ratio all	72.73
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	100.00
Vector-efficiency ratio mul	100.00
Vector-efficiency ratio add_sub	40.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	100.00
Vector-efficiency ratio other	100.00

Path /

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

Function	generate_chunk(int, global_variables&) [clone .omp_outlined]
Source file and lines	context.h:69-69
Module	exec

The loop is defined in:

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/generate_chunk.cpp:75-80

The related source loop is not unrolled or unrolled with no peel/tail loop.

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 72% of vector register length is used (average across all VPU instructions).

Details

51% of VPU instructions are used in vector version (process two or more data elements in vector registers):

0% of VPU loads are used in vector version.
20% of VPU addition or subtraction instructions are used in vector version.
60% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

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 12.00 cycles (1.67x speedup).

Slow data structures access

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

Details

Constant unknown stride: 8 occurrence(s)
Irregular (variable stride) or indirect: 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)
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

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

The binary loop does not contain any FP arithmetical operations. The binary loop is loading 1 bytes. The binary loop is storing 128 bytes.

General properties

nb instructions	36
nb uops	36
loop length	144
used w registers	0
used x registers	7
used b registers	0
used h registers	0
used s registers	0
used d registers	1
used q registers	0
used v registers	2
used z registers	17
nb stack references	0

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	0.50	0.50	0.50	0.25	1.00	0.25	12.50	12.50	2.00	2.00	2.67	2.67	2.67	0.00	0.00
cycles	0.50	0.50	0.50	0.25	1.00	0.25	12.50	12.50	2.00	2.00	2.67	2.67	2.67	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	0.87-20.00
Longest recurrence chain latency (RecMII)	2.00

Cycles summary

Front-end	4.50
Dispatch	12.50
DIV/SQRT	0.87-20.00
Data deps.	2.00
Overall L1	12.50-20.00

Vectorization ratios

all	51%
load	0%
store	100%
mul	100%
add-sub	20%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	100%
other	60%

Vector efficiency ratios

all	72%
load	25%
store	100%
mul	100%
add-sub	40%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	100%
other	100%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 20.00 cycles. At this rate:

0% of peak load performance is reached (0.05 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
13% of peak store performance is reached (6.40 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: 42ef80

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput	Vectorization
SEL Z20.D, P1, Z4.D, Z6.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
LDR D22, [X8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
ADD X14, X14, X9	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
SDIVR Z20.D, P0/M, Z20.D, Z19.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	7-20	0.87-20	vect (100.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
MOVPRFX Z21, Z19	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MLS Z21.D, P0/M, Z20.D, Z4.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
SXTW Z20.D, P0/M, Z20.D	1	0	0	0	0	0	0	0	1	0	0	0	0	0	2	1	vect (100.0%)
ADD Z19.D, Z19.D, Z5.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
MOVPRFX Z23, Z0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MUL Z23.D, P0/M, Z23.D, Z20.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
ADR Z24.D, [Z7, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z23.D, [Z24, Z23.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z24.D, [Z16, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ST1D {Z22.D}, P1, [V23.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
MOVPRFX Z23, Z1	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MUL Z23.D, P0/M, Z23.D, Z20.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
LDR D22, [X10]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
ADR Z23.D, [Z24, Z23.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z24.D, [Z17, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z21.D, [Z18, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
ST1D {Z22.D}, P1, [V23.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
MOVPRFX Z23, Z2	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MUL Z23.D, P0/M, Z23.D, Z20.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
LDR D22, [X11]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
MUL Z20.D, P0/M, Z20.D, Z3.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
ADR Z23.D, [Z24, Z23.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z20.D, [Z21, Z20.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
ST1D {Z22.D}, P1, [V23.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
LDR D22, [X12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
ST1D {Z22.D}, P1, [V20.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
WHILELO P1.D, X14, X13	1	0	0	0	0	1	0	0	0	0	0	0	0	0	3	2	N/A
B.MI 42ef80 <_Z14generate_chunkiR16global_variables.omp_outlined+0x120>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	generate_chunk(int, global_variables&) [clone .omp_outlined]
Source file and lines	context.h:69-69
Module	exec

The loop is defined in:

/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
/home/eoseret/qaas/qaas_runs/178-219-7589/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/generate_chunk.cpp:75-80

The related source loop is not unrolled or unrolled with no peel/tail loop.

gain
potential
hint
expert

Vectorization

Your loop is partially vectorized. Only 72% of vector register length is used (average across all VPU instructions).

Details

51% of VPU instructions are used in vector version (process two or more data elements in vector registers):

0% of VPU loads are used in vector version.
20% of VPU addition or subtraction instructions are used in vector version.
60% of VPU instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

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 12.00 cycles (1.67x speedup).

Slow data structures access

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

Details

Constant unknown stride: 8 occurrence(s)
Irregular (variable stride) or indirect: 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)
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.

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

The binary loop does not contain any FP arithmetical operations. The binary loop is loading 1 bytes. The binary loop is storing 128 bytes.

General properties

nb instructions	36
nb uops	36
loop length	144
used w registers	0
used x registers	7
used b registers	0
used h registers	0
used s registers	0
used d registers	1
used q registers	0
used v registers	2
used z registers	17
nb stack references	0

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	0.50	0.50	0.50	0.25	1.00	0.25	12.50	12.50	2.00	2.00	2.67	2.67	2.67	0.00	0.00
cycles	0.50	0.50	0.50	0.25	1.00	0.25	12.50	12.50	2.00	2.00	2.67	2.67	2.67	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	0.87-20.00
Longest recurrence chain latency (RecMII)	2.00

Cycles summary

Front-end	4.50
Dispatch	12.50
DIV/SQRT	0.87-20.00
Data deps.	2.00
Overall L1	12.50-20.00

Vectorization ratios

all	51%
load	0%
store	100%
mul	100%
add-sub	20%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	100%
other	60%

Vector efficiency ratios

all	72%
load	25%
store	100%
mul	100%
add-sub	40%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	100%
other	100%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 20.00 cycles. At this rate:

0% of peak load performance is reached (0.05 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
13% of peak store performance is reached (6.40 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: 42ef80

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	Latency	Recip. throughput	Vectorization
SEL Z20.D, P1, Z4.D, Z6.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
LDR D22, [X8]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
ADD X14, X14, X9	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	1	0.25	N/A
SDIVR Z20.D, P0/M, Z20.D, Z19.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	7-20	0.87-20	vect (100.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
MOVPRFX Z21, Z19	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MLS Z21.D, P0/M, Z20.D, Z4.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
SXTW Z20.D, P0/M, Z20.D	1	0	0	0	0	0	0	0	1	0	0	0	0	0	2	1	vect (100.0%)
ADD Z19.D, Z19.D, Z5.D	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
MOVPRFX Z23, Z0	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MUL Z23.D, P0/M, Z23.D, Z20.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
ADR Z24.D, [Z7, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z23.D, [Z24, Z23.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z24.D, [Z16, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ST1D {Z22.D}, P1, [V23.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
MOVPRFX Z23, Z1	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MUL Z23.D, P0/M, Z23.D, Z20.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
LDR D22, [X10]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
ADR Z23.D, [Z24, Z23.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z24.D, [Z17, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z21.D, [Z18, Z21.D,SXTW #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
ST1D {Z22.D}, P1, [V23.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
MOVPRFX Z23, Z2	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (100.0%)
MUL Z23.D, P0/M, Z23.D, Z20.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
LDR D22, [X11]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
MUL Z20.D, P0/M, Z20.D, Z3.D	1	0	0	0	0	0	0	1	0	0	0	0	0	0	5	2	vect (100.0%)
ADR Z23.D, [Z24, Z23.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
ADR Z20.D, [Z21, Z20.D,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	scal (25.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
ST1D {Z22.D}, P1, [V23.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
LDR D22, [X12]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (25.0%)
DUP Z22.D, Z22.D[0]	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	2	0.50	vect (100.0%)
ST1D {Z22.D}, P1, [V20.D]	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.50	0.50	0	6	2	vect (100.0%)
WHILELO P1.D, X14, X13	1	0	0	0	0	1	0	0	0	0	0	0	0	0	3	2	N/A
B.MI 42ef80 <_Z14generate_chunkiR16global_variables.omp_outlined+0x120>	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

Expensive FP math instructions/calls

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

Expensive FP math instructions/calls

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