OV - exec - Loop 174

	Loop Id: 174	Module: exec	Source: advec_mom.cpp:45-48 [...]	Coverage: 2.72%

0x41d380 ADD	X20, X1, X3

0x41d384 ADD	W2, W6, W3

0x41d388 SDIV	X28, X20, X25

0x41d38c ADD	X21, X20, #1

0x41d390 ADD	W20, W7, W3

0x41d394 ADD	X3, X3, #2

0x41d398 MADD	W20, W5, W28, W20

0x41d39c SDIV	X30, X21, X25

0x41d3a0 ADD	W21, W24, W28

0x41d3a4 SBFM	X21, X21, #0, #31

0x41d3a8 ADD	W26, W24, W30

0x41d3ac MADD	X27, X9, X21, XZR

0x41d3b0 SBFM	X26, X26, #0, #31

0x41d3b4 MADD	X18, X9, X26, XZR

0x41d3b8 ADD	X4, X8, X27,LSL #3

0x41d3bc ADD	X18, X8, X18,LSL #3

0x41d3c0 LDR	D0, [X4, X20,SXTW #3]    [14]

0x41d3c4 ADD	W4, W22, W28

0x41d3c8 SBFM	X4, X4, #0, #31

0x41d3cc MADD	X4, X11, X4, XZR

0x41d3d0 ADD	X4, X10, X4,LSL #3

0x41d3d4 LDR	D1, [X4, X20,SXTW #3]    [13]

0x41d3d8 MADD	X4, X11, X26, XZR

0x41d3dc ADD	X4, X10, X4,LSL #3

0x41d3e0 MADD	W19, W5, W30, W2

0x41d3e4 SBFM	X27, X19, #61, #31

0x41d3e8 ADD	X18, X18, X27

0x41d3ec ADD	W30, W22, W30

0x41d3f0 LD1	{V0.D[1]}, [X18]    [4]

0x41d3f4 SBFM	X18, X30, #0, #31

0x41d3f8 MADD	X18, X11, X18, XZR

0x41d3fc ADD	X4, X4, X27

0x41d400 ADD	X18, X10, X18,LSL #3

0x41d404 ADD	X18, X18, X27

0x41d408 LD1	{V1.D[1]}, [X18]    [3]

0x41d40c MADD	X18, X11, X21, XZR

0x41d410 ADD	X18, X10, X18,LSL #3

0x41d414 FADD	V0.2D, V1.2D, V0.2D

0x41d418 LDR	D1, [X18, X20,SXTW #3]    [2]

0x41d41c MADD	X18, X13, X21, XZR

0x41d420 ADD	X18, X12, X18,LSL #3

0x41d424 LD1	{V1.D[1]}, [X4]    [8]

0x41d428 MADD	X4, X13, X26, XZR

0x41d42c ADD	X4, X12, X4,LSL #3

0x41d430 ADD	X4, X4, X27

0x41d434 FSUB	V0.2D, V0.2D, V1.2D

0x41d438 STR	D0, [X18, X20,SXTW #3]    [1]

0x41d43c ST1	{V0.D[1]}, [X4]    [7]

0x41d440 MADD	X4, X15, X21, XZR

0x41d444 MADD	W18, W5, W28, W2

0x41d448 ADD	W2, W19, #1

0x41d44c MADD	X19, X15, X26, XZR

0x41d450 ADD	X4, X14, X4,LSL #3

0x41d454 ADD	X19, X14, X19,LSL #3

0x41d458 ADD	X2, X19, W2,SXTW #3

0x41d45c LDR	D1, [X4, X18,SXTW #3]    [6]

0x41d460 LDR	D2, [X4, X20,SXTW #3]    [6]

0x41d464 ADD	X18, X19, X27

0x41d468 LD1	{V1.D[1]}, [X2]    [10]

0x41d46c LD1	{V2.D[1]}, [X18]    [12]

0x41d470 MADD	X18, X17, X21, XZR

0x41d474 MADD	X2, X17, X26, XZR

0x41d478 ADD	X18, X16, X18,LSL #3

0x41d47c ADD	X2, X16, X2,LSL #3

0x41d480 ADD	X2, X2, X27

0x41d484 FADD	V0.2D, V1.2D, V0.2D

0x41d488 FSUB	V0.2D, V0.2D, V2.2D

0x41d48c STR	D0, [X18, X20,SXTW #3]    [11]

0x41d490 LDR	X18, [SP, #40]    [5]

0x41d494 ST1	{V0.D[1]}, [X2]    [9]

0x41d498 CMP	X18, X3

0x41d49c B.NE	41d380

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp: 45 - 48

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

45:     for (int j = (y_min - 2 + 1); j < (y_max + 2 + 2); j++) {

46:       for (int i = (x_min - 2 + 1); i < (x_max + 2 + 2); i++) {

47:         post_vol(i, j) = volume(i, j) + vol_flux_y(i + 0, j + 1) - vol_flux_y(i, j);

48:         pre_vol(i, j) = post_vol(i, j) + vol_flux_x(i + 1, j + 0) - vol_flux_x(i, j);

/home/hbollore/qaas/qaas-runs/174-169-1878/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]; }

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	3.40 - 8.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.19 - 2.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	2.35
Bottlenecks	P8, P10,
Function	advec_mom_kernel(int, int, int, int, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer1D&, clover::Buffer1D&, int, int, int) [clone .omp_outlined]
Source	advec_mom.cpp:45-48,context.h:69-69
Source loop unroll info	unrolled by 3
Source loop unroll confidence level	high
Unroll/vectorization loop type	NA
Unroll factor	3
CQA cycles	17.00 - 40.00
CQA cycles if no scalar integer	5.00
CQA cycles if FP arith vectorized	17.00 - 40.00
CQA cycles if fully vectorized	14.25 - 20.00
Front-end cycles	9.00
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	7.00
P3 cycles	7.00
P4 cycles	7.00
P5 cycles	7.00
P6 cycles	17.00
P7 cycles	7.00
P8 cycles	3.25
P9 cycles	3.25
P10 cycles	3.25
P11 cycles	3.25
P12 cycles	5.00
P13 cycles	5.00
P14 cycles	5.00
P15 cycles	0.00
P16 cycles	0.00
DIV/SQRT cycles	10.00 - 40.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	72.00
Nb uops	72.00
Nb loads	NA
Nb stores	4.00
Nb stack references	0.00
FLOP/cycle	0.47 - 0.20
Nb FLOP add-sub	8.00
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	3.40 - 8.00
Bytes prefetched	0.00
Bytes loaded	88.00
Bytes stored	48.00
Stride 0	1.00
Stride 1	0.00
Stride n	0.00
Stride unknown	7.00
Stride indirect	6.00
Vectorization ratio all	12.90
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	44.44
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	0.00
Vectorization ratio other	0.00
Vector-efficiency ratio all	58.06
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	63.89
Vector-efficiency ratio fma	41.67
Vector-efficiency ratio div_sqrt	50.00
Vector-efficiency ratio other	100.00

Metric	Value
CQA speedup if no scalar integer	3.40 - 8.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	1.19 - 2.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	2.35
Bottlenecks	P8, P10,
Function	advec_mom_kernel(int, int, int, int, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer1D&, clover::Buffer1D&, int, int, int) [clone .omp_outlined]
Source	advec_mom.cpp:45-48,context.h:69-69
Source loop unroll info	unrolled by 3
Source loop unroll confidence level	high
Unroll/vectorization loop type	NA
Unroll factor	3
CQA cycles	17.00 - 40.00
CQA cycles if no scalar integer	5.00
CQA cycles if FP arith vectorized	17.00 - 40.00
CQA cycles if fully vectorized	14.25 - 20.00
Front-end cycles	9.00
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	7.00
P3 cycles	7.00
P4 cycles	7.00
P5 cycles	7.00
P6 cycles	17.00
P7 cycles	7.00
P8 cycles	3.25
P9 cycles	3.25
P10 cycles	3.25
P11 cycles	3.25
P12 cycles	5.00
P13 cycles	5.00
P14 cycles	5.00
P15 cycles	0.00
P16 cycles	0.00
DIV/SQRT cycles	10.00 - 40.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	72.00
Nb uops	72.00
Nb loads	NA
Nb stores	4.00
Nb stack references	0.00
FLOP/cycle	0.47 - 0.20
Nb FLOP add-sub	8.00
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	3.40 - 8.00
Bytes prefetched	0.00
Bytes loaded	88.00
Bytes stored	48.00
Stride 0	1.00
Stride 1	0.00
Stride n	0.00
Stride unknown	7.00
Stride indirect	6.00
Vectorization ratio all	12.90
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	44.44
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	0.00
Vectorization ratio other	0.00
Vector-efficiency ratio all	58.06
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	63.89
Vector-efficiency ratio fma	41.67
Vector-efficiency ratio div_sqrt	50.00
Vector-efficiency ratio other	100.00

Path /

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

Function	advec_mom_kernel(int, int, int, int, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer1D&, clover::Buffer1D&, int, int, int) [clone .omp_outlined]
Source file and lines	advec_mom.cpp:45-48
Module	exec

The loop is defined in:

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp:45-48
/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69

The related source loop is unrolled by 3 (including vectorization).

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 40.00 to 5.00 cycles (8.00x speedup).

Workaround

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

Vectorization

Your loop is not vectorized. Only 58% 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 40.00 to 20.00 cycles (2.00x speedup).

Details

12% 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.
0% of VPU stores are used in vector version.
44% of VPU addition or subtraction instructions are used in vector version.
0% of VPU fused multiply-add instructions are used in vector version.
0% of VPU divide and square root 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
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)

No data for this section

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

The binary loop is composed of 8 FP arithmetical operations:

8: addition or subtraction

The binary loop is loading 88 bytes. The binary loop is storing 48 bytes.

Arithmetic intensity

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

General properties

nb instructions	72
loop length	288
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

9.00 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	7.00	7.00	7.00	7.00	17.00	7.00	3.25	3.25	3.25	3.25	5.00	5.00	5.00	0.00	0.00
cycles	0.50	0.50	7.00	7.00	7.00	7.00	17.00	7.00	3.25	3.25	3.25	3.25	5.00	5.00	5.00	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	10.00-40.00
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	9.00
Data deps.	1.00
Overall L1	17.00-40.00

Vectorization ratios

INT

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

all	100%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
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	12%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	44%
fma	0%
div/sqrt	0%
other	0%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 40.00 cycles. At this rate:

2% of peak load performance is reached (2.20 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (1.20 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 41d380

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
ADD X20, X1, X3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD W2, W6, W3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
SDIV X28, X20, X25	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	5-20	5-20	scal (50.0%)
ADD X21, X20, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD W20, W7, W3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X3, X3, #2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD W20, W5, W28, W20	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
SDIV X30, X21, X25	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	5-20	5-20	scal (50.0%)
ADD W21, W24, W28	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
SBFM X21, X21, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
ADD W26, W24, W30	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
MADD X27, X9, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	scal (50.0%)
SBFM X26, X26, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
MADD X18, X9, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X8, X27,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X18, X8, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LDR D0, [X4, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
ADD W4, W22, W28	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
SBFM X4, X4, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD X4, X11, X4, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X10, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LDR D1, [X4, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
MADD X4, X11, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X10, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD W19, W5, W30, W2	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
SBFM X27, X19, #61, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
ADD X18, X18, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD W30, W22, W30	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
LD1 {V0.D[1]}, [X18]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
SBFM X18, X30, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD X18, X11, X18, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X4, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X18, X10, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X18, X18, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LD1 {V1.D[1]}, [X18]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
MADD X18, X11, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X18, X10, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
FADD V0.2D, V1.2D, V0.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
LDR D1, [X18, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
MADD X18, X13, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X18, X12, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LD1 {V1.D[1]}, [X4]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
MADD X4, X13, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X12, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X4, X4, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
FSUB V0.2D, V0.2D, V1.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
STR D0, [X18, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	scal (50.0%)
ST1 {V0.D[1]}, [X4]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	4	1	scal (50.0%)
MADD X4, X15, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
MADD W18, W5, W28, W2	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD W2, W19, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD X19, X15, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	scal (50.0%)
ADD X4, X14, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X19, X14, X19,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X2, X19, W2,SXTW #3	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50	N/A
LDR D1, [X4, X18,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
LDR D2, [X4, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
ADD X18, X19, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LD1 {V1.D[1]}, [X2]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
LD1 {V2.D[1]}, [X18]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
MADD X18, X17, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
MADD X2, X17, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X18, X16, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X2, X16, X2,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X2, X2, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
FADD V0.2D, V1.2D, V0.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
FSUB V0.2D, V0.2D, V2.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
STR D0, [X18, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	scal (50.0%)
LDR X18, [SP, #40]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33	N/A
ST1 {V0.D[1]}, [X2]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	4	1	scal (50.0%)
CMP X18, X3	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	1	0.33	N/A
B.NE 41d380 <_Z16advec_mom_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_S2_S2_S2_S2_S2_S2_S2_S2_S2_RNS_8Buffer1DIdEES5_iii.omp_outlined+0x130>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	advec_mom_kernel(int, int, int, int, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer2D&, clover::Buffer1D&, clover::Buffer1D&, int, int, int) [clone .omp_outlined]
Source file and lines	advec_mom.cpp:45-48
Module	exec

The loop is defined in:

/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp:45-48
/home/hbollore/qaas/qaas-runs/174-169-1878/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69

The related source loop is unrolled by 3 (including vectorization).

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 40.00 to 5.00 cycles (8.00x speedup).

Workaround

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

Vectorization

Details

12% 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.
0% of VPU stores are used in vector version.
44% of VPU addition or subtraction instructions are used in vector version.
0% of VPU fused multiply-add instructions are used in vector version.
0% of VPU divide and square root 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
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)

No data for this section

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

The binary loop is composed of 8 FP arithmetical operations:

8: addition or subtraction

The binary loop is loading 88 bytes. The binary loop is storing 48 bytes.

Arithmetic intensity

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

General properties

nb instructions	72
loop length	288
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

9.00 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	7.00	7.00	7.00	7.00	17.00	7.00	3.25	3.25	3.25	3.25	5.00	5.00	5.00	0.00	0.00
cycles	0.50	0.50	7.00	7.00	7.00	7.00	17.00	7.00	3.25	3.25	3.25	3.25	5.00	5.00	5.00	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	10.00-40.00
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	9.00
Data deps.	1.00
Overall L1	17.00-40.00

Vectorization ratios

INT

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

all	100%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
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	12%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	44%
fma	0%
div/sqrt	0%
other	0%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 40.00 cycles. At this rate:

2% of peak load performance is reached (2.20 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
2% of peak store performance is reached (1.20 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 41d380

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput	Vectorization
ADD X20, X1, X3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD W2, W6, W3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
SDIV X28, X20, X25	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	5-20	5-20	scal (50.0%)
ADD X21, X20, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD W20, W7, W3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X3, X3, #2	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD W20, W5, W28, W20	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
SDIV X30, X21, X25	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	5-20	5-20	scal (50.0%)
ADD W21, W24, W28	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
SBFM X21, X21, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
ADD W26, W24, W30	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
MADD X27, X9, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	scal (50.0%)
SBFM X26, X26, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
MADD X18, X9, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X8, X27,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X18, X8, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LDR D0, [X4, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
ADD W4, W22, W28	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
SBFM X4, X4, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD X4, X11, X4, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X10, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LDR D1, [X4, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
MADD X4, X11, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X10, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD W19, W5, W30, W2	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	scal (25.0%)
SBFM X27, X19, #61, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (100.0%)
ADD X18, X18, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD W30, W22, W30	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (25.0%)
LD1 {V0.D[1]}, [X18]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
SBFM X18, X30, #0, #31	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD X18, X11, X18, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X4, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X18, X10, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X18, X18, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LD1 {V1.D[1]}, [X18]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
MADD X18, X11, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X18, X10, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
FADD V0.2D, V1.2D, V0.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
LDR D1, [X18, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
MADD X18, X13, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X18, X12, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LD1 {V1.D[1]}, [X4]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
MADD X4, X13, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X4, X12, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X4, X4, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
FSUB V0.2D, V0.2D, V1.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
STR D0, [X18, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	scal (50.0%)
ST1 {V0.D[1]}, [X4]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	4	1	scal (50.0%)
MADD X4, X15, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
MADD W18, W5, W28, W2	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD W2, W19, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
MADD X19, X15, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	scal (50.0%)
ADD X4, X14, X4,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X19, X14, X19,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X2, X19, W2,SXTW #3	1	0	0	0	0	0	0	0.50	0.50	0	0	0	0	0	0	0	2	0.50	N/A
LDR D1, [X4, X18,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
LDR D2, [X4, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	6	0.33	scal (50.0%)
ADD X18, X19, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
LD1 {V1.D[1]}, [X2]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
LD1 {V2.D[1]}, [X18]	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0.33	0.33	0.33	8	0.33	scal (50.0%)
MADD X18, X17, X21, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
MADD X2, X17, X26, XZR	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	2	1	N/A
ADD X18, X16, X18,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X2, X16, X2,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
ADD X2, X2, X27	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	1	0.17	N/A
FADD V0.2D, V1.2D, V0.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
FSUB V0.2D, V0.2D, V2.2D	1	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	2	0.25	vect (100.0%)
STR D0, [X18, X20,SXTW #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	2	0.50	scal (50.0%)
LDR X18, [SP, #40]	1	0	0	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	4	0.33	N/A
ST1 {V0.D[1]}, [X2]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	4	1	scal (50.0%)
CMP X18, X3	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	0	0	1	0.33	N/A
B.NE 41d380 <_Z16advec_mom_kerneliiiiRN6clover8Buffer2DIdEES2_S2_S2_S2_S2_S2_S2_S2_S2_S2_S2_S2_RNS_8Buffer1DIdEES5_iii.omp_outlined+0x130>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Report Configuration

Code clean check

Workaround

Vectorization

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

Cycles and memory resources usage

ASM code

Code clean check

Workaround

Vectorization

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

Cycles and memory resources usage

ASM code