OV - exec - Loop 572

	Loop Id: 572	Module: exec	Source: update_halo.cpp:582-583 [...]	Coverage: 0.02%

0x4347a0 MADD	X12, X20, X4, XZR

0x4347a4 ADD	X13, X21, X12

0x4347a8 ADD	X0, X17, X12,LSL #3

0x4347ac UBFM	X14, X12, #60, #59

0x4347b0 ADD	X5, X18, X13,LSL #3

0x4347b4 ADD	X1, X14, #8

0x4347b8 SUB	X15, X5, X0

0x4347bc SUB	X16, X15, #8

0x4347c0 UBFM	X30, X16, #3, #63

0x4347c4 ADD	X2, X30, #1

0x4347c8 ANDS	X6, X2, #4224

0x4347cc B.EQ	434894

0x4347d0 CMP	X6, #1

0x4347d4 B.EQ	434878

0x4347d8 CMP	X6, #2

0x4347dc B.EQ	434864

0x4347e0 CMP	X6, #3

0x4347e4 B.EQ	434850

0x4347e8 CMP	X6, #4

0x4347ec B.EQ	43483c

0x4347f0 CMP	X6, #5

0x4347f4 B.EQ	434828

0x4347f8 CMP	X6, #6

0x4347fc B.EQ	434814

0x434800 LDR	D0, [X0, #24]

0x434804 SUB	X7, X1, X0

0x434808 ADD	X0, X0, #8

0x43480c FNEG	D1, D0

0x434810 STR	D1, [X7, X3]

0x434814 LDR	D2, [X0, #24]

0x434818 SUB	X8, X1, X0

0x43481c ADD	X0, X0, #8

0x434820 FNEG	D3, D2

0x434824 STR	D3, [X8, X3]

0x434828 LDR	D4, [X0, #24]

0x43482c SUB	X9, X1, X0

0x434830 ADD	X0, X0, #8

0x434834 FNEG	D5, D4

0x434838 STR	D5, [X9, X3]

0x43483c LDR	D6, [X0, #24]

0x434840 SUB	X10, X1, X0

0x434844 ADD	X0, X0, #8

0x434848 FNEG	D7, D6

0x43484c STR	D7, [X10, X3]

0x434850 LDR	D16, [X0, #24]

0x434854 SUB	X11, X1, X0

0x434858 ADD	X0, X0, #8

0x43485c FNEG	D17, D16

0x434860 STR	D17, [X11, X3]

0x434864 LDR	D18, [X0, #24]

0x434868 SUB	X12, X1, X0

0x43486c ADD	X0, X0, #8

0x434870 FNEG	D19, D18

0x434874 STR	D19, [X12, X3]

0x434878 LDR	D20, [X0, #24]

0x43487c SUB	X13, X1, X0

0x434880 ADD	X0, X0, #8

0x434884 FNEG	D21, D20

0x434888 STR	D21, [X13, X3]

0x43488c CMP	X5, X0

0x434890 B.EQ	434944

0x434894 STP	X23, X24, [SP, #48]

  (571) 0x434898 ADD	X2, X0, #8

  (571) 0x43489c SUB	X24, X1, X0

  (571) 0x4348a0 SUB	X23, X1, X2

  (571) 0x4348a4 ADD	X14, X0, #16

  (571) 0x4348a8 LDR	D22, [X0, #24]

  (571) 0x4348ac ADD	X10, X0, #24

  (571) 0x4348b0 SUB	X30, X1, X14

  (571) 0x4348b4 ADD	X9, X0, #32

  (571) 0x4348b8 SUB	X16, X1, X10

  (571) 0x4348bc ADD	X8, X0, #40

  (571) 0x4348c0 SUB	X15, X1, X9

  (571) 0x4348c4 ADD	X7, X0, #48

  (571) 0x4348c8 SUB	X11, X1, X8

  (571) 0x4348cc ADD	X6, X0, #56

  (571) 0x4348d0 SUB	X13, X1, X7

  (571) 0x4348d4 SUB	X12, X1, X6

  (571) 0x4348d8 ADD	X0, X0, #64

  (571) 0x4348dc FNEG	D23, D22

  (571) 0x4348e0 STR	D23, [X24, X3]

  (571) 0x4348e4 LDR	D24, [X2, #24]

  (571) 0x4348e8 FNEG	D25, D24

  (571) 0x4348ec STR	D25, [X23, X3]

  (571) 0x4348f0 LDR	D26, [X2, #32]

  (571) 0x4348f4 FNEG	D27, D26

  (571) 0x4348f8 STR	D27, [X30, X3]

  (571) 0x4348fc LDR	D28, [X10, #24]

  (571) 0x434900 FNEG	D29, D28

  (571) 0x434904 STR	D29, [X16, X3]

  (571) 0x434908 LDR	D30, [X9, #24]

  (571) 0x43490c FNEG	D31, D30

  (571) 0x434910 STR	D31, [X15, X3]

  (571) 0x434914 LDR	D0, [X8, #24]

  (571) 0x434918 FNEG	D1, D0

  (571) 0x43491c STR	D1, [X11, X3]

  (571) 0x434920 LDR	D2, [X7, #24]

  (571) 0x434924 FNEG	D3, D2

  (571) 0x434928 STR	D3, [X13, X3]

  (571) 0x43492c LDR	D4, [X6, #24]

  (571) 0x434930 FNEG	D5, D4

  (571) 0x434934 STR	D5, [X12, X3]

  (571) 0x434938 CMP	X5, X0

  (571) 0x43493c B.NE	434898

0x434940 LDP	X23, X24, [SP, #48]

0x434944 ADD	X4, X4, #1

0x434948 ADD	W5, W22, W4

0x43494c CMP	W19, W5

0x434950 B.GT	4347a0

/home/hbollore/qaas-runs/170-300-2310/intel/CloverLeafCXX/build/CloverLeafCXX/src/omp/context.h: 69 - 69

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

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

/home/hbollore/qaas-runs/170-300-2310/intel/CloverLeafCXX/build/CloverLeafCXX/src/omp/update_halo.cpp: 582 - 583

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

582:         for (int j = 0; j < depth; ++j) {

583:           field.mass_flux_x(1 - j, k) = -field.mass_flux_x(1 + 2 + j, k);

Path /

Metric	Value
CQA speedup if no scalar integer	1.81
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P2, P3, P4, P5,
Function	update_halo_kernel(int, int, int, int, std::array const&, std::array const&, field_type&, int const*, int) [clone ._omp_fn.50]
Source	context.h:69-69,update_halo.cpp:582-583
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	8.75
CQA cycles if no scalar integer	4.83
CQA cycles if FP arith vectorized	8.75
CQA cycles if fully vectorized	2.19
Front-end cycles	8.38
DIV/SQRT cycles	4.50
P0 cycles	4.50
P1 cycles	8.75
P2 cycles	8.75
P3 cycles	8.75
P4 cycles	8.75
P5 cycles	3.50
P6 cycles	3.50
P7 cycles	3.50
P8 cycles	3.50
P9 cycles	5.33
P10 cycles	5.33
P11 cycles	5.33
P12 cycles	0.50
P13 cycles	0.50
P14 cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	67.00
Nb uops	67.00
Nb loads	NA
Nb stores	8.00
Nb stack references	0.00
FLOP/cycle	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	16.46
Bytes prefetched	0.00
Bytes loaded	72.00
Bytes stored	72.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	25.00

Metric	Value
CQA speedup if no scalar integer	1.81
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P2, P3, P4, P5,
Function	update_halo_kernel(int, int, int, int, std::array const&, std::array const&, field_type&, int const*, int) [clone ._omp_fn.50]
Source	context.h:69-69,update_halo.cpp:582-583
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	8.75
CQA cycles if no scalar integer	4.83
CQA cycles if FP arith vectorized	8.75
CQA cycles if fully vectorized	2.19
Front-end cycles	8.38
DIV/SQRT cycles	4.50
P0 cycles	4.50
P1 cycles	8.75
P2 cycles	8.75
P3 cycles	8.75
P4 cycles	8.75
P5 cycles	3.50
P6 cycles	3.50
P7 cycles	3.50
P8 cycles	3.50
P9 cycles	5.33
P10 cycles	5.33
P11 cycles	5.33
P12 cycles	0.50
P13 cycles	0.50
P14 cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	67.00
Nb uops	67.00
Nb loads	NA
Nb stores	8.00
Nb stack references	0.00
FLOP/cycle	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	16.46
Bytes prefetched	0.00
Bytes loaded	72.00
Bytes stored	72.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	0.00
Vectorization ratio store	0.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	25.00
Vector-efficiency ratio store	25.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	NA
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	update_halo_kernel(int, int, int, int, std::array const&, std::array const&, field_type&, int const*, int) [clone ._omp_fn.50]
Source file and lines	update_halo.cpp:582-583
Module	exec

The loop is defined in:

/home/hbollore/qaas-runs/170-300-2310/intel/CloverLeafCXX/build/CloverLeafCXX/src/omp/context.h:69
/home/hbollore/qaas-runs/170-300-2310/intel/CloverLeafCXX/build/CloverLeafCXX/src/omp/update_halo.cpp:582-583

Warnings:

Non-innermost loop: analyzing only self part (ignoring child loops).
Ignoring paths for analysis
Too many paths. Rerun with max-paths=15
RecMII not computed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 15 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

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 8.75 to 4.83 cycles (1.81x 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. 4 data elements could be processed at once in vector registers. By vectorizing your loop, you can lower the cost of an iteration from 8.75 to 2.19 cycles (4.00x speedup).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a 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 does not contain any FP arithmetical operations. The binary loop is loading 72 bytes. The binary loop is storing 72 bytes.

General properties

nb instructions	67
loop length	268
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

8.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	4.50	4.50	8.75	8.75	8.75	8.75	3.50	3.50	3.50	3.50	5.33	5.33	5.33	0.50	0.50
cycles	4.50	4.50	8.75	8.75	8.75	8.75	3.50	3.50	3.50	3.50	5.33	5.33	5.33	0.50	0.50

Execution ports to units layout:

P0:
P1:
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

Cycles summary

Front-end	8.38
Overall L1	8.75

Vectorization ratios

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.75 cycles. At this rate:

8% of peak load performance is reached (8.23 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
17% of peak store performance is reached (8.23 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 4347a0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
MADD X12, X20, X4, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
ADD X13, X21, X12	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X17, X12,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
UBFM X14, X12, #60, #59	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X5, X18, X13,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X1, X14, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
SUB X15, X5, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
SUB X16, X15, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
UBFM X30, X16, #3, #63	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X2, X30, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ANDS X6, X2, #4224	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434894 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x194>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434878 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x178>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434864 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x164>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434850 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 43483c <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x13c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434828 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x128>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #6	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434814 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x114>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR D0, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X7, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D1, D0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D1, [X7, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D2, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X8, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D3, D2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D3, [X8, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D4, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X9, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D5, D4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D5, [X9, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D6, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X10, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D7, D6	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D7, [X10, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D16, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X11, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D17, D16	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D17, [X11, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D18, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X12, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D19, D18	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D19, [X12, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D20, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X13, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D21, D20	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D21, [X13, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
CMP X5, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434944 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x244>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
STP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
LDP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
ADD X4, X4, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD W5, W22, W4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP W19, W5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.GT 4347a0 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0xa0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50

Function	update_halo_kernel(int, int, int, int, std::array const&, std::array const&, field_type&, int const*, int) [clone ._omp_fn.50]
Source file and lines	update_halo.cpp:582-583
Module	exec

The loop is defined in:

/home/hbollore/qaas-runs/170-300-2310/intel/CloverLeafCXX/build/CloverLeafCXX/src/omp/context.h:69
/home/hbollore/qaas-runs/170-300-2310/intel/CloverLeafCXX/build/CloverLeafCXX/src/omp/update_halo.cpp:582-583

Warnings:

Non-innermost loop: analyzing only self part (ignoring child loops).
Ignoring paths for analysis
Too many paths. Rerun with max-paths=15
RecMII not computed since number of paths is unknown or > max_paths

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

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 8.75 to 4.83 cycles (1.81x 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. 4 data elements could be processed at once in vector registers. By vectorizing your loop, you can lower the cost of an iteration from 8.75 to 2.19 cycles (4.00x speedup).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a 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 does not contain any FP arithmetical operations. The binary loop is loading 72 bytes. The binary loop is storing 72 bytes.

General properties

nb instructions	67
loop length	268
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

8.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	4.50	4.50	8.75	8.75	8.75	8.75	3.50	3.50	3.50	3.50	5.33	5.33	5.33	0.50	0.50
cycles	4.50	4.50	8.75	8.75	8.75	8.75	3.50	3.50	3.50	3.50	5.33	5.33	5.33	0.50	0.50

Execution ports to units layout:

P0:
P1:
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

Cycles summary

Front-end	8.38
Overall L1	8.75

Vectorization ratios

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

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 8.75 cycles. At this rate:

8% of peak load performance is reached (8.23 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
17% of peak store performance is reached (8.23 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 4347a0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
MADD X12, X20, X4, XZR	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
ADD X13, X21, X12	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X17, X12,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
UBFM X14, X12, #60, #59	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X5, X18, X13,LSL #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X1, X14, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
SUB X15, X5, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
SUB X16, X15, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
UBFM X30, X16, #3, #63	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X2, X30, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ANDS X6, X2, #4224	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434894 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x194>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434878 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x178>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #2	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434864 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x164>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #3	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434850 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x150>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 43483c <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x13c>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434828 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x128>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
CMP X6, #6	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434814 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x114>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
LDR D0, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X7, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D1, D0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D1, [X7, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D2, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X8, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D3, D2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D3, [X8, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D4, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X9, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D5, D4	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D5, [X9, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D6, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X10, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D7, D6	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D7, [X10, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D16, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X11, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D17, D16	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D17, [X11, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D18, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X12, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D19, D18	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D19, [X12, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
LDR D20, [X0, #24]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	6	0.33
SUB X13, X1, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, #8	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
FNEG D21, D20	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	2	0.25
STR D21, [X13, X3]	1	0	0	0	0	0	0	0.50	0.50	0	0	0.50	0.50	0	0	0	2	0.50
CMP X5, X0	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.EQ 434944 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0x244>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50
STP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
LDP X23, X24, [SP, #48]	1	0	0	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	1
ADD X4, X4, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD W5, W22, W4	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP W19, W5	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.GT 4347a0 <_Z18update_halo_kerneliiiiRKSt5arrayIiLm4EES2_R10field_typePKii._omp_fn.50+0xa0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

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

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

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code