OV - exec - Loop 1206

0x437be0 LDR	X1, [X11, X16,LSL #3]    [1]

0x437be4 ADD	X16, X16, #1

0x437be8 SUBS	X18, X18, #1

0x437bec LDR	X1, [X12, X1,LSL #3]    [3]

0x437bf0 STR	X26, [X19, X1,LSL #3]    [2]

0x437bf4 B.NE	437be0

/home/hbollore/qaas-runs/171-233-5044/intel/AMG/build/AMG/AMG/parcsr_ls/par_multi_interp.c: 1612 - 1615

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

1612:              for (j=j_start; j < j_end; j++)

1613:              {

1614:                 k1 = P_diag_pass[1][j];

1615:                 tmp_marker[C_array[k1]] = i1;

Coverage (%)	Name	Source Location	Module

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	NA
CQA speedup if fully vectorized	NA
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.17
Bottlenecks	P10,
Function	.omp_outlined..20#0x437930
Source	par_multi_interp.c:1612-1615
Source loop unroll info	unrolled by 4
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	1
CQA cycles	1.17
CQA cycles if no scalar integer	1.17
CQA cycles if FP arith vectorized	NA
CQA cycles if fully vectorized	NA
Front-end cycles	0.75
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.50
P5 cycles	0.00
P6 cycles	0.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	1.17
P10 cycles	0.83
P11 cycles	1.00
P12 cycles	0.50
P13 cycles	0.50
P14 cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	6.00
Nb uops	6.00
Nb loads	NA
Nb stores	1.00
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	20.57
Bytes prefetched	0.00
Bytes loaded	16.00
Bytes stored	8.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	NA
Vectorization ratio load	NA
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	NA
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	NA

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	NA
CQA speedup if fully vectorized	NA
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.17
Bottlenecks	P10,
Function	.omp_outlined..20#0x437930
Source	par_multi_interp.c:1612-1615
Source loop unroll info	unrolled by 4
Source loop unroll confidence level	max
Unroll/vectorization loop type	NA
Unroll factor	1
CQA cycles	1.17
CQA cycles if no scalar integer	1.17
CQA cycles if FP arith vectorized	NA
CQA cycles if fully vectorized	NA
Front-end cycles	0.75
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	0.50
P2 cycles	0.50
P3 cycles	0.50
P4 cycles	0.50
P5 cycles	0.00
P6 cycles	0.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	1.17
P10 cycles	0.83
P11 cycles	1.00
P12 cycles	0.50
P13 cycles	0.50
P14 cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	6.00
Nb uops	6.00
Nb loads	NA
Nb stores	1.00
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	20.57
Bytes prefetched	0.00
Bytes loaded	16.00
Bytes stored	8.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	NA
Vectorization ratio load	NA
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	NA
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	NA
Vector-efficiency ratio all	NA
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	NA
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	NA

Path /

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

Function	.omp_outlined..20#0x437930
Source file and lines	par_multi_interp.c:1612-1615
Module	exec

The loop is defined in /home/hbollore/qaas-runs/171-233-5044/intel/AMG/build/AMG/AMG/parcsr_ls/par_multi_interp.c:1612-1615.

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

gain
potential
hint
expert

Unrolling/vectorization cost

This loop is peel/tail of a unrolled/vectorized loop. If its cost is not negligible compared to the main (unrolled/vectorized) loop, unrolling/vectorization is counterproductive due to low trip count.

Details

The more iterations the main loop is processing, the higher the trip count must be to amortize peel/tail overhead.

Workaround

use profile-guided optimization flags of your compiler
hardcode most frequent values of loop bounds by adding specialized paths.:
- For instance, replace for (i=0; i<n; i++) foo(i) with: switch (n) { case (4): for (i=0; i<4; i++) foo(i); break; case (6): for (i=0; i<6; i++) foo(i); break; default : for (i=0; i<n; i++) foo(i); break; }

Vectorization

Your loop is not vectorized.

Details

No VPU instruction: only general purpose registers are used and data elements are processed one at a time.

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 16 bytes. The binary loop is storing 8 bytes.

General properties

nb instructions	6
loop length	24
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

0.75 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	0.00	0.00	0.00	0.00	1.17	0.83	1.00	0.50	0.50
cycles	0.50	0.50	0.50	0.50	0.50	0.50	0.00	0.00	0.00	0.00	1.17	0.83	1.00	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	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	0.75
Data deps.	1.00
Overall L1	1.17

Vectorization ratios

No vectorizable/vectorized instructions

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 1.17 cycles. At this rate:

14% of peak load performance is reached (13.71 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
14% of peak store performance is reached (6.86 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 437be0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P10	P11	P12	P13	P14	Latency	Recip. throughput
LDR X1, [X11, X16,LSL #3]	1	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X16, X16, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	1	0.25
SUBS X18, X18, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	1	0.33
LDR X1, [X12, X1,LSL #3]	1	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
STR X26, [X19, X1,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
B.NE 437be0 <.omp_outlined..20+0x2b0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50

Function	.omp_outlined..20#0x437930
Source file and lines	par_multi_interp.c:1612-1615
Module	exec

The loop is defined in /home/hbollore/qaas-runs/171-233-5044/intel/AMG/build/AMG/AMG/parcsr_ls/par_multi_interp.c:1612-1615.

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

gain
potential
hint
expert

Unrolling/vectorization cost

This loop is peel/tail of a unrolled/vectorized loop. If its cost is not negligible compared to the main (unrolled/vectorized) loop, unrolling/vectorization is counterproductive due to low trip count.

Details

The more iterations the main loop is processing, the higher the trip count must be to amortize peel/tail overhead.

Workaround

use profile-guided optimization flags of your compiler
hardcode most frequent values of loop bounds by adding specialized paths.:
- For instance, replace for (i=0; i<n; i++) foo(i) with: switch (n) { case (4): for (i=0; i<4; i++) foo(i); break; case (6): for (i=0; i<6; i++) foo(i); break; default : for (i=0; i<n; i++) foo(i); break; }

Vectorization

Your loop is not vectorized.

Details

No VPU instruction: only general purpose registers are used and data elements are processed one at a time.

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 16 bytes. The binary loop is storing 8 bytes.

General properties

nb instructions	6
loop length	24
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

0.75 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	0.00	0.00	0.00	0.00	1.17	0.83	1.00	0.50	0.50
cycles	0.50	0.50	0.50	0.50	0.50	0.50	0.00	0.00	0.00	0.00	1.17	0.83	1.00	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	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	0.75
Data deps.	1.00
Overall L1	1.17

Vectorization ratios

No vectorizable/vectorized instructions

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 1.17 cycles. At this rate:

14% of peak load performance is reached (13.71 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
14% of peak store performance is reached (6.86 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 437be0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P10	P11	P12	P13	P14	Latency	Recip. throughput
LDR X1, [X11, X16,LSL #3]	1	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
ADD X16, X16, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	1	0.25
SUBS X18, X18, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	1	0.33
LDR X1, [X12, X1,LSL #3]	1	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33
STR X26, [X19, X1,LSL #3]	1	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50
B.NE 437be0 <.omp_outlined..20+0x2b0>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	1	0.50

Report Configuration

Unrolling/vectorization cost

Details

Workaround

Vectorization

Details

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

Unrolling/vectorization cost

Details

Workaround

Vectorization

Details

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