OV - exec - Loop 2844

0x481f80 STR	X23, [X5, X26,LSL #3]

0x481f84 STR	X20, [X7, X26,LSL #3]

0x481f88 ORR	X20, XZR, X24

0x481f8c ORR	X23, XZR, X6

0x481f90 ORR	X26, XZR, X30

0x481f94 ORR	X19, XZR, X21

0x481f98 CMP	X30, X21

0x481f9c B.EQ	482180

0x481fa0 LDUR	X8, [X29, #496]

0x481fa4 LDR	X8, [X8]

0x481fa8 CBZ	X8, 481fb4

0x481fac ADD	X6, X23, #1

0x481fb0 STR	X23, [X0, X26,LSL #3]

0x481fb4 LDR	X8, [X15]

0x481fb8 CBZ	X8, 482080

0x481fbc LDUR	X8, [X29, #488]

0x481fc0 ADD	X13, X8, X26,LSL #3

0x481fc4 LDP	X14, X8, [X13]

0x481fc8 CMP	X14, X8

0x481fcc B.GE	482080

0x481fd0 LDR	X8, [SP, #72]

0x481fd4 ORR	X24, XZR, X20

0x481fd8 LDR	X19, [X8]

0x481fdc LDUR	X8, [X29, #448]

0x481fe0 LDR	X27, [X8]

0x481fe4 B	481ff8

  (2848) 0x481fe8 LDR	X8, [X13, #8]

  (2848) 0x481fec ADD	X14, X14, #1

  (2848) 0x481ff0 CMP	X14, X8

  (2848) 0x481ff4 B.GE	482084

  (2848) 0x481ff8 LDR	X9, [X17, X14,LSL #3]

  (2848) 0x481ffc ADD	X8, X9, #1

  (2848) 0x482000 LDR	X28, [X18, X9,LSL #3]

  (2848) 0x482004 LDR	X10, [X18, X8,LSL #3]

  (2848) 0x482008 B	482010

    (2850) 0x48200c ADD	X28, X28, #1

    (2850) 0x482010 CMP	X28, X10

    (2850) 0x482014 B.GE	482040

    (2850) 0x482018 LDR	X12, [X22]

    (2850) 0x48201c LDR	X25, [X19, X28,LSL #3]

    (2850) 0x482020 ADD	X12, X25, X12

    (2850) 0x482024 LDR	X25, [X0, X12,LSL #3]

    (2850) 0x482028 CMP	X25, X20

    (2850) 0x48202c B.GE	48200c

    (2850) 0x482030 STR	X24, [X0, X12,LSL #3]

    (2850) 0x482034 ADD	X24, X24, #1

    (2850) 0x482038 LDR	X10, [X18, X8,LSL #3]

    (2850) 0x48203c B	48200c

  (2848) 0x482040 LDR	X9, [X1, X9,LSL #3]

  (2848) 0x482044 LDR	X10, [X1, X8,LSL #3]

  (2848) 0x482048 B	482050

    (2849) 0x48204c ADD	X9, X9, #1

    (2849) 0x482050 CMP	X9, X10

    (2849) 0x482054 B.GE	481fe8

    (2849) 0x482058 LDR	X12, [X27, X9,LSL #3]

    (2849) 0x48205c LDR	X25, [X0, X12,LSL #3]

    (2849) 0x482060 CMP	X25, X23

    (2849) 0x482064 B.GE	48204c

    (2849) 0x482068 STR	X6, [X0, X12,LSL #3]

    (2849) 0x48206c ADD	X6, X6, #1

    (2849) 0x482070 LDR	X10, [X1, X8,LSL #3]

    (2849) 0x482074 B	48204c

0x482080 ORR	X24, XZR, X20

0x482084 ADD	X30, X26, #1

0x482088 LDR	X28, [X2, X26,LSL #3]

0x48208c LDR	X8, [X2, X30,LSL #3]

0x482090 CMP	X28, X8

0x482094 B.GE	481f80

/home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_mv/par_csr_matop.c: 127 - 242

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

127:     for (i1 = ns; i1 < ne; i1++)

[...]

135:       if ( allsquare ) {

136:          B_marker[i1] = jj_count_diag;

137:          jj_count_diag++;

[...]

144:       if (num_cols_offd_A)

145:       {

146:          for (jj2 = A_offd_i[i1]; jj2 < A_offd_i[i1+1]; jj2++)

147:          {

148:             i2 = A_offd_j[jj2];

[...]

154:             for (jj3 = B_ext_offd_i[i2]; jj3 < B_ext_offd_i[i2+1]; jj3++)

155:             {

156:                i3 = num_cols_diag_B+B_ext_offd_j[jj3];

[...]

164:                if (B_marker[i3] < jj_row_begin_offd)

165:                {

166:                   B_marker[i3] = jj_count_offd;

167:                   jj_count_offd++;

168:                }

169:             }

170:             for (jj3 = B_ext_diag_i[i2]; jj3 < B_ext_diag_i[i2+1]; jj3++)

171:             {

172:                i3 = B_ext_diag_j[jj3];

173:

174:                if (B_marker[i3] < jj_row_begin_diag)

175:                {

176:                   B_marker[i3] = jj_count_diag;

177:                   jj_count_diag++;

[...]

187:       for (jj2 = A_diag_i[i1]; jj2 < A_diag_i[i1+1]; jj2++)

[...]

241:       (*C_diag_i)[i1] = jj_row_begin_diag;

242:       (*C_offd_i)[i1] = jj_row_begin_offd;

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

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	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	2.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P12,
Function	hypre_ParMatmul_RowSizes.omp_outlined
Source	par_csr_matop.c:127-127,par_csr_matop.c:135-137,par_csr_matop.c:144-146,par_csr_matop.c:170-170,par_csr_matop.c:187-187,par_csr_matop.c:241-242
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	4.83
CQA cycles if no scalar integer	4.83
CQA cycles if FP arith vectorized	4.83
CQA cycles if fully vectorized	2.42
Front-end cycles	4.00
P0 cycles	3.00
P1 cycles	3.00
P2 cycles	2.08
P3 cycles	2.08
P4 cycles	2.00
P5 cycles	2.00
P6 cycles	1.92
P7 cycles	1.92
P8 cycles	0.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	4.83
P13 cycles	4.50
P14 cycles	4.67
P15 cycles	1.50
P16 cycles	1.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	32.00
Nb uops	32.00
Nb loads	NA
Nb stores	3.00
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	24.83
Bytes prefetched	0.00
Bytes loaded	96.00
Bytes stored	24.00
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	50.00
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	50.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	50.00

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	2.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.04
Bottlenecks	P12,
Function	hypre_ParMatmul_RowSizes.omp_outlined
Source	par_csr_matop.c:127-127,par_csr_matop.c:135-137,par_csr_matop.c:144-146,par_csr_matop.c:170-170,par_csr_matop.c:187-187,par_csr_matop.c:241-242
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	4.83
CQA cycles if no scalar integer	4.83
CQA cycles if FP arith vectorized	4.83
CQA cycles if fully vectorized	2.42
Front-end cycles	4.00
P0 cycles	3.00
P1 cycles	3.00
P2 cycles	2.08
P3 cycles	2.08
P4 cycles	2.00
P5 cycles	2.00
P6 cycles	1.92
P7 cycles	1.92
P8 cycles	0.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	4.83
P13 cycles	4.50
P14 cycles	4.67
P15 cycles	1.50
P16 cycles	1.50
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	32.00
Nb uops	32.00
Nb loads	NA
Nb stores	3.00
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	24.83
Bytes prefetched	0.00
Bytes loaded	96.00
Bytes stored	24.00
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	50.00
Vector-efficiency ratio load	50.00
Vector-efficiency ratio store	50.00
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	50.00
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	50.00

Path /

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

Function	hypre_ParMatmul_RowSizes.omp_outlined
Source file and lines	par_csr_matop.c:127-242
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_mv/par_csr_matop.c:127,135-137,144-148,154-156,164-177,187,241-242.

Analyzed code is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_mv/par_csr_matop.c:127,135-137,144-146,170,187,241-242.

Warnings:

Non-innermost loop: analyzing only self part (ignoring child loops).
Ignoring paths for analysis
Failed to get the number of paths
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. 2 data elements could be processed at once in vector registers. By vectorizing your loop, you can lower the cost of an iteration from 4.83 to 2.42 cycles (2.00x 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 96 bytes. The binary loop is storing 24 bytes.

General properties

nb instructions	32
loop length	128
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

4.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	3.00	3.00	2.08	2.08	2.00	2.00	1.92	1.92	0.00	0.00	0.00	0.00	4.83	4.50	4.67	1.50	1.50
cycles	3.00	3.00	2.08	2.08	2.00	2.00	1.92	1.92	0.00	0.00	0.00	0.00	4.83	4.50	4.67	1.50	1.50

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

NA

Cycles summary

Front-end	4.00
Overall L1	4.83

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 4.83 cycles. At this rate:

20% of peak load performance is reached (19.86 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
10% of peak store performance is reached (4.97 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 481f80

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P12	P13	P14	P15	P16	Latency	Recip. throughput	Vectorization
STR X23, [X5, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
STR X20, [X7, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ORR X20, XZR, X24	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
ORR X23, XZR, X6	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
ORR X26, XZR, X30	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	N/A
ORR X19, XZR, X21	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
CMP X30, X21	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	1	0.33	scal (50.0%)
B.EQ 482180 <hypre_ParMatmul_RowSizes.omp_outlined+0x380>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDUR X8, [X29, #496]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X8, [X8]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
CBZ X8, 481fb4 <hypre_ParMatmul_RowSizes.omp_outlined+0x1b4>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD X6, X23, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
STR X23, [X0, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDR X8, [X15]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
CBZ X8, 482080 <hypre_ParMatmul_RowSizes.omp_outlined+0x280>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDUR X8, [X29, #488]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD X13, X8, X26,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	N/A
LDP X14, X8, [X13]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
CMP X14, X8	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	1	0.33	scal (50.0%)
B.GE 482080 <hypre_ParMatmul_RowSizes.omp_outlined+0x280>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X8, [SP, #72]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ORR X24, XZR, X20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
LDR X19, [X8]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDUR X8, [X29, #448]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X27, [X8]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
B 481ff8 <hypre_ParMatmul_RowSizes.omp_outlined+0x1f8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR X24, XZR, X20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X30, X26, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	N/A
LDR X28, [X2, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X8, [X2, X30,LSL #3]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
CMP X28, X8	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	1	0.33	N/A
B.GE 481f80 <hypre_ParMatmul_RowSizes.omp_outlined+0x180>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

Function	hypre_ParMatmul_RowSizes.omp_outlined
Source file and lines	par_csr_matop.c:127-242
Module	exec

The loop is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_mv/par_csr_matop.c:127,135-137,144-148,154-156,164-177,187,241-242.

Analyzed code is defined in /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/build/AMG/AMG/parcsr_mv/par_csr_matop.c:127,135-137,144-146,170,187,241-242.

Warnings:

Non-innermost loop: analyzing only self part (ignoring child loops).
Ignoring paths for analysis
Failed to get the number of paths
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. 2 data elements could be processed at once in vector registers. By vectorizing your loop, you can lower the cost of an iteration from 4.83 to 2.42 cycles (2.00x 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 96 bytes. The binary loop is storing 24 bytes.

General properties

nb instructions	32
loop length	128
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

4.00 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	P16
uops	3.00	3.00	2.08	2.08	2.00	2.00	1.92	1.92	0.00	0.00	0.00	0.00	4.83	4.50	4.67	1.50	1.50
cycles	3.00	3.00	2.08	2.08	2.00	2.00	1.92	1.92	0.00	0.00	0.00	0.00	4.83	4.50	4.67	1.50	1.50

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

NA

Cycles summary

Front-end	4.00
Overall L1	4.83

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 4.83 cycles. At this rate:

20% of peak load performance is reached (19.86 out of 96.00 bytes loaded per cycle (GB/s @ 1GHz))
10% of peak store performance is reached (4.97 out of 48.00 bytes stored per cycle (GB/s @ 1GHz))

ASM code

In the binary file, the address of the loop is: 481f80

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P12	P13	P14	P15	P16	Latency	Recip. throughput	Vectorization
STR X23, [X5, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
STR X20, [X7, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
ORR X20, XZR, X24	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
ORR X23, XZR, X6	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
ORR X26, XZR, X30	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	N/A
ORR X19, XZR, X21	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
CMP X30, X21	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	1	0.33	scal (50.0%)
B.EQ 482180 <hypre_ParMatmul_RowSizes.omp_outlined+0x380>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDUR X8, [X29, #496]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X8, [X8]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
CBZ X8, 481fb4 <hypre_ParMatmul_RowSizes.omp_outlined+0x1b4>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ADD X6, X23, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
STR X23, [X0, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.50	0.50	0	0.50	0.50	1	0.50	scal (50.0%)
LDR X8, [X15]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
CBZ X8, 482080 <hypre_ParMatmul_RowSizes.omp_outlined+0x280>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDUR X8, [X29, #488]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ADD X13, X8, X26,LSL #3	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	N/A
LDP X14, X8, [X13]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.50	N/A
CMP X14, X8	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	1	0.33	scal (50.0%)
B.GE 482080 <hypre_ParMatmul_RowSizes.omp_outlined+0x280>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
LDR X8, [SP, #72]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
ORR X24, XZR, X20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
LDR X19, [X8]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
LDUR X8, [X29, #448]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X27, [X8]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	scal (50.0%)
B 481ff8 <hypre_ParMatmul_RowSizes.omp_outlined+0x1f8>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A
ORR X24, XZR, X20	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	scal (50.0%)
ADD X30, X26, #1	1	0	0	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	1	0.17	N/A
LDR X28, [X2, X26,LSL #3]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
LDR X8, [X2, X30,LSL #3]	1	0	0	0	0	0	0	0	0	0.33	0.33	0.33	0	0	4	0.33	N/A
CMP X28, X8	1	0	0	0.25	0.25	0	0	0.25	0.25	0	0	0	0	0	1	0.33	N/A
B.GE 481f80 <hypre_ParMatmul_RowSizes.omp_outlined+0x180>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	1	0.50	N/A

▶Scalability Runs Description

Run 1x1	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_NUM_THREADS: 1OMP_PLACES: threads
Run 1x2	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 2OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x4	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 4OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x8	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 8OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x16	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 16OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x24	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 24OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x32	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 32OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x40	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 40OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x48	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 48OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x56	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 56OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x64	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 64OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x72	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 72OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x80	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 80OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x88	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 88OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads
Run 1x96	Number processes: 1Number nodes: 1Run Command: <executable> -n 400 400 400MPI Command: mpirun -n <number_processes> --bind-to core --map-by package:PE=96 --rank-by fill --report-bindings Dataset: Run Directory: /home/hbollore/qaas/qaas-runs/174-161-6712/intel/AMG/run/oneview_runs/multicore/armclang_5/oneview_run_1741633338OMP_NUM_THREADS: 96OMP_PROC_BIND: spreadOMP_DISPLAY_AFFINITY: TRUEOMP_AFFINITY_FORMAT: 'OMP: pid %P tid %i thread %n bound to OS proc set {%A}'OMP_DISPLAY_ENV: TRUEOMP_PLACES: threads

(1x1) Efficiency	(1x1) Potential Speed-Up (%)	(1x2) Efficiency	(1x2) Potential Speed-Up (%)	(1x4) Efficiency	(1x4) Potential Speed-Up (%)	(1x8) Efficiency	(1x8) Potential Speed-Up (%)	(1x16) Efficiency	(1x16) Potential Speed-Up (%)	(1x24) Efficiency	(1x24) Potential Speed-Up (%)	(1x32) Efficiency	(1x32) Potential Speed-Up (%)	(1x40) Efficiency	(1x40) Potential Speed-Up (%)	(1x48) Efficiency	(1x48) Potential Speed-Up (%)	(1x56) Efficiency	(1x56) Potential Speed-Up (%)	(1x64) Efficiency	(1x64) Potential Speed-Up (%)	(1x72) Efficiency	(1x72) Potential Speed-Up (%)	(1x80) Efficiency	(1x80) Potential Speed-Up (%)	(1x88) Efficiency	(1x88) Potential Speed-Up (%)	(1x96) Efficiency	(1x96) Potential Speed-Up (%)
1	0	2.97	0	4.98	0	9.36	0	17.86	0	30.82	0	36.54	0	36.15	0	65.04	0	48.84	0	57.52	0	83.89	0	69.7	0	81.92	0	88.74	0

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency

Run	Number of threads	Efficiency (ideal is 1)	Speedup	Ideal Speedup	Time (s)	Coverage (%)
1x1	1	1	1	1	0.38999995589256	0.14660058915615
1x2	2	2.97	2.97	2	0.14499999582767	0.092690140008926
1x4	4	4.98	4.98	4	0.094999983906746	0.10360082238913
1x8	8	9.36	9.36	8	0.054999995976686	0.09442250430584
1x16	15	17.86	17.86	16	0.03999999538064	0.075790107250214
1x24	22	30.82	30.82	24	0.029999993741512	0.051099788397551
1x32	26	36.54	36.54	32	0.024999998509884	0.04701329022646
1x40	34	36.15	36.15	40	0.024999998509884	0.044712979346514
1x48	29	65.04	65.04	48	0.01999999769032	0.027054730802774
1x56	41	48.84	48.84	56	0.025000000372529	0.039639197289944
1x64	45	57.52	57.52	64	0.024999998509884	0.03481825068593
1x72	45	83.89	83.89	72	0.010000000707805	0.023902289569378
1x80	55	69.7	69.7	80	0.015000000596046	0.031259369105101
1x88	48	81.92	81.92	88	0.015000000596046	0.024271856993437
1x96	51	88.74	88.74	96	0.01999999769032	0.022498112171888

Report Configuration

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

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

▶Scalability Runs Description

▼Columns Filter

Speed-Up and Efficiency

▶Detailed Speed-Up and Efficiency