| Loop Id: 56 | Module: attention-native | Source: attention.cpp:26-33 | Coverage: 0.16% |
|---|
| Loop Id: 56 | Module: attention-native | Source: attention.cpp:26-33 | Coverage: 0.16% |
|---|
0x12560 ORR X14, XZR, XZR |
0x12564 MOVI D1, #0 |
0x12568 ADD W15, W13, W14 |
0x1256c MADD W14, W14, W25, W12 |
0x12570 LDR S2, [X20, W15,UXTW #2] |
0x12574 LDR S3, [X22, W14,UXTW #2] |
0x12578 FCVT D2, S2 |
0x1257c FCVT D3, S3 |
0x12580 FMADD D1, D2, D3, D1 |
0x12584 FCVT S1, D1 |
0x12588 ADD W14, W13, W12 |
0x1258c ADD X12, X12, #1 |
0x12590 CMP X12, X0 |
0x12594 STR S1, [X21, W14,UXTW #2] |
0x12598 B.EQ 12544 |
0x1259c CMP X0, #1 |
0x125a0 B.EQ 12560 |
0x125a4 MOVI D1, #0 |
0x125a8 ORR X14, XZR, XZR |
0x125ac ORR X15, XZR, X12 |
0x125b0 ORR W16, WZR, W11 |
0x125b4 HINT #0 |
0x125b8 HINT #0 |
0x125bc HINT #0 |
(57) 0x125c0 SUB W17, W16, #1 |
(57) 0x125c4 LDR S3, [X22, W15,UXTW #2] |
(57) 0x125c8 ADD X14, X14, #2 |
(57) 0x125cc LDR S2, [X20, W17,UXTW #2] |
(57) 0x125d0 ADD W17, W25, W15 |
(57) 0x125d4 ADD X15, X15, X19 |
(57) 0x125d8 CMP X9, X14 |
(57) 0x125dc FCVT D3, S3 |
(57) 0x125e0 FCVT D2, S2 |
(57) 0x125e4 FMADD D1, D2, D3, D1 |
(57) 0x125e8 LDR S2, [X20, W16,UXTW #2] |
(57) 0x125ec LDR S3, [X22, W17,UXTW #2] |
(57) 0x125f0 ADD W16, W16, #2 |
(57) 0x125f4 FCVT D2, S2 |
(57) 0x125f8 FCVT D3, S3 |
(57) 0x125fc FMADD D1, D2, D3, D1 |
(57) 0x12600 B.NE 125c0 |
0x12604 TBNZ W25, #0, 12568 |
0x12608 B 12584 |
/home/eoseret/llm-attention/attention.cpp: 26 - 33 |
-------------------------------------------------------------------------------- |
26: for (unsigned int i = 0; i < M; ++i) { |
27: for (unsigned int j = 0; j < N; ++j) { //vectorized |
28: double sum = 0.0; |
29: #pragma clang loop vectorize(enable) |
30: for (unsigned int k = 0; k < K; ++k) { //vectorized |
31: sum += (double)A[i * K + k] * (double)B[k * N + j]; |
32: } |
33: C[i * N + j] = alpha * static_cast<float>(sum); |
| Coverage (%) | Name | Source Location | Module |
|---|---|---|---|
| ►100.00+ | __libc_start_call_main | libc.so.6 | |
| ○ | __libc_start_main | libc.so.6 | |
| ○ | _start | attention-native |
| 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 | 1.99 |
| CQA speedup if FP arith vectorized | 1.32 |
| CQA speedup if fully vectorized | 6.58 |
| CQA speedup if no inter-iteration dependency | NA |
| CQA speedup if next bottleneck killed | 1.15 |
| Bottlenecks | |
| Function | main |
| Source | attention.cpp:26-27,attention.cpp:30-33 |
| Source loop unroll info | NA |
| Source loop unroll confidence level | NA |
| Unroll/vectorization loop type | NA |
| Unroll factor | NA |
| CQA cycles | 2.21 |
| CQA cycles if no scalar integer | 1.11 |
| CQA cycles if FP arith vectorized | 1.67 |
| CQA cycles if fully vectorized | 0.34 |
| Front-end cycles | 2.13 |
| P0 cycles | 1.50 |
| P1 cycles | 1.50 |
| P2 cycles | 1.92 |
| P3 cycles | 1.92 |
| P4 cycles | 1.92 |
| P5 cycles | 1.92 |
| P6 cycles | 1.33 |
| P7 cycles | 1.25 |
| P8 cycles | 1.25 |
| P9 cycles | 1.17 |
| P10 cycles | 0.94 |
| P11 cycles | 0.72 |
| P12 cycles | 0.67 |
| P13 cycles | 0.00 |
| P14 cycles | 0.00 |
| DIV/SQRT cycles | 0.00 |
| Inter-iter dependencies cycles | 1 |
| FE+BE cycles (UFS) | NA |
| Stall cycles (UFS) | NA |
| Nb insns | 19.00 |
| Nb uops | 17.00 |
| Nb loads | NA |
| Nb stores | 1.00 |
| Nb stack references | 0.00 |
| FLOP/cycle | 0.60 |
| Nb FLOP add-sub | 0.00 |
| Nb FLOP mul | 0.00 |
| Nb FLOP fma | 0.67 |
| Nb FLOP div | 0.00 |
| Nb FLOP rcp | 0.00 |
| Nb FLOP sqrt | 0.00 |
| Nb FLOP rsqrt | 0.00 |
| Bytes/cycle | 3.11 |
| Bytes prefetched | 0.00 |
| Bytes loaded | 4.00 |
| Bytes stored | 3.00 |
| 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 | 0.00 |
| Vectorization ratio div_sqrt | NA |
| Vectorization ratio other | 0.00 |
| Vector-efficiency ratio all | 19.45 |
| Vector-efficiency ratio load | 12.50 |
| Vector-efficiency ratio store | 12.50 |
| Vector-efficiency ratio mul | NA |
| Vector-efficiency ratio add_sub | 25.00 |
| Vector-efficiency ratio fma | 25.00 |
| Vector-efficiency ratio div_sqrt | NA |
| Vector-efficiency ratio other | 20.42 |
| Metric | Value |
|---|---|
| CQA speedup if no scalar integer | 1.82 |
| CQA speedup if FP arith vectorized | 1.42 |
| CQA speedup if fully vectorized | 6.48 |
| CQA speedup if no inter-iteration dependency | NA |
| CQA speedup if next bottleneck killed | 1.21 |
| Bottlenecks | micro-operation queue, |
| Function | main |
| Source | attention.cpp:26-27,attention.cpp:30-33 |
| Source loop unroll info | NA |
| Source loop unroll confidence level | NA |
| Unroll/vectorization loop type | NA |
| Unroll factor | NA |
| CQA cycles | 2.13 |
| CQA cycles if no scalar integer | 1.17 |
| CQA cycles if FP arith vectorized | 1.50 |
| CQA cycles if fully vectorized | 0.33 |
| Front-end cycles | 2.13 |
| P0 cycles | 1.00 |
| P1 cycles | 1.00 |
| P2 cycles | 1.75 |
| P3 cycles | 1.75 |
| P4 cycles | 1.75 |
| P5 cycles | 1.75 |
| P6 cycles | 1.50 |
| P7 cycles | 1.50 |
| P8 cycles | 1.50 |
| P9 cycles | 1.50 |
| P10 cycles | 1.17 |
| P11 cycles | 0.83 |
| P12 cycles | 1.00 |
| P13 cycles | 0.00 |
| P14 cycles | 0.00 |
| DIV/SQRT cycles | 0.00 |
| Inter-iter dependencies cycles | 1 |
| FE+BE cycles (UFS) | NA |
| Stall cycles (UFS) | NA |
| Nb insns | 17.00 |
| Nb uops | 17.00 |
| Nb loads | NA |
| Nb stores | 1.00 |
| Nb stack references | 0.00 |
| FLOP/cycle | 0.94 |
| Nb FLOP add-sub | 0.00 |
| Nb FLOP mul | 0.00 |
| Nb FLOP fma | 1.00 |
| Nb FLOP div | 0.00 |
| Nb FLOP rcp | 0.00 |
| Nb FLOP sqrt | 0.00 |
| Nb FLOP rsqrt | 0.00 |
| Bytes/cycle | 4.24 |
| Bytes prefetched | 0.00 |
| Bytes loaded | 6.00 |
| Bytes stored | 3.00 |
| 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 | 0.00 |
| Vectorization ratio div_sqrt | NA |
| Vectorization ratio other | 0.00 |
| Vector-efficiency ratio all | 18.75 |
| Vector-efficiency ratio load | 12.50 |
| Vector-efficiency ratio store | 12.50 |
| Vector-efficiency ratio mul | NA |
| Vector-efficiency ratio add_sub | 25.00 |
| Vector-efficiency ratio fma | 25.00 |
| Vector-efficiency ratio div_sqrt | NA |
| Vector-efficiency ratio other | 20.00 |
| Metric | Value |
|---|---|
| CQA speedup if no scalar integer | 2.00 |
| CQA speedup if FP arith vectorized | 1.00 |
| CQA speedup if fully vectorized | 6.40 |
| CQA speedup if no inter-iteration dependency | NA |
| CQA speedup if next bottleneck killed | 1.14 |
| Bottlenecks | P0, P1, |
| Function | main |
| Source | attention.cpp:26-27,attention.cpp:30-33 |
| Source loop unroll info | NA |
| Source loop unroll confidence level | NA |
| Unroll/vectorization loop type | NA |
| Unroll factor | NA |
| CQA cycles | 2.00 |
| CQA cycles if no scalar integer | 1.00 |
| CQA cycles if FP arith vectorized | 2.00 |
| CQA cycles if fully vectorized | 0.31 |
| Front-end cycles | 1.75 |
| P0 cycles | 2.00 |
| P1 cycles | 2.00 |
| P2 cycles | 1.75 |
| P3 cycles | 1.75 |
| P4 cycles | 1.75 |
| P5 cycles | 1.75 |
| P6 cycles | 1.00 |
| P7 cycles | 0.75 |
| P8 cycles | 0.75 |
| P9 cycles | 0.50 |
| P10 cycles | 0.50 |
| P11 cycles | 0.50 |
| P12 cycles | 0.00 |
| P13 cycles | 0.00 |
| P14 cycles | 0.00 |
| DIV/SQRT cycles | 0.00 |
| Inter-iter dependencies cycles | 1 |
| FE+BE cycles (UFS) | NA |
| Stall cycles (UFS) | NA |
| Nb insns | 17.00 |
| Nb uops | 14.00 |
| 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 | 1.50 |
| Bytes prefetched | 0.00 |
| Bytes loaded | 0.00 |
| Bytes stored | 3.00 |
| Stride 0 | 0.00 |
| Stride 1 | 0.00 |
| Stride n | 0.00 |
| Stride unknown | 1.00 |
| Stride indirect | 1.00 |
| Vectorization ratio all | 0.00 |
| Vectorization ratio load | NA |
| 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 | 21.43 |
| Vector-efficiency ratio load | NA |
| Vector-efficiency ratio store | 12.50 |
| 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 | 22.50 |
| Metric | Value |
|---|---|
| CQA speedup if no scalar integer | 2.14 |
| CQA speedup if FP arith vectorized | 1.67 |
| CQA speedup if fully vectorized | 6.81 |
| CQA speedup if no inter-iteration dependency | NA |
| CQA speedup if next bottleneck killed | 1.11 |
| Bottlenecks | micro-operation queue, |
| Function | main |
| Source | attention.cpp:26-27,attention.cpp:30-33 |
| Source loop unroll info | NA |
| Source loop unroll confidence level | NA |
| Unroll/vectorization loop type | NA |
| Unroll factor | NA |
| CQA cycles | 2.50 |
| CQA cycles if no scalar integer | 1.17 |
| CQA cycles if FP arith vectorized | 1.50 |
| CQA cycles if fully vectorized | 0.37 |
| Front-end cycles | 2.50 |
| P0 cycles | 1.50 |
| P1 cycles | 1.50 |
| P2 cycles | 2.25 |
| P3 cycles | 2.25 |
| P4 cycles | 2.25 |
| P5 cycles | 2.25 |
| P6 cycles | 1.50 |
| P7 cycles | 1.50 |
| P8 cycles | 1.50 |
| P9 cycles | 1.50 |
| P10 cycles | 1.17 |
| P11 cycles | 0.83 |
| P12 cycles | 1.00 |
| P13 cycles | 0.00 |
| P14 cycles | 0.00 |
| DIV/SQRT cycles | 0.00 |
| Inter-iter dependencies cycles | 1 |
| FE+BE cycles (UFS) | NA |
| Stall cycles (UFS) | NA |
| Nb insns | 23.00 |
| Nb uops | 20.00 |
| Nb loads | NA |
| Nb stores | 1.00 |
| Nb stack references | 0.00 |
| FLOP/cycle | 0.80 |
| Nb FLOP add-sub | 0.00 |
| Nb FLOP mul | 0.00 |
| Nb FLOP fma | 1.00 |
| Nb FLOP div | 0.00 |
| Nb FLOP rcp | 0.00 |
| Nb FLOP sqrt | 0.00 |
| Nb FLOP rsqrt | 0.00 |
| Bytes/cycle | 3.60 |
| Bytes prefetched | 0.00 |
| Bytes loaded | 6.00 |
| Bytes stored | 3.00 |
| Stride 0 | 0.00 |
| Stride 1 | 0.00 |
| Stride n | 0.00 |
| Stride unknown | 1.00 |
| Stride indirect | 1.00 |
| 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 | 0.00 |
| Vectorization ratio div_sqrt | NA |
| Vectorization ratio other | 0.00 |
| Vector-efficiency ratio all | 18.18 |
| Vector-efficiency ratio load | 12.50 |
| Vector-efficiency ratio store | 12.50 |
| Vector-efficiency ratio mul | NA |
| Vector-efficiency ratio add_sub | 25.00 |
| Vector-efficiency ratio fma | 25.00 |
| Vector-efficiency ratio div_sqrt | NA |
| Vector-efficiency ratio other | 18.75 |
| Path / |
| Function | main |
| Source file and lines | attention.cpp:26-33 |
| Module | attention-native |
 |
| nb instructions | 19 |
| nb uops | 17 |
| loop length | 76 |
| used w registers | 6.67 |
| used x registers | 7 |
| used b registers | 0 |
| used h registers | 0 |
| used s registers | 2.33 |
| used d registers | 2.33 |
| used q registers | 0 |
| used v registers | 0 |
| used z registers | 0 |
| nb stack references | 0 |
| micro-operation queue | 2.13 cycles |
| front end | 2.13 cycles |
| P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| uops | 1.50 | 1.50 | 1.92 | 1.92 | 1.92 | 1.92 | 1.33 | 1.25 | 1.25 | 1.17 | 0.94 | 0.72 | 0.67 | 0.00 | 0.00 |
| cycles | 1.50 | 1.50 | 1.92 | 1.92 | 1.92 | 1.92 | 1.33 | 1.25 | 1.25 | 1.17 | 0.94 | 0.72 | 0.67 | 0.00 | 0.00 |
| Cycles executing div or sqrt instructions | NA |
| Longest recurrence chain latency (RecMII) | 1.00 |
| Front-end | 2.13 |
| Dispatch | 2.00 |
| Data deps. | 1.00 |
| Overall L1 | 2.21 |
| all | 0% |
| load | 0% |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | 0% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | 0% |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | 0% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 19% |
| load | 12% |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 22% |
| all | 20% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | 25% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 19% |
| all | 19% |
| load | 12% |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | 25% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 20% |
| Function | main |
| Source file and lines | attention.cpp:26-33 |
| Module | attention-native |
|
| nb instructions | 17 |
| nb uops | 17 |
| loop length | 68 |
| used w registers | 5 |
| used x registers | 7 |
| used b registers | 0 |
| used h registers | 0 |
| used s registers | 3 |
| used d registers | 3 |
| used q registers | 0 |
| used v registers | 0 |
| used z registers | 0 |
| nb stack references | 0 |
| micro-operation queue | 2.13 cycles |
| front end | 2.13 cycles |
| P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| uops | 1.00 | 1.00 | 1.75 | 1.75 | 1.75 | 1.75 | 1.50 | 1.50 | 1.50 | 1.50 | 1.17 | 0.83 | 1.00 | 0.00 | 0.00 |
| cycles | 1.00 | 1.00 | 1.75 | 1.75 | 1.75 | 1.75 | 1.50 | 1.50 | 1.50 | 1.50 | 1.17 | 0.83 | 1.00 | 0.00 | 0.00 |
| Cycles executing div or sqrt instructions | NA |
| Longest recurrence chain latency (RecMII) | 1.00 |
| Front-end | 2.13 |
| Dispatch | 1.75 |
| Data deps. | 1.00 |
| Overall L1 | 2.13 |
| all | 0% |
| load | 0% |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | 0% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | 0% |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | 0% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 18% |
| load | 12% |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 25% |
| all | 18% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | 25% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 16% |
| all | 18% |
| load | 12% |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | 25% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 20% |
| Instruction | Nb FU | P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | Latency | Recip. throughput | Vectorization |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ORR X14, XZR, XZR | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| MOVI D1, #0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 2 | 0.25 | scal (25.0%) |
| ADD W15, W13, W14 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| MADD W14, W14, W25, W12 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 1 | N/A |
| LDR S2, [X20, W15,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.33 | 0.33 | 0.33 | 0 | 0 | 6 | 0.33 | scal (12.5%) |
| LDR S3, [X22, W14,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.33 | 0.33 | 0.33 | 0 | 0 | 6 | 0.33 | scal (12.5%) |
| FCVT D2, S2 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (12.5%) |
| FCVT D3, S3 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (12.5%) |
| FMADD D1, D2, D3, D1 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 4 | 0.25 | scal (25.0%) |
| FCVT S1, D1 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (25.0%) |
| ADD W14, W13, W12 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| ADD X12, X12, #1 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | scal (25.0%) |
| CMP X12, X0 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 | scal (25.0%) |
| STR S1, [X21, W14,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0.50 | 0 | 0 | 0.50 | 0.50 | 0 | 0 | 0 | 2 | 0.50 | scal (12.5%) |
| B.EQ 12544 <main+0x1544> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| CMP X0, #1 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 | N/A |
| B.EQ 12560 <main+0x1560> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| Function | main |
| Source file and lines | attention.cpp:26-33 |
| Module | attention-native |
|
| nb instructions | 17 |
| nb uops | 14 |
| loop length | 68 |
| used w registers | 7 |
| used x registers | 6 |
| used b registers | 0 |
| used h registers | 0 |
| used s registers | 1 |
| used d registers | 1 |
| used q registers | 0 |
| used v registers | 0 |
| used z registers | 0 |
| nb stack references | 0 |
| micro-operation queue | 1.75 cycles |
| front end | 1.75 cycles |
| P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| uops | 2.00 | 2.00 | 1.75 | 1.75 | 1.75 | 1.75 | 1.00 | 0.75 | 0.75 | 0.50 | 0.50 | 0.50 | 0.00 | 0.00 | 0.00 |
| cycles | 2.00 | 2.00 | 1.75 | 1.75 | 1.75 | 1.75 | 1.00 | 0.75 | 0.75 | 0.50 | 0.50 | 0.50 | 0.00 | 0.00 | 0.00 |
| Cycles executing div or sqrt instructions | NA |
| Longest recurrence chain latency (RecMII) | 1.00 |
| Front-end | 1.75 |
| Dispatch | 2.00 |
| Data deps. | 1.00 |
| Overall L1 | 2.00 |
| all | 0% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | NA (no fma vectorizable/vectorized instructions) |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | NA (no load vectorizable/vectorized instructions) |
| 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% |
| all | 20% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 21% |
| all | 25% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | NA (no fma vectorizable/vectorized instructions) |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 25% |
| all | 21% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 22% |
| Instruction | Nb FU | P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | Latency | Recip. throughput | Vectorization |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| FCVT S1, D1 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (25.0%) |
| ADD W14, W13, W12 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| ADD X12, X12, #1 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | scal (25.0%) |
| CMP X12, X0 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 | scal (25.0%) |
| STR S1, [X21, W14,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0.50 | 0 | 0 | 0.50 | 0.50 | 0 | 0 | 0 | 2 | 0.50 | scal (12.5%) |
| B.EQ 12544 <main+0x1544> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| CMP X0, #1 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 | N/A |
| B.EQ 12560 <main+0x1560> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| MOVI D1, #0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 2 | 0.25 | scal (25.0%) |
| ORR X14, XZR, XZR | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| ORR X15, XZR, X12 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | scal (25.0%) |
| ORR W16, WZR, W11 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | scal (12.5%) |
| HINT #0 | N/A | ||||||||||||||||||
| HINT #0 | N/A | ||||||||||||||||||
| HINT #0 | N/A | ||||||||||||||||||
| TBNZ W25, #0, 12568 <main+0x1568> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| B 12584 <main+0x1584> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| Function | main |
| Source file and lines | attention.cpp:26-33 |
| Module | attention-native |
|
| nb instructions | 23 |
| nb uops | 20 |
| loop length | 92 |
| used w registers | 8 |
| used x registers | 8 |
| used b registers | 0 |
| used h registers | 0 |
| used s registers | 3 |
| used d registers | 3 |
| used q registers | 0 |
| used v registers | 0 |
| used z registers | 0 |
| nb stack references | 0 |
| micro-operation queue | 2.50 cycles |
| front end | 2.50 cycles |
| P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| uops | 1.50 | 1.50 | 2.25 | 2.25 | 2.25 | 2.25 | 1.50 | 1.50 | 1.50 | 1.50 | 1.17 | 0.83 | 1.00 | 0.00 | 0.00 |
| cycles | 1.50 | 1.50 | 2.25 | 2.25 | 2.25 | 2.25 | 1.50 | 1.50 | 1.50 | 1.50 | 1.17 | 0.83 | 1.00 | 0.00 | 0.00 |
| Cycles executing div or sqrt instructions | NA |
| Longest recurrence chain latency (RecMII) | 1.00 |
| Front-end | 2.50 |
| Dispatch | 2.25 |
| Data deps. | 1.00 |
| Overall L1 | 2.50 |
| all | 0% |
| load | 0% |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | 0% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 0% |
| load | 0% |
| store | 0% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 0% |
| fma | 0% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 0% |
| all | 17% |
| load | 12% |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | NA (no fma vectorizable/vectorized instructions) |
| other | 20% |
| all | 18% |
| load | NA (no load vectorizable/vectorized instructions) |
| store | NA (no store vectorizable/vectorized instructions) |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | NA (no add-sub vectorizable/vectorized instructions) |
| fma | 25% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 16% |
| all | 18% |
| load | 12% |
| store | 12% |
| mul | NA (no mul vectorizable/vectorized instructions) |
| add-sub | 25% |
| fma | 25% |
| div/sqrt | NA (no div/sqrt vectorizable/vectorized instructions) |
| other | 18% |
| Instruction | Nb FU | P0 | P1 | P2 | P3 | P4 | P5 | P6 | P7 | P8 | P9 | P10 | P11 | P12 | P13 | P14 | Latency | Recip. throughput | Vectorization |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ADD W15, W13, W14 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| MADD W14, W14, W25, W12 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 1 | N/A |
| LDR S2, [X20, W15,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.33 | 0.33 | 0.33 | 0 | 0 | 6 | 0.33 | scal (12.5%) |
| LDR S3, [X22, W14,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.33 | 0.33 | 0.33 | 0 | 0 | 6 | 0.33 | scal (12.5%) |
| FCVT D2, S2 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (12.5%) |
| FCVT D3, S3 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (12.5%) |
| FMADD D1, D2, D3, D1 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 4 | 0.25 | scal (25.0%) |
| FCVT S1, D1 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 0.50 | scal (25.0%) |
| ADD W14, W13, W12 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| ADD X12, X12, #1 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | scal (25.0%) |
| CMP X12, X0 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 | scal (25.0%) |
| STR S1, [X21, W14,UXTW #2] | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.50 | 0.50 | 0 | 0 | 0.50 | 0.50 | 0 | 0 | 0 | 2 | 0.50 | scal (12.5%) |
| B.EQ 12544 <main+0x1544> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| CMP X0, #1 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.33 | N/A |
| B.EQ 12560 <main+0x1560> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |
| MOVI D1, #0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 2 | 0.25 | scal (25.0%) |
| ORR X14, XZR, XZR | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| ORR X15, XZR, X12 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | N/A |
| ORR W16, WZR, W11 | 1 | 0 | 0 | 0.25 | 0.25 | 0.25 | 0.25 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.25 | scal (12.5%) |
| HINT #0 | N/A | ||||||||||||||||||
| HINT #0 | N/A | ||||||||||||||||||
| HINT #0 | N/A | ||||||||||||||||||
| TBNZ W25, #0, 12568 <main+0x1568> | 1 | 0.50 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.50 | N/A |