Benchmarks 2024 11 26 TVM GCC O3 spike_rv64 - tum-ei-eda/muriscv-nn GitHub Wiki
Setup
Simulator
- Spike (
riscv-isa-sim
) (ISS, CPI=1)- Spike :
eb0a3e2b0a7c57522928be39de95cd9f8c6dc636
- Spike PK :
fix-gcc14-rvv
- Spike :
Toolchains
- RISC-V GCC:
- Scalar:
riscv32-unknown-elf-gcc (g8b4bb54e6c4) 14.2.1 20241118
- Vector:
riscv32-unknown-elf-gcc (g8b4bb54e6c4) 14.2.1 20241118
- Packed: Self compiled using patches found in https://github.com/riscv-collab/riscv-gcc/pull/258 and https://github.com/riscvarchive/riscv-binutils-gdb/pull/257
- Scalar:
Models
-
MLPerfTiny Benchmark
-
TODO: others!
Frameworks
-
MLonMCU :
develop
-
TVM :
Nightly Pre-Build
Miscellaneous
- Used
-Os
flag for compilation. - Benchmarks generated using MLonMCU deployment tool with minimal efforts.
- Memory metrics are reported in Bytes
Results (Framework: tvm, Backend: tvmaot, Toolchain: gcc, Flags: -O3, Target: spike_rv64 )
aww
)
Audio Wake Words (Cycles (Speedup) | Total ROM (rel.) | Total RAM (rel.) | VLEN | Layout | Kernels | Mode | Arch | Unroll | Auto-Vectorization |
---|---|---|---|---|---|---|---|---|---|
13511672.0 ( 1.2x ) |
121738 ( 1.297 ) |
61328 ( 2.909 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
20733186.0 ( 0.8x ) |
99988 ( 1.065 ) |
61328 ( 2.909 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
10096760.0 ( 1.6x ) |
144132 ( 1.536 ) |
61328 ( 2.909 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
10032731.0 ( 1.6x ) |
145468 ( 1.55 ) |
61344 ( 2.91 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
8296691.0 ( 1.9x ) |
139538 ( 1.487 ) |
61328 ( 2.909 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7754541.0 ( 2.1x ) |
137074 ( 1.461 ) |
61328 ( 2.909 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7917578.0 ( 2.0x ) |
137028 ( 1.46 ) |
61328 ( 2.909 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7915526.0 ( 2.0x ) |
137292 ( 1.463 ) |
61328 ( 2.909 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7865207.0 ( 2.0x ) |
109482 ( 1.167 ) |
61328 ( 2.909 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
4864793.0 ( 3.3x ) |
114598 ( 1.221 ) |
61328 ( 2.909 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
3353908.0 ( 4.8x ) |
107522 ( 1.146 ) |
61328 ( 2.909 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
2489591.0 ( 6.4x ) |
105514 ( 1.124 ) |
61328 ( 2.909 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
2114480.0 ( 7.6x ) |
101926 ( 1.086 ) |
61328 ( 2.909 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
2113093.0 ( 7.6x ) |
102258 ( 1.09 ) |
61328 ( 2.909 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
15975582.0 ( Base ) |
93852 ( Base ) |
21080 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
15989442.0 ( 1.0x ) |
90798 ( 0.967 ) |
25544 ( 1.212 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
8895421.0 ( 1.8x ) |
108020 ( 1.151 ) |
21080 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
8959671.0 ( 1.8x ) |
108808 ( 1.159 ) |
21080 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
8700589.0 ( 1.8x ) |
109162 ( 1.163 ) |
21080 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
8515796.0 ( 1.9x ) |
110706 ( 1.18 ) |
21080 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
8265442.0 ( 1.9x ) |
113124 ( 1.205 ) |
21080 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
8007224.0 ( 2.0x ) |
116680 ( 1.243 ) |
21080 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
9088588.0 ( 1.8x ) |
96100 ( 1.024 ) |
25544 ( 1.212 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
6810301.0 ( 2.3x ) |
96088 ( 1.024 ) |
25544 ( 1.212 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
5599038.0 ( 2.9x ) |
96100 ( 1.024 ) |
25544 ( 1.212 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
5562855.0 ( 2.9x ) |
96110 ( 1.024 ) |
25544 ( 1.212 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
5564717.0 ( 2.9x ) |
96096 ( 1.024 ) |
25544 ( 1.212 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
5566584.0 ( 2.9x ) |
96108 ( 1.024 ) |
25544 ( 1.212 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8339460.0 ( 1.9x ) |
105822 ( 1.128 ) |
25544 ( 1.212 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
7727170.0 ( 2.1x ) |
105966 ( 1.129 ) |
25544 ( 1.212 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
7187335.0 ( 2.2x ) |
106388 ( 1.134 ) |
25544 ( 1.212 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6730347.0 ( 2.4x ) |
107922 ( 1.15 ) |
25544 ( 1.212 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6470226.0 ( 2.5x ) |
110326 ( 1.176 ) |
25544 ( 1.212 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6212510.0 ( 2.6x ) |
113880 ( 1.213 ) |
25544 ( 1.212 ) |
4096 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
resnet
)
Image Classification (Cycles (Speedup) | Total ROM (rel.) | Total RAM (rel.) | VLEN | Layout | Kernels | Mode | Arch | Unroll | Auto-Vectorization |
---|---|---|---|---|---|---|---|---|---|
56632189.0 ( 1.0x ) |
224702 ( 1.632 ) |
110240 ( 1.921 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
76261064.0 ( 0.7x ) |
207576 ( 1.507 ) |
110240 ( 1.921 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
31133846.0 ( 1.7x ) |
236952 ( 1.721 ) |
110240 ( 1.921 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
33688790.0 ( 1.6x ) |
235992 ( 1.714 ) |
110256 ( 1.921 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
28404147.0 ( 1.9x ) |
234988 ( 1.706 ) |
110240 ( 1.921 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
29094485.0 ( 1.9x ) |
230628 ( 1.675 ) |
110240 ( 1.921 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
29508402.0 ( 1.8x ) |
228100 ( 1.656 ) |
110240 ( 1.921 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
29501932.0 ( 1.8x ) |
228296 ( 1.658 ) |
110240 ( 1.921 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
47313821.0 ( 1.1x ) |
209312 ( 1.52 ) |
110240 ( 1.921 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
23760370.0 ( 2.3x ) |
211114 ( 1.533 ) |
110240 ( 1.921 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
16653762.0 ( 3.3x ) |
213156 ( 1.548 ) |
110240 ( 1.921 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
30388900.0 ( 1.8x ) |
211018 ( 1.532 ) |
110240 ( 1.921 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
49254100.0 ( 1.1x ) |
209124 ( 1.519 ) |
110240 ( 1.921 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
73973850.0 ( 0.7x ) |
207908 ( 1.51 ) |
110240 ( 1.921 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
54274852.0 ( Base ) |
137704 ( Base ) |
57384 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
71752662.0 ( 0.8x ) |
137038 ( 0.995 ) |
57384 ( 1.0 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
17939078.0 ( 3.0x ) |
152240 ( 1.106 ) |
57384 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
11425909.0 ( 4.8x ) |
152338 ( 1.106 ) |
57384 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
8231462.0 ( 6.6x ) |
152304 ( 1.106 ) |
57384 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
6739916.0 ( 8.1x ) |
152290 ( 1.106 ) |
57384 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
6059365.0 ( 9.0x ) |
152316 ( 1.106 ) |
57384 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
5727599.0 ( 9.5x ) |
152336 ( 1.106 ) |
57384 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
31989492.0 ( 1.7x ) |
144176 ( 1.047 ) |
57384 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
19583162.0 ( 2.8x ) |
144168 ( 1.047 ) |
57384 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
13950358.0 ( 3.9x ) |
144146 ( 1.047 ) |
57384 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
11204088.0 ( 4.8x ) |
144170 ( 1.047 ) |
57384 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
9048522.0 ( 6.0x ) |
144174 ( 1.047 ) |
57384 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8470693.0 ( 6.4x ) |
144182 ( 1.047 ) |
57384 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
29171431.0 ( 1.9x ) |
151546 ( 1.101 ) |
57384 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
17068861.0 ( 3.2x ) |
151580 ( 1.101 ) |
57384 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
10958411.0 ( 5.0x ) |
151624 ( 1.101 ) |
57384 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
8088806.0 ( 6.7x ) |
151566 ( 1.101 ) |
57384 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6778484.0 ( 8.0x ) |
151568 ( 1.101 ) |
57384 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6139517.0 ( 8.8x ) |
151584 ( 1.101 ) |
57384 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
toycar
)
Anomaly Detection (Cycles (Speedup) | Total ROM (rel.) | Total RAM (rel.) | VLEN | Layout | Kernels | Mode | Arch | Unroll | Auto-Vectorization |
---|---|---|---|---|---|---|---|---|---|
1643134.0 ( 1.0x ) |
584258 ( 1.872 ) |
7416 ( 1.121 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
1643134.0 ( 1.0x ) |
584258 ( 1.872 ) |
7416 ( 1.121 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
880851.0 ( 1.9x ) |
593736 ( 1.903 ) |
7416 ( 1.121 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
790655.0 ( 2.1x ) |
592568 ( 1.899 ) |
7416 ( 1.121 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
742308.0 ( 2.3x ) |
588962 ( 1.887 ) |
7416 ( 1.121 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
952259.0 ( 1.8x ) |
624224 ( 2.0 ) |
7416 ( 1.121 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
748413.0 ( 2.2x ) |
594282 ( 1.904 ) |
7416 ( 1.121 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
769138.0 ( 2.2x ) |
597032 ( 1.913 ) |
7416 ( 1.121 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
880851.0 ( 1.9x ) |
593736 ( 1.903 ) |
7416 ( 1.121 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
790655.0 ( 2.1x ) |
592568 ( 1.899 ) |
7416 ( 1.121 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
742308.0 ( 2.3x ) |
588962 ( 1.887 ) |
7416 ( 1.121 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
952259.0 ( 1.8x ) |
624224 ( 2.0 ) |
7416 ( 1.121 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
748413.0 ( 2.2x ) |
594282 ( 1.904 ) |
7416 ( 1.121 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
769138.0 ( 2.2x ) |
597032 ( 1.913 ) |
7416 ( 1.121 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1678619.0 ( Base ) |
312052 ( Base ) |
6616 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
1678625.0 ( 1.0x ) |
312058 ( 1.0 ) |
6616 ( 1.0 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
765431.0 ( 2.2x ) |
312824 ( 1.002 ) |
6616 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
583637.0 ( 2.9x ) |
312822 ( 1.002 ) |
6616 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
493021.0 ( 3.4x ) |
312928 ( 1.003 ) |
6616 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
447785.0 ( 3.7x ) |
312848 ( 1.003 ) |
6616 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
425172.0 ( 3.9x ) |
312828 ( 1.002 ) |
6616 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
413864.0 ( 4.1x ) |
312812 ( 1.002 ) |
6616 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
971179.0 ( 1.7x ) |
314348 ( 1.007 ) |
6616 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
711170.0 ( 2.4x ) |
314352 ( 1.007 ) |
6616 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
581167.0 ( 2.9x ) |
314348 ( 1.007 ) |
6616 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
516943.0 ( 3.2x ) |
314344 ( 1.007 ) |
6616 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
510717.0 ( 3.3x ) |
314352 ( 1.007 ) |
6616 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
508437.0 ( 3.3x ) |
314354 ( 1.007 ) |
6616 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
765431.0 ( 2.2x ) |
312828 ( 1.002 ) |
6616 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
583639.0 ( 2.9x ) |
312826 ( 1.002 ) |
6616 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
493019.0 ( 3.4x ) |
312924 ( 1.003 ) |
6616 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
447782.0 ( 3.7x ) |
312846 ( 1.003 ) |
6616 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
425172.0 ( 3.9x ) |
312830 ( 1.002 ) |
6616 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
413861.0 ( 4.1x ) |
312804 ( 1.002 ) |
6616 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
vww
)
Visual Wake Words (Cycles (Speedup) | Total ROM (rel.) | Total RAM (rel.) | VLEN | Layout | Kernels | Mode | Arch | Unroll | Auto-Vectorization |
---|---|---|---|---|---|---|---|---|---|
40551985.0 ( 1.1x ) |
581980 ( 1.779 ) |
182840 ( 2.089 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
57502325.0 ( 0.8x ) |
522892 ( 1.598 ) |
182840 ( 2.089 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
25951548.0 ( 1.8x ) |
636484 ( 1.946 ) |
182840 ( 2.089 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
27267072.0 ( 1.7x ) |
635414 ( 1.942 ) |
182856 ( 2.089 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
22538563.0 ( 2.0x ) |
623666 ( 1.906 ) |
182840 ( 2.089 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
21027228.0 ( 2.2x ) |
605246 ( 1.85 ) |
182840 ( 2.089 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
20864202.0 ( 2.2x ) |
590850 ( 1.806 ) |
182840 ( 2.089 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
20920371.0 ( 2.2x ) |
590548 ( 1.805 ) |
182840 ( 2.089 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
25876909.0 ( 1.8x ) |
545438 ( 1.667 ) |
182840 ( 2.089 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
18305974.0 ( 2.5x ) |
554390 ( 1.695 ) |
182840 ( 2.089 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
14375562.0 ( 3.2x ) |
571156 ( 1.746 ) |
182840 ( 2.089 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
12468767.0 ( 3.7x ) |
553384 ( 1.692 ) |
182840 ( 2.089 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
15555948.0 ( 2.9x ) |
543748 ( 1.662 ) |
182840 ( 2.089 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
15531828.0 ( 2.9x ) |
545956 ( 1.669 ) |
182840 ( 2.089 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
45528018.0 ( Base ) |
327142 ( Base ) |
87520 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
46833650.0 ( 1.0x ) |
324180 ( 0.991 ) |
87520 ( 1.0 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
25427730.0 ( 1.8x ) |
342494 ( 1.047 ) |
87520 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
21368084.0 ( 2.1x ) |
343096 ( 1.049 ) |
87520 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
19380866.0 ( 2.3x ) |
343464 ( 1.05 ) |
87520 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
18442187.0 ( 2.5x ) |
344988 ( 1.055 ) |
87520 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
18068115.0 ( 2.5x ) |
347422 ( 1.062 ) |
87520 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
17932805.0 ( 2.5x ) |
350962 ( 1.073 ) |
87520 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
25398515.0 ( 1.8x ) |
329414 ( 1.007 ) |
87520 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
18530087.0 ( 2.5x ) |
329402 ( 1.007 ) |
87520 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
15908678.0 ( 2.9x ) |
329378 ( 1.007 ) |
87520 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
14815866.0 ( 3.1x ) |
329390 ( 1.007 ) |
87520 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
14728700.0 ( 3.1x ) |
329394 ( 1.007 ) |
87520 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
14730565.0 ( 3.1x ) |
329388 ( 1.007 ) |
87520 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
24361953.0 ( 1.9x ) |
340442 ( 1.041 ) |
87520 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
18213324.0 ( 2.5x ) |
340418 ( 1.041 ) |
87520 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
15324435.0 ( 3.0x ) |
340842 ( 1.042 ) |
87520 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
13538455.0 ( 3.4x ) |
342418 ( 1.047 ) |
87520 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
13163429.0 ( 3.5x ) |
344812 ( 1.054 ) |
87520 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
13016921.0 ( 3.5x ) |
348344 ( 1.065 ) |
87520 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
Original data
Click here to download the raw files for this benchmark.