Benchmarks 2024 11 26 TVM LLVM Os 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:
-
LLVM/Clang:
clang version 18.1.8 (https://github.com/llvm/llvm-project.git 3b5b5c1ec4a3095ab096dd780e84d7ab81f3d7ff)
- Linker: lld (TODO)
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: llvm, Flags: -Os, Target: spike_rv64 )
aww
)
Audio Wake Words (Cycles (Speedup) | Total ROM (rel.) | Total RAM (rel.) | VLEN | Layout | Kernels | Mode | Arch | Unroll | Auto-Vectorization |
---|---|---|---|---|---|---|---|---|---|
28356271.0 ( 0.6x ) |
104722 ( 1.208 ) |
61368 ( 2.915 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
24004451.0 ( 0.7x ) |
99776 ( 1.15 ) |
61288 ( 2.911 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
5890709.0 ( 2.7x ) |
103714 ( 1.196 ) |
61296 ( 2.911 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
5796380.0 ( 2.7x ) |
103358 ( 1.192 ) |
61296 ( 2.911 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
5792380.0 ( 2.7x ) |
103318 ( 1.191 ) |
61296 ( 2.911 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
5848725.0 ( 2.7x ) |
103574 ( 1.194 ) |
61296 ( 2.911 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
5847925.0 ( 2.7x ) |
103534 ( 1.194 ) |
61296 ( 2.911 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
5892724.0 ( 2.7x ) |
103818 ( 1.197 ) |
61296 ( 2.911 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
9587040.0 ( 1.6x ) |
100828 ( 1.163 ) |
61288 ( 2.911 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7929144.0 ( 2.0x ) |
100928 ( 1.164 ) |
61288 ( 2.911 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7100055.0 ( 2.2x ) |
100898 ( 1.163 ) |
61288 ( 2.911 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7843373.0 ( 2.0x ) |
119332 ( 1.376 ) |
61288 ( 2.911 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7854560.0 ( 2.0x ) |
119322 ( 1.376 ) |
61288 ( 2.911 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
7878894.0 ( 2.0x ) |
119360 ( 1.376 ) |
61288 ( 2.911 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
15705418.0 ( Base ) |
86726 ( Base ) |
21056 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
15650160.0 ( 1.0x ) |
85928 ( 0.991 ) |
25520 ( 1.212 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
7146391.0 ( 2.2x ) |
89814 ( 1.036 ) |
21056 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
6265600.0 ( 2.5x ) |
89348 ( 1.03 ) |
21056 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
5827896.0 ( 2.7x ) |
89348 ( 1.03 ) |
21056 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
5610909.0 ( 2.8x ) |
89348 ( 1.03 ) |
21056 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
5610909.0 ( 2.8x ) |
89348 ( 1.03 ) |
21056 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
5614639.0 ( 2.8x ) |
89348 ( 1.03 ) |
21056 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
6017576.0 ( 2.6x ) |
88016 ( 1.015 ) |
25520 ( 1.212 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4883519.0 ( 3.2x ) |
87590 ( 1.01 ) |
25520 ( 1.212 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4325223.0 ( 3.6x ) |
87590 ( 1.01 ) |
25520 ( 1.212 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4286092.0 ( 3.7x ) |
87590 ( 1.01 ) |
25520 ( 1.212 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4286092.0 ( 3.7x ) |
87590 ( 1.01 ) |
25520 ( 1.212 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4287957.0 ( 3.7x ) |
87590 ( 1.01 ) |
25520 ( 1.212 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8326965.0 ( 1.9x ) |
88586 ( 1.021 ) |
25520 ( 1.212 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
7451311.0 ( 2.1x ) |
88160 ( 1.017 ) |
25520 ( 1.212 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
7013607.0 ( 2.2x ) |
88160 ( 1.017 ) |
25520 ( 1.212 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6796620.0 ( 2.3x ) |
88160 ( 1.017 ) |
25520 ( 1.212 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6796620.0 ( 2.3x ) |
88160 ( 1.017 ) |
25520 ( 1.212 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
6800350.0 ( 2.3x ) |
88160 ( 1.017 ) |
25520 ( 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 |
---|---|---|---|---|---|---|---|---|---|
123435843.0 ( 0.5x ) |
212766 ( 1.585 ) |
110216 ( 1.921 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
100169412.0 ( 0.6x ) |
205998 ( 1.535 ) |
110216 ( 1.921 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
24209094.0 ( 2.3x ) |
212044 ( 1.58 ) |
110216 ( 1.921 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
24198053.0 ( 2.3x ) |
212020 ( 1.58 ) |
110216 ( 1.921 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
24042297.0 ( 2.3x ) |
211574 ( 1.576 ) |
110216 ( 1.921 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
24036663.0 ( 2.3x ) |
211584 ( 1.576 ) |
110216 ( 1.921 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
24047583.0 ( 2.3x ) |
211638 ( 1.577 ) |
110216 ( 1.921 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
24216233.0 ( 2.3x ) |
212154 ( 1.581 ) |
110216 ( 1.921 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
35933045.0 ( 1.6x ) |
206850 ( 1.541 ) |
110216 ( 1.921 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
34755076.0 ( 1.6x ) |
207750 ( 1.548 ) |
110216 ( 1.921 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
34013415.0 ( 1.6x ) |
208056 ( 1.55 ) |
110216 ( 1.921 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
34291886.0 ( 1.6x ) |
211168 ( 1.573 ) |
110216 ( 1.921 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
34303075.0 ( 1.6x ) |
211160 ( 1.573 ) |
110216 ( 1.921 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
34332536.0 ( 1.6x ) |
211194 ( 1.573 ) |
110216 ( 1.921 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
55736927.0 ( Base ) |
134226 ( Base ) |
57376 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
71562670.0 ( 0.8x ) |
133618 ( 0.995 ) |
57376 ( 1.0 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
25741863.0 ( 2.2x ) |
138106 ( 1.029 ) |
57376 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
17778716.0 ( 3.1x ) |
137588 ( 1.025 ) |
57376 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
14046096.0 ( 4.0x ) |
137588 ( 1.025 ) |
57376 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
12347539.0 ( 4.5x ) |
137588 ( 1.025 ) |
57376 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
11519964.0 ( 4.8x ) |
137588 ( 1.025 ) |
57376 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
10899749.0 ( 5.1x ) |
137588 ( 1.025 ) |
57376 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
15261403.0 ( 3.7x ) |
136156 ( 1.014 ) |
57376 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
9678205.0 ( 5.8x ) |
135638 ( 1.011 ) |
57376 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
7132365.0 ( 7.8x ) |
135638 ( 1.011 ) |
57376 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
5890237.0 ( 9.5x ) |
135638 ( 1.011 ) |
57376 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4963838.0 ( 11.2x ) |
135638 ( 1.011 ) |
57376 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
4716611.0 ( 11.8x ) |
135638 ( 1.011 ) |
57376 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
19854766.0 ( 2.8x ) |
136814 ( 1.019 ) |
57376 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
13562787.0 ( 4.1x ) |
136296 ( 1.015 ) |
57376 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
10613527.0 ( 5.3x ) |
136296 ( 1.015 ) |
57376 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
9271834.0 ( 6.0x ) |
136296 ( 1.015 ) |
57376 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
8616474.0 ( 6.5x ) |
136296 ( 1.015 ) |
57376 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
8128684.0 ( 6.9x ) |
136296 ( 1.015 ) |
57376 ( 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 |
---|---|---|---|---|---|---|---|---|---|
3202427.0 ( 0.5x ) |
578702 ( 1.851 ) |
7408 ( 1.121 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
3202427.0 ( 0.5x ) |
578702 ( 1.851 ) |
7408 ( 1.121 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
1041694.0 ( 1.6x ) |
577898 ( 1.848 ) |
7408 ( 1.121 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1041457.0 ( 1.6x ) |
577774 ( 1.848 ) |
7408 ( 1.121 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1041694.0 ( 1.6x ) |
577898 ( 1.848 ) |
7408 ( 1.121 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1041457.0 ( 1.6x ) |
577774 ( 1.848 ) |
7408 ( 1.121 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1013976.0 ( 1.6x ) |
578660 ( 1.85 ) |
7408 ( 1.121 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
1662786.0 ( Base ) |
312718 ( Base ) |
6608 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
1662787.0 ( 1.0x ) |
312720 ( 1.0 ) |
6608 ( 1.0 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
628443.0 ( 2.6x ) |
313474 ( 1.002 ) |
6608 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
510980.0 ( 3.3x ) |
313058 ( 1.001 ) |
6608 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
452260.0 ( 3.7x ) |
313058 ( 1.001 ) |
6608 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
422901.0 ( 3.9x ) |
313060 ( 1.001 ) |
6608 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
408219.0 ( 4.1x ) |
313056 ( 1.001 ) |
6608 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
406381.0 ( 4.1x ) |
313066 ( 1.001 ) |
6608 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
625472.0 ( 2.7x ) |
314186 ( 1.005 ) |
6608 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
501954.0 ( 3.3x ) |
313772 ( 1.003 ) |
6608 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
442769.0 ( 3.8x ) |
313772 ( 1.003 ) |
6608 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
413514.0 ( 4.0x ) |
313772 ( 1.003 ) |
6608 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
409831.0 ( 4.1x ) |
313778 ( 1.003 ) |
6608 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
407949.0 ( 4.1x ) |
313780 ( 1.003 ) |
6608 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
628443.0 ( 2.6x ) |
313474 ( 1.002 ) |
6608 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
510980.0 ( 3.3x ) |
313058 ( 1.001 ) |
6608 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
452259.0 ( 3.7x ) |
313056 ( 1.001 ) |
6608 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
422900.0 ( 3.9x ) |
313058 ( 1.001 ) |
6608 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
408221.0 ( 4.1x ) |
313060 ( 1.001 ) |
6608 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
406381.0 ( 4.1x ) |
313060 ( 1.001 ) |
6608 ( 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 |
---|---|---|---|---|---|---|---|---|---|
81866644.0 ( 0.5x ) |
536648 ( 1.679 ) |
182976 ( 2.091 ) |
0 | NCHW | TVM | Fallback | RV64GC | 0 | - |
69686529.0 ( 0.6x ) |
519012 ( 1.623 ) |
182816 ( 2.089 ) |
0 | NHWC | TVM | Fallback | RV64GC | 0 | - |
18255670.0 ( 2.4x ) |
532050 ( 1.664 ) |
182840 ( 2.089 ) |
128 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
18199896.0 ( 2.4x ) |
531504 ( 1.662 ) |
182840 ( 2.089 ) |
256 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
18179623.0 ( 2.4x ) |
531288 ( 1.662 ) |
182840 ( 2.089 ) |
512 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
17986729.0 ( 2.5x ) |
530490 ( 1.659 ) |
182840 ( 2.089 ) |
1024 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
18012143.0 ( 2.4x ) |
530716 ( 1.66 ) |
182840 ( 2.089 ) |
2048 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
18268689.0 ( 2.4x ) |
532208 ( 1.665 ) |
182840 ( 2.089 ) |
4096 | NCHW | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
28854346.0 ( 1.5x ) |
521786 ( 1.632 ) |
182816 ( 2.089 ) |
128 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
24036439.0 ( 1.8x ) |
521668 ( 1.632 ) |
182816 ( 2.089 ) |
256 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
22239138.0 ( 2.0x ) |
522704 ( 1.635 ) |
182816 ( 2.089 ) |
512 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
22221099.0 ( 2.0x ) |
550762 ( 1.723 ) |
182816 ( 2.089 ) |
1024 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
22424692.0 ( 2.0x ) |
578930 ( 1.811 ) |
182816 ( 2.089 ) |
2048 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
22522523.0 ( 2.0x ) |
589050 ( 1.842 ) |
182816 ( 2.089 ) |
4096 | NHWC | TVM | Fallback | RV64GCV | 0 | Loop+SLP |
44083240.0 ( Base ) |
319714 ( Base ) |
87512 ( Base ) |
0 | NHWC | muRISCV-NN | Scalar | RV64GC | 0 | - |
45097701.0 ( 1.0x ) |
319032 ( 0.998 ) |
87512 ( 1.0 ) |
0 | NHWC | muRISCV-NN | Vector (Portable) | RV64GC | 0 | - |
18474951.0 ( 2.4x ) |
323724 ( 1.013 ) |
87512 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
15487671.0 ( 2.8x ) |
322328 ( 1.008 ) |
87512 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
14061513.0 ( 3.1x ) |
322336 ( 1.008 ) |
87512 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
13537305.0 ( 3.3x ) |
322332 ( 1.008 ) |
87512 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
13323558.0 ( 3.3x ) |
322332 ( 1.008 ) |
87512 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
13307090.0 ( 3.3x ) |
322336 ( 1.008 ) |
87512 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Scalar | RV64GCV | 0 | Loop+SLP |
13473675.0 ( 3.3x ) |
321952 ( 1.007 ) |
87512 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
10152615.0 ( 4.3x ) |
320696 ( 1.003 ) |
87512 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8861159.0 ( 5.0x ) |
320696 ( 1.003 ) |
87512 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8359020.0 ( 5.3x ) |
320696 ( 1.003 ) |
87512 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8311245.0 ( 5.3x ) |
320696 ( 1.003 ) |
87512 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
8313110.0 ( 5.3x ) |
320696 ( 1.003 ) |
87512 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector | RV64GCV | 0 | - |
22084820.0 ( 2.0x ) |
322522 ( 1.009 ) |
87512 ( 1.0 ) |
128 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
19186160.0 ( 2.3x ) |
321266 ( 1.005 ) |
87512 ( 1.0 ) |
256 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
17799168.0 ( 2.5x ) |
321266 ( 1.005 ) |
87512 ( 1.0 ) |
512 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
17274961.0 ( 2.6x ) |
321266 ( 1.005 ) |
87512 ( 1.0 ) |
1024 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
17059348.0 ( 2.6x ) |
321262 ( 1.005 ) |
87512 ( 1.0 ) |
2048 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
17044745.0 ( 2.6x ) |
321266 ( 1.005 ) |
87512 ( 1.0 ) |
4096 | NHWC | muRISCV-NN | Vector (Portable) | RV64GCV | 0 | Loop+SLP |
Original data
Click here to download the raw files for this benchmark.