KernelBot user 44
Competition author · gpumode-kernelbot/user/44 · license unknown
Records held
14as of 2026-10-01 09:09 UTC
Entries
26
Runs
191
Is this you? Claim this project ›
A claim grants attribution and metadata maintenance, never the right to edit evidence. Evidence claims are reviewed by a maintainer.
Current records
Operation · workload
History
Record
Implementation
Hardware
Since
AMD DeepSeek-style MoE layerfp16 · [1, 8192, 7168]
10.4ms
0.7% faster · was 10.4 ms
AMD Instinct MI300X
2025-06-03
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [1024, 7168] · n 512 · seed 6563
50.1µs
0.5% faster · was 50.4 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 1536]
112.6µs
0.4% faster · was 113.1 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 7168] · n 1536 · seed 6543
214.5µs
0.5% faster · was 215.5 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [1024, 2304]
79.4µs
1.3% faster · was 80.4 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [1024, 2048]
76.6µs
1.4% faster · was 77.7 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [1024, 7168] · n 576 · seed 12346
55.1µs
0.7% faster · was 55.5 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [1024, 7168] · n 1536 · seed 8135
79.0µs
0.8% faster · was 79.6 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 512]
74.5µs
0.1% faster · was 74.6 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 2304]
308.7µs
0.1% faster · was 308.9 µs
AMD Instinct MI300X
2025-06-02
Show all 14 records ›Showing all 14 records ⌄
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 7168] · n 4608 · seed 65436
566.3µs
0.2% faster · was 567.5 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 2048]
299.2µs
0.0% faster · was 299.3 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [6144, 256]
77.4µs
0.1% faster · was 77.5 µs
AMD Instinct MI300X
2025-06-02
AMD FP8 blockwise GEMMfp32/fp8_e4m3fnuz · [1024, 7168] · n 4608 · seed 7531
142.2µs
0.2% faster · was 142.4 µs
AMD Instinct MI300X
2025-06-02
Measured entries
Implementation
Operation
Runtime
Best latency
Evidence
Availability
Show all 26 entries ›Showing all 26 entries ⌄
amd-mixture-of-experts · submission 30960KernelBot user 44
python
4.23ms
Reported
license unknown · source
amd-mixture-of-experts · submission 30957KernelBot user 44
python
4.30ms
Reported
license unknown · source
amd-mixture-of-experts · submission 30943KernelBot user 44
python
4.49ms
Reported
license unknown · source
amd-mixture-of-experts · submission 30429KernelBot user 44
python
4.67ms
Reported
license unknown · source
amd-mixture-of-experts · submission 30424KernelBot user 44
python
4.72ms
Reported
license unknown · source
amd-mixture-of-experts · submission 30076KernelBot user 44
python
4.80ms
Reported
license unknown · source
amd-mixture-of-experts · submission 29947KernelBot user 44
python
4.93ms
Reported
license unknown · source
amd-mixture-of-experts · submission 30027KernelBot user 44
python
4.94ms
Reported
license unknown · source
amd-mixture-of-experts · submission 29937KernelBot user 44
python
5.16ms
Reported
license unknown · source
amd-mixture-of-experts · submission 29919KernelBot user 44
python
5.31ms
Reported
license unknown · source
amd-mixture-of-experts · submission 29927KernelBot user 44
python
5.31ms
Reported
license unknown · source
amd-mixture-of-experts · submission 29917KernelBot user 44
python
5.32ms
Reported
license unknown · source
Record activity
No record transitions on this GPU in the trailing year.
Sources: GPU MODE KernelBot (2025-06-03) · June 9 Researcher Reciprocity License v1.0last observed 2025-06-03JSON