deepseek-r1
best known per operation on the selected GPU
19 operations · 7 families · 4,432 eligible runs
0 of 19 operations have a usable best known on AMD Instinct MI355X · 19 without one
Best known on AMD Instinct MI355X
Operation · workload
Best known
Record
Trust
Observed
moe1 operation
AMD MXFP4 MoEsuite of 7 casessubmission 745202not usable69.9µsReported · license unknown · source2026-04-06
No entry in this cohort passes the usability policy (no install recipe, license unknown); the fastest known is shown. Measured by the source's own harness. Ranked by its metric under ranking-v1. 781 more entries in this cohort.
source mirroredlicense unknownno install recipeCohort on the operation page →View source →Vendor the source →Run detail →
mla-attention1 operation
AMD mixed-precision MLA decodesuite of 8 casessubmission 754087not usable8.05µsReported · license unknown · source2026-04-07
No entry in this cohort passes the usability policy (no install recipe, license unknown); the fastest known is shown. Measured by the source's own harness. Ranked by its metric under ranking-v1. 765 more entries in this cohort.
source mirroredlicense unknownno install recipeCohort on the operation page →View source →Vendor the source →Run detail →
Coverage gaps on AMD Instinct MI355X
Operation
Family
Status
mla-paged-attention
0 eligible runs on AMD Instinct MI355X · measured on NVIDIA B200
mla-paged-attention
0 eligible runs on AMD Instinct MI355X · measured on NVIDIA B200
mla-ragged
0 eligible runs on AMD Instinct MI355X · measured on NVIDIA B200
moe
0 eligible runs on AMD Instinct MI355X · measured on AMD Instinct MI300X
moe
0 eligible runs on AMD Instinct MI355X · measured on NVIDIA B200
mla-attention
measured · no usable entry (no install recipe, license unknown)
No eligible evidence yet for these operations. Challenges →
Related tags: deepseek-r1-0528
Sources: GPU MODE KernelBot (2026-04-07) · June 9 Researcher Reciprocity License v1.0 · FlashInfer-Bench (2026-04-06) · Apache-2.0How comparison worksJSON