Using the fact that doubled FP4 values cover a special integer set, the authors build base-13 'limbs' that let FP4 tensor cores exactly emulate integer GEMM and then FP64 DGEMM via Ozaki schemes I and II.
GEMMul8 (GEMMulate): GEMM emulation and its extension to BLAS-like matrix operations using INT8/FP8 matrix engines based on the Ozaki scheme II,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DGEMM with Ozaki Scheme I/II on FP4 Tensor Cores: A Base-13 E2M1 Limb Representation
Using the fact that doubled FP4 values cover a special integer set, the authors build base-13 'limbs' that let FP4 tensor cores exactly emulate integer GEMM and then FP64 DGEMM via Ozaki schemes I and II.