ViperEkura
4244df2785
perf: pure FP8 fwd/bwd and lean non-transposed GEMM
- drop the fused kernel; forward/backward are quantize + a pre-quantized GEMM
- rename module fp8_mm -> fp8_ops (mm.cu -> ops.cu)
- kernels/launchers fp8_gemm_kernel / launch_fp8_gemm; drop PqTraits/gather_trans/pack_fp8x4_vector
- remove the in-kernel transposed-operand branches (TransA/TransB)
- backward: quantize g once (amax_g here), explicit fp8 transposes, fast non-transposed GEMMs (dX = g@w^T, dW = g^T@x^T)
- each pass uses a single FP8 format (E4M3 fwd / E5M2 bwd)
2026-08-23 15:38:30 +08:00
..
2026-08-19 20:55:13 +08:00
2026-08-19 16:17:40 +08:00
2026-08-23 15:38:30 +08:00
2026-08-23 14:47:02 +08:00
2026-08-20 11:34:59 +08:00
2026-08-01 09:20:54 +08:00
2026-08-19 16:17:40 +08:00
2026-08-05 22:20:29 +08:00
2026-08-22 21:08:06 +08:00
2026-07-30 09:38:20 +08:00
2026-08-20 11:34:59 +08:00
2026-08-19 20:55:13 +08:00
2026-07-30 09:38:20 +08:00
2026-08-20 01:37:00 +08:00
2026-05-24 20:35:44 +08:00
2026-07-28 10:36:17 +08:00