0
Fork 0
mirror of https://github.com/zml/zml.git synced 2026-11-16 19:11:09 +00:00
zml/platforms
Kevin Barre c40fa52d2d
zml/moe [Platform:tcl]: support block-128 FP8 SwiGLU experts (#1053)
- Select TCL for plain E4M3FN block-128 experts with matching BF16 or
  F32 scales. Preserve rejection of unsupported quantization.

- Quantize activations per block and apply checkpoint scales during
  native projections. Keep BF16 materialization and F32 down and
  routing accumulation.

- Reuse local expert routing and AllReduce for expert parallelism.
  Choose expert-stationary for small FP8 token counts and use the
  existing activation-stationary and blockwise recipes for larger
  counts.

- Share FP8 and BF16 dispatch preparation while retaining the token
  padding and BF16 SiLU rounding workarounds.
History 2026-11-12 13:47:48 +01:00
..
cpu zml/kernels/flash_attention/metal: port the kernels to the Metal DSL (#1054) 2026-11-10 18:51:08 +01:00
cuda zml/kernels/flash_attention/metal: port the kernels to the Metal DSL (#1054) 2026-11-10 18:51:08 +01:00
furiosa zml/moe [Platform:tcl]: support block-128 FP8 SwiGLU experts (#1053) 2026-11-12 13:47:48 +01:00
metal zml/kernels/flash_attention/metal: port the kernels to the Metal DSL (#1054) 2026-11-10 18:51:08 +01:00
neuron ci: add supported platforms with bumped and audited deps (#967) 2026-10-20 14:06:49 +02:00
oneapi platforms/oneapi: bump to IGC 2.41.15 (#996) 2026-10-29 11:15:39 +01:00
rocm zml/kernels/flash_attention/metal: port the kernels to the Metal DSL (#1054) 2026-11-10 18:51:08 +01:00
tpu dsl: unify the kernel DSLs (#991) 2026-10-27 17:53:31 +01:00
BUILD.bazel zml/platform: Add furiosa platform (#871) 2026-10-09 12:21:16 +02:00
packages.bzl platforms/cuda: add arm64 support (#493) 2026-04-22 12:37:48 +02:00
platforms.zig zig 0.17 (second attempt) (#923) 2026-10-13 17:11:38 +02:00