IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /max/get-started.md). For the complete documentation index, see llms.txt.
Skip to main content
For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo function

enqueue_fp8_materialize

def enqueue_fp8_materialize[out_type: DType](out_w: TileTensor[out_type, Storage=out_w.Storage, address_space=out_w.address_space, linear_idx_type=out_w.linear_idx_type], weight: TileTensor[DType.float8_e4m3fn, Storage=weight.Storage, address_space=weight.address_space, linear_idx_type=weight.linear_idx_type], ctx: DeviceContext)

Enqueues fp8_materialize_kernel: FP8 weight -> dense [N, K] out_type.

out_w is [N, K], weight is [N, K]. The grid is (ceil(K/16), ceil(N/16)) threadgroups of 16x16 threads; bounds are checked per thread so ragged K/N are fine. The FP8 sibling of enqueue_fp4_materialize.

Was this page helpful?