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 struct
Struct_ep_dispatch_mxfp4
struct Struct_ep_dispatch_mxfp4
Registers the ep.dispatch.mxfp4 graph op with the graph compiler.
Implemented traits
Methods
execute
static def execute[input_dtype: DType, dispatch_dtype: DType, dispatch_scale_dtype: DType, hidden_size: Int, top_k: Int, n_experts: Int, max_token_per_rank: Int, n_gpus_per_node: Int, n_nodes: Int, fused_shared_expert: Bool, skip_a2a: Bool, allreduce_world_size: Int, //, target: StringSlice[ImmStaticOrigin], *, fuse_a_scale_preshuffle: Bool = False, max_padded_M: Int = Int(0)](output_tokens: ManagedTensorSlice[IOSpec[_, _].Output, static_spec=output_tokens.static_spec], output_scales: ManagedTensorSlice[IOSpec[_, _].Output, static_spec=output_scales.static_spec], row_offsets: ManagedTensorSlice[IOSpec[_, _].Output, static_spec=row_offsets.static_spec], expert_ids: ManagedTensorSlice[IOSpec[_, _].Output, static_spec=expert_ids.static_spec], src_info: ManagedTensorSlice[IOSpec[_, _].Output, static_spec=src_info.static_spec], atomic_counters: ManagedTensorSlice[IOSpec[_, _].MutableInput, static_spec=atomic_counters.static_spec], input_tokens: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=input_tokens.static_spec], topk_ids: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=topk_ids.static_spec], send_ptrs: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=send_ptrs.static_spec], recv_ptrs: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=recv_ptrs.static_spec], recv_count_ptrs: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=recv_count_ptrs.static_spec], context: DeviceContext)
Execute the fused Expert Parallelism MXFP4 dispatch kernel. Tokens are dispatched in MXFP4 format.
When fuse_a_scale_preshuffle=True (KS224 up-proj fusion), the
wait-side copy writes the E8M0 activation scale directly into the
up-proj grouped matmul's per-expert fixed-stride scale_4d slot
layout (slot stride max_padded_M * K_SCALES), dropping the standalone
preshuffle. The scales output tensor must then be slot-sized
([n_local_experts * max_padded_M, K_SCALES]).
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!