[Kernel] Integrate CUTLASS MoE kernel with PPLX (#18762)

Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
This commit is contained in:
ElizaWszola
2025-06-06 18:26:11 -07:00
committed by GitHub
co-authored by Tyler Michael Smith
parent 6e0cd10f72
commit 84166fee97
26 changed files with 917 additions and 408 deletions
+10 -1
View File
@@ -236,7 +236,8 @@ void cutlass_moe_mm(
torch::Tensor const& b_tensors, torch::Tensor const& a_scales,
torch::Tensor const& b_scales, torch::Tensor const& expert_offsets,
torch::Tensor const& problem_sizes, torch::Tensor const& a_strides,
torch::Tensor const& b_strides, torch::Tensor const& c_strides);
torch::Tensor const& b_strides, torch::Tensor const& c_strides,
bool per_act_token, bool per_out_ch);
void cutlass_fp4_group_mm(
torch::Tensor& output, const torch::Tensor& a, const torch::Tensor& b,
@@ -251,6 +252,14 @@ void get_cutlass_moe_mm_data(
const int64_t num_experts, const int64_t n, const int64_t k,
const std::optional<torch::Tensor>& blockscale_offsets);
void get_cutlass_pplx_moe_mm_data(torch::Tensor& expert_offsets,
torch::Tensor& problem_sizes1,
torch::Tensor& problem_sizes2,
const torch::Tensor& expert_num_tokens,
const int64_t num_local_experts,
const int64_t padded_m, const int64_t n,
const int64_t k);
void cutlass_scaled_mm_azp(torch::Tensor& out, torch::Tensor const& a,
torch::Tensor const& b,
torch::Tensor const& a_scales,