45 lines
2.1 KiB
Plaintext
45 lines
2.1 KiB
Plaintext
#include <flashinfer/attention/decode.cuh>
|
|
#include <flashinfer/attention/scheduler.cuh>
|
|
#include <optional>
|
|
|
|
#include "mla_config.inc"
|
|
#include "pytorch_conversion_utils.h"
|
|
#include "pytorch_extension_utils.h"
|
|
|
|
using namespace flashinfer;
|
|
|
|
at::Tensor BatchDecodeWithPagedKVCachePlanMLA(at::Tensor float_workspace_buffer,
|
|
at::Tensor int_workspace_buffer,
|
|
at::Tensor page_locked_int_workspace_buffer,
|
|
at::Tensor indptr, int64_t batch_size,
|
|
int64_t num_qo_heads, int64_t page_size,
|
|
bool enable_cuda_graph) {
|
|
const c10::cuda::OptionalCUDAGuard device_guard(float_workspace_buffer.device());
|
|
const cudaStream_t stream = c10::cuda::getCurrentCUDAStream();
|
|
|
|
size_t float_workspace_size_in_bytes =
|
|
float_workspace_buffer.size(0) * float_workspace_buffer.element_size();
|
|
size_t int_workspace_size_in_bytes =
|
|
int_workspace_buffer.size(0) * int_workspace_buffer.element_size();
|
|
|
|
DecodePlanInfo plan_info;
|
|
cudaStream_t stream = reinterpret_cast<cudaStream_t>(cuda_stream);
|
|
|
|
auto work_estimation_func =
|
|
BatchDecodeWithPagedKVCacheWorkEstimationDispatchedMLA<HEAD_DIM_CKV, HEAD_DIM_KPE,
|
|
AttentionVariant, Params>;
|
|
cudaError_t status =
|
|
DecodePlan<HEAD_DIM_CKV, flashinfer::PosEncodingMode::kRoPELlama, AttentionVariant, Params>(
|
|
static_cast<void*>(float_workspace_buffer.data_ptr()), float_workspace_size_in_bytes,
|
|
static_cast<void*>(int_workspace_buffer.data_ptr()),
|
|
static_cast<void*>(page_locked_int_workspace_buffer.data_ptr()),
|
|
int_workspace_size_in_bytes, plan_info, static_cast<IdType*>(indptr.data_ptr()),
|
|
batch_size, num_qo_heads, page_size, enable_cuda_graph, /*stream=*/stream,
|
|
work_estimation_func);
|
|
|
|
TORCH_CHECK(status == cudaSuccess, "BatchDecodeWithPagedKVCachePlanMLA failed with error ",
|
|
cudaGetErrorString(status));
|
|
|
|
return vec_to_tensor(plan_info.ToVector());
|
|
}
|