wehub-resource-sync
125 lines
5.3 KiB
C++
125 lines
5.3 KiB
C++
// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
#include "paddle/phi/backends/xpu/enforce_xpu.h"
|
|
#include "paddle/phi/backends/xpu/xpu_context.h"
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
#include "paddle/utils/optional.h"
|
|
|
|
namespace phi {
|
|
#ifndef MAX_NUM_EXPERTS
|
|
#define MAX_NUM_EXPERTS 64
|
|
#endif
|
|
|
|
template <typename T, typename Context>
|
|
void dispatch_tokens_zip(const Context &dev_ctx,
|
|
const DenseTensor &unzipped_tokens,
|
|
const DenseTensor &zipped_expertwise_rowmap,
|
|
const DenseTensor &expert_routemap_topk,
|
|
const DenseTensor &unzipped_token_probs,
|
|
DenseTensor *zipped_tokens,
|
|
DenseTensor *zipped_probs_topk,
|
|
const int total_zipped_tokens_num,
|
|
const int num_experts,
|
|
const int token_length,
|
|
const int topk,
|
|
const bool MP) {
|
|
using XPU_BF16 = typename XPUTypeTrait<phi::bfloat16>::Type;
|
|
// Map data types to C++ types
|
|
if (unzipped_token_probs.dtype() == DataType::FLOAT32) {
|
|
int r = xpu::moe_unpermute(
|
|
dev_ctx.x_context(),
|
|
reinterpret_cast<const XPU_BF16 *>(
|
|
unzipped_tokens.data<phi::bfloat16>()),
|
|
reinterpret_cast<const int *>(zipped_expertwise_rowmap.data<int>()),
|
|
reinterpret_cast<const int *>(expert_routemap_topk.data<int>()),
|
|
reinterpret_cast<const float *>(unzipped_token_probs.data<float>()),
|
|
reinterpret_cast<XPU_BF16 *>(zipped_tokens->data<phi::bfloat16>()),
|
|
zipped_probs_topk->data<float>(),
|
|
total_zipped_tokens_num,
|
|
num_experts,
|
|
token_length,
|
|
topk,
|
|
MP,
|
|
unzipped_tokens.dims()[0]);
|
|
PADDLE_ENFORCE_XDNN_SUCCESS(r, "moe_unpermute");
|
|
}
|
|
}
|
|
|
|
template <typename T, typename Context>
|
|
void MoeUnpermuteKernel(const Context &dev_ctx,
|
|
const DenseTensor &unzipped_tokens,
|
|
const DenseTensor &zipped_expertwise_rowmap,
|
|
const DenseTensor &expert_routemap_topk,
|
|
const DenseTensor &unzipped_token_probs,
|
|
const int total_zipped_tokens_num,
|
|
const int num_experts,
|
|
const bool MP,
|
|
const bool using_weighted_combine,
|
|
DenseTensor *zipped_tokens,
|
|
DenseTensor *zipped_probs_topk) {
|
|
PADDLE_ENFORCE_EQ(
|
|
using_weighted_combine,
|
|
false,
|
|
common::errors::Unimplemented("moe_unpermute on XPU does not support "
|
|
"using_weighted_combine=true yet."));
|
|
const int64_t cols = unzipped_tokens.dims()[1];
|
|
PADDLE_ENFORCE_LE(cols,
|
|
std::numeric_limits<int32_t>::max(),
|
|
common::errors::InvalidArgument(
|
|
"unzipped_tokens.dims()[1] should be less than "
|
|
"INT_MAX, received unzipped_tokens.dims()[1]: (%ld)",
|
|
cols));
|
|
PADDLE_ENFORCE_LE(
|
|
num_experts,
|
|
MAX_NUM_EXPERTS,
|
|
common::errors::InvalidArgument(
|
|
"Currently we support no more than (%ld), received num_expert: "
|
|
"(%ld). Please check input "
|
|
"value.",
|
|
MAX_NUM_EXPERTS,
|
|
num_experts));
|
|
const int64_t topk = expert_routemap_topk.dims()[1];
|
|
PADDLE_ENFORCE_LE(
|
|
topk,
|
|
std::numeric_limits<int32_t>::max(),
|
|
common::errors::InvalidArgument(
|
|
"topk should be less than INT_MAX, received topk: (%ld)", topk));
|
|
dev_ctx.template Alloc<T>(zipped_tokens);
|
|
dev_ctx.template Alloc<float>(zipped_probs_topk);
|
|
if (unzipped_tokens.numel() == 0) return; // 0-size tensor
|
|
void *zipped_probs_topk_ptr =
|
|
reinterpret_cast<void *>(zipped_probs_topk->data<float>());
|
|
PADDLE_ENFORCE_XPU_SUCCESS(
|
|
cudaMemsetAsync(zipped_probs_topk_ptr,
|
|
0,
|
|
sizeof(float) * int64_t(total_zipped_tokens_num) * topk,
|
|
reinterpret_cast<cudaStream_t>(dev_ctx.stream())));
|
|
dispatch_tokens_zip<T, Context>(dev_ctx,
|
|
unzipped_tokens,
|
|
zipped_expertwise_rowmap,
|
|
expert_routemap_topk,
|
|
unzipped_token_probs,
|
|
zipped_tokens,
|
|
zipped_probs_topk,
|
|
total_zipped_tokens_num,
|
|
num_experts,
|
|
static_cast<int>(cols),
|
|
static_cast<int>(topk),
|
|
MP);
|
|
}
|
|
} // namespace phi
|
|
|
|
PD_REGISTER_KERNEL(
|
|
moe_unpermute, XPU, ALL_LAYOUT, phi::MoeUnpermuteKernel, phi::bfloat16) {}
|