Skip to content

add num_worst_token #59

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

add num_worst_token #59

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions include/mori/ops/dispatch_combine/dispatch_combine.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,6 +79,7 @@ struct EpDispatchCombineConfig {
int maxNumInpTokenPerRank{128};
int numExpertPerRank{1};
int numExpertPerToken{2};
int numWorstToken{0};
int warpNumPerBlock{1};
int blockNum{1};
// If true, use external buffer which incurs extra copy overhead; otherwise, the kernel assumes
Expand All @@ -98,6 +99,9 @@ struct EpDispatchCombineConfig {
}

inline __host__ __device__ int MaxNumTokensToRecv() const {
if (numWorstToken != 0) {
return numWorstToken;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if return worldSize * numWorstToken will be better?

Copy link
Contributor

@isytwu isytwu Sep 4, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Perhaps using min(numWorstToken, worldSize * MaxNumTokensToRecvPerRank()) could prevent users from passing large values. And should MaxNumTokensToSend() also be changed to add numWorstToken?

}
return worldSize * MaxNumTokensToRecvPerRank();
}
};
Expand Down
2 changes: 2 additions & 0 deletions python/mori/ops/dispatch_combine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,6 +45,7 @@ class EpDispatchCombineConfig:
num_experts_per_token: int
warp_num_per_block: int = 8
block_num: int = 80
num_worst_token: int = 0
use_external_inp_buf: bool = True
kernel_type: EpDispatchCombineKernelType = EpDispatchCombineKernelType.IntraNode

Expand All @@ -71,6 +72,7 @@ def __init__(self, config):
num_experts_per_token=config.num_experts_per_token,
warp_num_per_block=config.warp_num_per_block,
block_num=config.block_num,
num_worst_token=config.num_worst_token,
use_external_inp_buf=config.use_external_inp_buf,
)
)
Expand Down
2 changes: 2 additions & 0 deletions src/ops/dispatch_combine/dispatch_combine.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,10 +66,12 @@ mori::application::SymmMemObjPtr ShmemMallocAndReturnMemObjPtr(size_t size, unsi
void EpDispatchCombineHandle::InitializeShmemBuf() {
size_t maxTokenSize = static_cast<ssize_t>(config.MaxNumTokensToRecv()) * config.hiddenDim *
config.maxTokenTypeSize;

size_t maxStagingTokSize = static_cast<ssize_t>(config.MaxNumTokensToRecv()) *
(config.hiddenDim * config.maxTokenTypeSize +
(sizeof(float) + sizeof(index_t)) * config.numExpertPerToken +
config.scaleDim * config.scaleTypeSize);

shmemInpTokMemObj = ShmemMallocAndReturnMemObjPtr(maxStagingTokSize, hipDeviceMallocUncached);
shmemOutTokMemObj = ShmemMallocAndReturnMemObjPtr(maxTokenSize, hipDeviceMallocUncached);
shmemStagingTokMemObj = ShmemMallocAndReturnMemObjPtr(maxStagingTokSize, hipDeviceMallocUncached);
Expand Down
7 changes: 4 additions & 3 deletions src/pybind/mori.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -234,13 +234,13 @@ void RegisterMoriOps(py::module_& m) {
.export_values();

pybind11::class_<mori::moe::EpDispatchCombineConfig>(m, "EpDispatchCombineConfig")
.def(pybind11::init<int, int, int, int, int, int, int, int, int, int, int, bool>(),
.def(pybind11::init<int, int, int, int, int, int, int, int, int, int, int, int, bool>(),
py::arg("rank") = 0, py::arg("world_size") = 0, py::arg("hidden_dim") = 0,
py::arg("scale_dim") = 0, py::arg("scale_type_size") = 0,
py::arg("max_token_type_size") = 0, py::arg("max_num_inp_token_per_rank") = 0,
py::arg("num_experts_per_rank") = 0, py::arg("num_experts_per_token") = 0,
py::arg("warp_num_per_block") = 0, py::arg("block_num") = 0,
py::arg("use_external_inp_buf") = true)
py::arg("num_worst_token") = 0, py::arg("warp_num_per_block") = 0,
py::arg("block_num") = 0, py::arg("use_external_inp_buf") = true)
.def_readwrite("rank", &mori::moe::EpDispatchCombineConfig::rank)
.def_readwrite("world_size", &mori::moe::EpDispatchCombineConfig::worldSize)
.def_readwrite("hidden_dim", &mori::moe::EpDispatchCombineConfig::hiddenDim)
Expand All @@ -252,6 +252,7 @@ void RegisterMoriOps(py::module_& m) {
.def_readwrite("num_experts_per_rank", &mori::moe::EpDispatchCombineConfig::numExpertPerRank)
.def_readwrite("num_experts_per_token",
&mori::moe::EpDispatchCombineConfig::numExpertPerToken)
.def_readwrite("num_worst_token", &mori::moe::EpDispatchCombineConfig::numWorstToken)
.def_readwrite("warp_num_per_block", &mori::moe::EpDispatchCombineConfig::warpNumPerBlock)
.def_readwrite("block_num", &mori::moe::EpDispatchCombineConfig::blockNum);

Expand Down
2 changes: 2 additions & 0 deletions tests/python/ops/bench_dispatch_combine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -189,6 +189,7 @@ def _bench_dispatch_combine(
num_experts_per_rank=16,
num_experts_per_token=8,
):
num_worst_token = max_num_inp_token_per_rank * world_size
config = mori.ops.EpDispatchCombineConfig(
data_type=data_type,
rank=rank,
Expand All @@ -200,6 +201,7 @@ def _bench_dispatch_combine(
max_num_inp_token_per_rank=max_num_inp_token_per_rank,
num_experts_per_rank=num_experts_per_rank,
num_experts_per_token=num_experts_per_token,
num_worst_token=num_worst_token,
warp_num_per_block=16,
block_num=80,
use_external_inp_buf=False,
Expand Down