Skip to content

[TRTLLM-10126][feat] Increase topk upper limit to 22 for NVLinkOneSid… #10229

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
nv-guomingz wants to merge 1 commit into
base: main
Choose a base branch
from
+71 −2
Open

[TRTLLM-10126][feat] Increase topk upper limit to 22 for NVLinkOneSid… #10229

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
70 changes: 69 additions & 1 deletion cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,12 @@ namespace kernels::moe_comm
#define SWITCH_TOP_K(top_k, TOP_K, ...) \
switch (top_k) \
{ \
case 22: \
{ \
constexpr int TOP_K = 22; \
__VA_ARGS__; \
break; \
} \
case 16: \
{ \
constexpr int TOP_K = 16; \
Expand Down Expand Up @@ -654,7 +660,69 @@ __device__ void vectorized_combine_impl(
acc[k].load(recv_buffer + base_token + offset);
}
// Reduce acc[TOP_K] into acc[0]
if constexpr (TOP_K == 16)
if constexpr (TOP_K == 22)
{
T* a0 = reinterpret_cast<T*>(&acc[0]);
T* a1 = reinterpret_cast<T*>(&acc[1]);
T* a2 = reinterpret_cast<T*>(&acc[2]);
T* a3 = reinterpret_cast<T*>(&acc[3]);
T* a4 = reinterpret_cast<T*>(&acc[4]);
T* a5 = reinterpret_cast<T*>(&acc[5]);
T* a6 = reinterpret_cast<T*>(&acc[6]);
T* a7 = reinterpret_cast<T*>(&acc[7]);
T* a8 = reinterpret_cast<T*>(&acc[8]);
T* a9 = reinterpret_cast<T*>(&acc[9]);
T* a10 = reinterpret_cast<T*>(&acc[10]);
T* a11 = reinterpret_cast<T*>(&acc[11]);
T* a12 = reinterpret_cast<T*>(&acc[12]);
T* a13 = reinterpret_cast<T*>(&acc[13]);
T* a14 = reinterpret_cast<T*>(&acc[14]);
T* a15 = reinterpret_cast<T*>(&acc[15]);
T* a16 = reinterpret_cast<T*>(&acc[16]);
T* a17 = reinterpret_cast<T*>(&acc[17]);
T* a18 = reinterpret_cast<T*>(&acc[18]);
T* a19 = reinterpret_cast<T*>(&acc[19]);
T* a20 = reinterpret_cast<T*>(&acc[20]);
T* a21 = reinterpret_cast<T*>(&acc[21]);
#pragma unroll
for (int j = 0; j < elems_per_vec; ++j)
{
a0[j] += a1[j];
a2[j] += a3[j];
a4[j] += a5[j];
a6[j] += a7[j];
a8[j] += a9[j];
a10[j] += a11[j];
a12[j] += a13[j];
a14[j] += a15[j];
a16[j] += a17[j];
a18[j] += a19[j];
a20[j] += a21[j];
}
#pragma unroll
for (int j = 0; j < elems_per_vec; ++j)
{
a0[j] += a2[j];
a4[j] += a6[j];
a8[j] += a10[j];
a12[j] += a14[j];
a16[j] += a18[j];
}
#pragma unroll
for (int j = 0; j < elems_per_vec; ++j)
{
a0[j] += a4[j];
a8[j] += a12[j];
a16[j] += a20[j];
}
#pragma unroll
for (int j = 0; j < elems_per_vec; ++j)
{
a0[j] += a8[j];
a0[j] += a16[j];
}
}
else if constexpr (TOP_K == 16)
{
T* a0 = reinterpret_cast<T*>(&acc[0]);
T* a1 = reinterpret_cast<T*>(&acc[1]);
Expand Down
2 changes: 1 addition & 1 deletion cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ namespace kernels::moe_comm
{

// Configuration constants
static constexpr int kMaxTopK = 16; // Maximum top-k experts per token
static constexpr int kMaxTopK = 22; // Maximum top-k experts per token
static constexpr int kMaxPayloads = 4; // Maximum number of different payload types
static constexpr int kMaxRanks = 64; // Maximum supported EP size

Expand Down
1 change: 1 addition & 0 deletions tests/unittest/_torch/multi_gpu/test_moe_a2a.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -565,6 +565,7 @@ def test_dispatch(self, mpi_pool_executor, all_num_tokens, top_k):
(2, [100, 50], 2),
(4, [32, 32, 32, 32], 4),
(4, [32, 32, 32, 32], 10), # (top_k=10 is used by Qwen3-next)
(4, [32, 32, 32, 32], 22),
(4, [1, 1, 1, 1], 2),
(8, [640, 640, 640, 640, 640, 640, 640, 640], 4),
(4, [32, 0, 16, 0], 2),
Expand Down