Skip to content
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

Draft: Use fused push_send_recv kernel for TP AG and RS overlaps #1200

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Draft: Use fused push_send_recv kernel for TP AG and RS overlaps #1200

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8367bad
Used fused push_send_recv kernel for AG overlap
erhoo82 Sep 21, 2024
5cad985
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Sep 26, 2024
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
113 changes: 77 additions & 36 deletions transformer_engine/pytorch/csrc/comm_gemm_overlap.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -666,6 +666,7 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
int _num_comm_sm;
int _cga_size;
bool _atomic_gemm;
bool _use_fused_sendrecv;

UbufP2PCommOverlap(torch::Tensor sample, int myrank, int numranks, int mylocal, int numlocal,
int mynode, int numnodes, int tp_size, int num_comm_sm, int comm_cga_size,
Expand Down Expand Up @@ -727,6 +728,10 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
at::cuda::getStreamFromExternal(stream, stream_main.device_index()));
}

// Set to use a single kernel for push_send and push_recv
const char *use_fused_sendrecv = std::getenv("NVTE_UB_FUSED_SEND_RECV");
_use_fused_sendrecv = use_fused_sendrecv != nullptr && use_fused_sendrecv[0] == '1';

// Set the number of SMs for GEMM with margin
cudaDeviceProp prop;
cudaGetDeviceProperties(&prop, 0);
Expand Down Expand Up @@ -943,9 +948,10 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {

at::cuda::CUDAStream stream_main = at::cuda::getCurrentCUDAStream();
NVTE_CHECK_CUDA(cudaEventRecord(_start_compute, (cudaStream_t)stream_main));

NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _start_compute, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_recv, _start_compute, 0));
if (!_use_fused_sendrecv) {
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_recv, _start_compute, 0));
}
for (size_t i = 0; i < _stream_compute.size(); i++) {
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_compute[i], _start_compute, 0));
}
Expand All @@ -959,13 +965,20 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
int send_offset = comm_bytes * send_chunk_id;
int recv_offset = comm_bytes * recv_chunk_id;
int peer_rank = (_tp_id % 2 == 0) ? _next_rank : _prev_rank;
userbuffers_send(_ub_reg, send_offset, _ub_reg, send_offset, comm_bytes, _ub_comm, peer_rank,
(cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, recv_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm, peer_rank,
(cudaStream_t)_stream_recv);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _stop_recv, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_compute[0], _stop_recv, 0));
if (_use_fused_sendrecv) {
userbuffers_sendrecv(_ub_reg, _ub_reg, send_offset, send_offset, comm_bytes, _ub_comm,
peer_rank, peer_rank, (cudaStream_t)_stream_send);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_send));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_compute[0], _stop_recv, 0));
} else {
userbuffers_send(_ub_reg, send_offset, _ub_reg, send_offset, comm_bytes, _ub_comm,
peer_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, recv_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
peer_rank, (cudaStream_t)_stream_recv);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _stop_recv, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_compute[0], _stop_recv, 0));
}

int local_rank_round2 = (_tp_id % 2 == 0) ? _tp_id : _tp_id - 1;
const int next_rank = (_tp_size + _tp_id + 2) % _tp_size + _rank_round_tp;
Expand Down Expand Up @@ -998,14 +1011,22 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {

if (i < num_steps - 1) {
// P2P communication
userbuffers_send(_ub_reg, send_offset, _ub_reg, send_offset, comm_bytes * 2, _ub_comm,
next_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, recv_offset, _ub_reg, recv_offset, comm_bytes * 2, _ub_comm,
prev_rank, (cudaStream_t)_stream_recv);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _stop_recv, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent(
(cudaStream_t)_stream_compute[(i + 1) % _stream_compute.size()], _stop_recv, 0));
if (_use_fused_sendrecv) {
userbuffers_sendrecv(_ub_reg, _ub_reg, send_offset, send_offset, comm_bytes * 2,
_ub_comm, next_rank, prev_rank, (cudaStream_t)_stream_send);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_send));
NVTE_CHECK_CUDA(cudaStreamWaitEvent(
(cudaStream_t)_stream_compute[(i + 1) % _stream_compute.size()], _stop_recv, 0));
} else {
userbuffers_send(_ub_reg, send_offset, _ub_reg, send_offset, comm_bytes * 2, _ub_comm,
next_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, recv_offset, _ub_reg, recv_offset, comm_bytes * 2, _ub_comm,
prev_rank, (cudaStream_t)_stream_recv);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _stop_recv, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent(
(cudaStream_t)_stream_compute[(i + 1) % _stream_compute.size()], _stop_recv, 0));
}
} else if (B_copy.numel() > 0) {
assert(B_copy.numel() == _ubufs[_tp_id].numel());
assert(B_copy.element_size() == _ubufs[_tp_id].element_size());
Expand Down Expand Up @@ -1043,14 +1064,22 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {

if (i < _tp_size - 1) {
// P2P communication
userbuffers_send(_ub_reg, send_offset, _ub_reg, send_offset, comm_bytes, _ub_comm,
_next_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, recv_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
_prev_rank, (cudaStream_t)_stream_recv);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _stop_recv, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent(
(cudaStream_t)_stream_compute[(i + 1) % _stream_compute.size()], _stop_recv, 0));
if (_use_fused_sendrecv) {
userbuffers_sendrecv(_ub_reg, _ub_reg, send_offset, send_offset, comm_bytes, _ub_comm,
_next_rank, _prev_rank, (cudaStream_t)_stream_send);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_send));
NVTE_CHECK_CUDA(cudaStreamWaitEvent(
(cudaStream_t)_stream_compute[(i + 1) % _stream_compute.size()], _stop_recv, 0));
} else {
userbuffers_send(_ub_reg, send_offset, _ub_reg, send_offset, comm_bytes, _ub_comm,
_next_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, recv_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
_prev_rank, (cudaStream_t)_stream_recv);
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _stop_recv, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent(
(cudaStream_t)_stream_compute[(i + 1) % _stream_compute.size()], _stop_recv, 0));
}
} else if (B_copy.numel() > 0) {
assert(B_copy.numel() == _ubufs[_tp_id].numel());
assert(B_copy.element_size() == _ubufs[_tp_id].element_size());
Expand All @@ -1066,8 +1095,10 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
}
NVTE_CHECK_CUDA(cudaEventRecord(_stop_send, (cudaStream_t)_stream_send));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)stream_main, _stop_send, 0));
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)stream_main, _stop_recv, 0));
if (!_use_fused_sendrecv) {
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)stream_main, _stop_recv, 0));
}
at::cuda::setCurrentCUDAStream(stream_main);
_ub_comm->sms = ori_sms;

Expand Down Expand Up @@ -1190,7 +1221,9 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
at::cuda::CUDAStream stream_main = at::cuda::getCurrentCUDAStream();
NVTE_CHECK_CUDA(cudaEventRecord(_start_compute, (cudaStream_t)stream_main));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _start_compute, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_recv, _start_compute, 0));
if (!_use_fused_sendrecv) {
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_recv, _start_compute, 0));
}
for (size_t i = 0; i < _stream_compute.size(); i++) {
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_compute[i], _start_compute, 0));
}
Expand Down Expand Up @@ -1218,12 +1251,18 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
int recv_rank = (_tp_size + _tp_id - i) % _tp_size + _rank_round_tp;
NVTE_CHECK_CUDA(cudaEventRecord(
_start_comm, (cudaStream_t)_stream_compute[(i - 1) % _stream_compute.size()]));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _start_comm, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_recv, _start_comm, 0));
userbuffers_send(_ub_reg, send_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
send_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, send_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
recv_rank, (cudaStream_t)_stream_recv);
if (_use_fused_sendrecv) {
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _start_comm, 0));
userbuffers_sendrecv(_ub_reg, _ub_reg, send_offset, recv_offset, comm_bytes, _ub_comm,
send_rank, recv_rank, (cudaStream_t)_stream_send);
} else {
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_send, _start_comm, 0));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)_stream_recv, _start_comm, 0));
userbuffers_send(_ub_reg, send_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
send_rank, (cudaStream_t)_stream_send);
userbuffers_recv(_ub_reg, send_offset, _ub_reg, recv_offset, comm_bytes, _ub_comm,
recv_rank, (cudaStream_t)_stream_recv);
}
}
}
at::cuda::setCurrentCUDAStream(stream_main);
Expand All @@ -1233,8 +1272,10 @@ struct UbufP2PCommOverlap : torch::CustomClassHolder, UbufBase {
}
NVTE_CHECK_CUDA(cudaEventRecord(_stop_send, (cudaStream_t)_stream_send));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)stream_main, _stop_send, 0));
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)stream_main, _stop_recv, 0));
if (!_use_fused_sendrecv) {
NVTE_CHECK_CUDA(cudaEventRecord(_stop_recv, (cudaStream_t)_stream_recv));
NVTE_CHECK_CUDA(cudaStreamWaitEvent((cudaStream_t)stream_main, _stop_recv, 0));
}

// Reduce GEMM output chunks
char *reduce_buf_ptr = reinterpret_cast<char *>(_ubufs[_tp_size - 1].data_ptr());
Expand Down
2 changes: 1 addition & 1 deletion transformer_engine/pytorch/csrc/userbuffers/userbuffers.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2277,7 +2277,7 @@ void userbuffers_sendrecv(const int srchandler, const int dsthandler, const size

void *send_srcptr = reinterpret_cast<char *>(comm->mem_ptr[srchandler]) + send_offset;
void *send_dstptr =
reinterpret_cast<char *>(comm->peer_ptr[dsthandler][send_peerlocal]) + send_offset;
reinterpret_cast<char *>(comm->peer_ptr[dsthandler][send_peerlocal]) + recv_offset;

if (comm->use_ce) {
// kuserbuffers_inc<<<1, 1, 0, stream>>>(reinterpret_cast<int *>(ce_send_start_ptr));
Expand Down