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

Jump to bottom

UCT/CUDA_IPC: Increase MAX_PEERS and MAX_STREAMS #10258

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
+2 −2
Open

UCT/CUDA_IPC: Increase MAX_PEERS and MAX_STREAMS #10258

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5b4ddba
UCT/CUDA_IPC: increase max peers and max streams
Akshay-Venkatesh Oct 24, 2024
7d2815d
UCT/CUDA_IPC: make max_stream and max_peers consistent by default
Akshay-Venkatesh Oct 30, 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
2 changes: 1 addition & 1 deletion src/uct/cuda/cuda_ipc/cuda_ipc_iface.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ static ucs_config_field_t uct_cuda_ipc_iface_config_table[] = {
"Max number of event completions to pick during cuda events polling",
ucs_offsetof(uct_cuda_ipc_iface_config_t, max_poll), UCS_CONFIG_TYPE_UINT},

{"MAX_STREAMS", "16",
{"MAX_STREAMS", "128",
Copy link
Contributor

Choose a reason for hiding this comment

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

can we use UCS_PP_MAKE_STRING(UCT_CUDA_IPC_MAX_PEERS) here?
what happens if the user sets MAX_STREAMS config to a value larger than UCT_CUDA_IPC_MAX_PEERS?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Having larger MAX_STREAMS than UCT_CUDA_IPC_MAX_PEERS allows cuda-ipc transport to round-robin select among more number of streams. It wouldn't be a functional or performance issue in my opinion although it is unlikely to improve performance.

"Max number of CUDA streams to make concurrent progress on",
ucs_offsetof(uct_cuda_ipc_iface_config_t, max_streams), UCS_CONFIG_TYPE_UINT},

Expand Down
2 changes: 1 addition & 1 deletion src/uct/cuda/cuda_ipc/cuda_ipc_iface.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@
#include "cuda_ipc_cache.h"


#define UCT_CUDA_IPC_MAX_PEERS 16
#define UCT_CUDA_IPC_MAX_PEERS 128


typedef struct uct_cuda_ipc_iface {
Expand Down
Loading