Skip to content

Commit

Permalink
Merge pull request #294 from Tencent/jiaruifang/accurate_timer
Browse files Browse the repository at this point in the history
accurate timer
  • Loading branch information
feifeibear authored Dec 23, 2021
2 parents bc9683f + 1a74991 commit b1265ee
Show file tree
Hide file tree
Showing 3 changed files with 7 additions and 2 deletions.
5 changes: 4 additions & 1 deletion CHANGE_LOG.md
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
## v0.4.5 Dec. 2021
refactory the files in example and adding chunk size searching.
Evaluate on 8 nodes of SuperPod. Fix bugs in multi-GPU mem tracer.

## v0.4.5 Dec. 2021
Refactor the files in example and add chunk size searching.


### v0.4.4 Dec. 2021
Expand Down
2 changes: 2 additions & 0 deletions patrickstar/utils/global_timer.py
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,7 @@
# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

import time
import torch

# from .logging import logger
from .singleton_meta import SingletonMeta
Expand Down Expand Up @@ -57,6 +58,7 @@ def start_profile(self, key):
def finish_profile(self, key):
if not self.start_flag:
return
torch.cuda.current_stream().synchronize()
if key in self.elapse_stat:
self.elapse_stat[key] += time.time() - self.start_time[key]
else:
Expand Down
2 changes: 1 addition & 1 deletion setup.py
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,7 @@ def fetch_requirements(path):

setup(
name="patrickstar",
version="0.4.5",
version="0.4.6",
description="PatrickStart library",
long_description="PatrickStar: Parallel Training of Large Language Models via a Chunk-based Parameter Server",
long_description_content_type="text/markdown",
Expand Down

0 comments on commit b1265ee

Please sign in to comment.