marsupialtail commented on PR #13640:
URL: https://github.com/apache/arrow/pull/13640#issuecomment-1227765779
Really hope can pass CI now, this is a script you can use to test
performance:
```
SIZE = 1024 * 1024
N = 100 * 32
import time
import os
import pyarrow.fs as fs
#da = fs.LocalFileSystem(use_directio=True)
da = fs.LocalFileSystem(use_directio=True)
#da = fs.LocalFileSystem(reuse=False)
s = da.open_output_stream("bump")
a = bytes("1","utf-8") * SIZE
start = time.time()
for i in range(N):
s.write(a)
s.close()
os.sync()
print(time.time() - start)
```
On my NVME SSD direct IO is about 50% faster than non-direct IO writing this
file.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
