g2gps commented on PR #10478:
URL: https://github.com/apache/nuttx/pull/10478#issuecomment-1705805031

   > @g2gps any idea why for 64KB block the write performance with copy it 
worst? Is it because maybe cache misses?
   
   It's because the amount of data copied is larger than the total cache size 
I've allocated, so the cached block driver needs to start unloading dirty 
blocks to the underlying EMMC. If I repeat the test with a smaller file, the 
64KB block size is faster.
   
   ```
   nsh> dd if=/dev/zero of=test3 bs=65536 count=16
   1048576 bytes copied, 65856 usec, 15549 KB/s
   nsh> dd if=test3 of=test4 bs=65536
   1048576 bytes copied, 313765 usec, 3263 KB/s
   ```
   
   In general, the performance of a read/write/copy operation is going to 
depend on if the data is already cached, and when it is not, if any existing 
cached data has been marked dirty and needs to be synchronised. 


-- 
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]

Reply via email to