Baunsgaard commented on PR #1843:
URL: https://github.com/apache/systemds/pull/1843#issuecomment-1723219006
In the results you have above the time of the `Heavy hitter instructions:`
is what i would use to compare. The rest of the time reported is all overhead
of Java , I/O and other compilation related elements.
But i would modify the script you use. A example of a measuring script:
```R
res = read($1)
print(sum(res))
for(i in 1:$2) {
res2 = OPERATION(res)
}
print(sum(res2))
```
This reads the file in the first argument and then repeats the operation the
number of times specified in the second argument. Because of the way we execute
in SystemDS the sum before the loop make sure res is actually read from disk.
--
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: dev-unsubscr...@systemds.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
