Use threads #4

New Issue

tastytea · 2021-05-25T19:34:48+02:00

tastytea commented

2021-05-25 19:34:48 +02:00

We should probably make use of threads where possible. It should be safe to use a separate thread for each file in the EPUB. Could be too much overhead though. Maybe it is better to use one thread per EPUB.

tastytea added the

enhancement

label 2021-05-25 19:34:48 +02:00

tastytea commented

2021-05-26 15:18:27 +02:00

I tested a bit with time ./build/src/epubgrep -EC2 'a' <files> > /dev/null. I made 2 passes each time and took the average.

3 files:
- Without threads: 10.7 seconds
- With 1 thread per file: 7.2 seconds
6 files:
- Without threads: 42.2 seconds
- With 1 thread per file: 29.1 seconds

That's 31-33% improvement.

With 6 books and time ./build/src/epubgrep -EiC2 '(aber|doch)' <files> > /dev/null, 4 passes:

Without threads: 0.8 seconds
With 1 thread per file: 0.6 seconds

25% improvement.

I tested a bit with `time ./build/src/epubgrep -EC2 'a' <files> > /dev/null`. I made 2 passes each time and took the average. * 3 files: * Without threads: 10.7 seconds * With 1 thread per file: 7.2 seconds * 6 files: * Without threads: 42.2 seconds * With 1 thread per file: 29.1 seconds That's 31-33% improvement. With 6 books and `time ./build/src/epubgrep -EiC2 '(aber|doch)' <files> > /dev/null`, 4 passes: * Without threads: 0.8 seconds * With 1 thread per file: 0.6 seconds 25% improvement.

tastytea referenced this issue from a commit

2021-05-26 18:05:44 +02:00

Use threads if more than one input file is searched.

tastytea closed this issue

2021-05-26 18:05:44 +02:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: tastytea/epubgrep#4