-
-
Notifications
You must be signed in to change notification settings - Fork 860
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
JpegEncoder - Optimize Some Low Hanging Fruit #1533
Conversation
This is all working on my SB2. I'll have to crack open the mac to see what's happening here. 😢 |
Definitely a Jit issue. Removing inlining changed the error. I’ll manually inline everything into the method. |
Codecov Report
@@ Coverage Diff @@
## master #1533 +/- ##
==========================================
- Coverage 83.52% 83.47% -0.05%
==========================================
Files 742 742
Lines 32802 32830 +28
Branches 3671 3667 -4
==========================================
+ Hits 27399 27406 +7
- Misses 4689 4709 +20
- Partials 714 715 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
Everything works now. Will try and recreate the JIT issue and open an issue upstream |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM with one suggestion.
Is this code just for dividing your 8x8 blocks by the quantization table? If so, why not calculate the reciprocal of the QT values up front so you can multiply instead? Also, you can probably use |
I'll be honest @saucecontrol and say it's been a while since I looked at all this so I've forgotten a lot of how jpeg works. I'm just limiting changes to known bottlenecks (that part was actually the least of the three). Regarding |
Ha, I've just started re-learning everything I've ever semi-understood about JPEG. Didn't expect the quantization to be much of a bottleneck, but it seems you got some good gains from widening the vectors, so it may be worth looking into some other time. |
Man... I cannot wait to see what you bring to the table. |
Getting so close now...
|
JpegEncoder - Optimize Some Low Hanging Fruit
Prerequisites
Description
I had a go at optimizing some of the methods identified by @tkp1n here
#1476 (comment)
BEFORE
AFTER