More precision for fastCos and fastSin #264
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
fastCos and fastSin had a max error of ~0.2 on [-pi pi]. The updated version is ~0.000007.
| @@ -67,6 +73,12 @@ namespace glm | |||
| } | |||
|
|
|||
| template <typename genType> | |||
| GLM_FUNC_QUALIFIER genType three_half_pi() | |||
There was a problem hiding this comment.
This name is confusing. I had to make calculation to understand what it meant.
"three_over_two_pi" follows better GLM precedent.
|
It looks good. It could be valuable to had some performance tests in a similar fashion than https://github.com/g-truc/glm/blob/vectorize/test/gtc/gtc_bitfield.cpp. But really, the fast trigonometry is pretty broken right now so regardless, it's a very valuable addition. Thanks, |
Groovounet
added a commit
that referenced
this pull request
Nov 14, 2014
More precision for fastCos and fastSin #264
|
I didn't realized that your pull request was against 0.9.5. I'll manually pull it for 0.9.6. Thanks, |
Groovounet
pushed a commit
that referenced
this pull request
Nov 14, 2014
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I tested the precision of glm::fastCos and glm::fastSin against glm::cos and glm::sin on the range [-pi pi]. The max error is ~0.2 on my computer.
I've made some changes and now it's ~0.000007. It's also slightly faster and work on a wider range.
To give some context, I had a problem of cross platform float determinism whith glm::cos and glm::sin, because theses methods rely on .
Using the fastCos and fastSin solve my determinism problem, but the accuracy was not good enough anymore.
If you are ok with the code, I can improve the rest of the fast_trigonometry.
For reference, I used the same method as the one found in "A Guide to Approximations" by Jack G. Ganssle