test: Add f64.{add,sub} off by one unit tests #786
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
Codecov Report
@@ Coverage Diff @@
## master #786 +/- ##
==========================================
+ Coverage 99.22% 99.23% +0.01%
==========================================
Files 79 79
Lines 12060 12394 +334
==========================================
+ Hits 11966 12299 +333
- Misses 94 95 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
|
gumb0
reviewed
Apr 20, 2021
| // The i386 x87 FPU performs all arithmetic with 80-bit extended precision by default | ||
| // and in the end rounds it to the required type. This causes issues for f64 types | ||
| // because results may be different than when doing computation with 64-bit precision. | ||
| // f32 is not affected. |
There was a problem hiding this comment.
Some sources says that float has too low precision to observe different results. This is confirmed by testfloat tests.
gumb0
reviewed
Apr 20, 2021
| auto instance = instantiate(parse(wasm)); | ||
|
|
||
| constexpr auto a = 0x1.0008000008000p60; | ||
| constexpr auto b = 0x1.0000000081fffp07; |
There was a problem hiding this comment.
If you know the easy way to do this, maybe it would be helpful to comment with "de-normalized" exponent 60 value of this.
I mean, as I understand, it's something like this:
Suggested change
| constexpr auto b = 0x1.0000000081fffp07; | |
| constexpr auto a = 0x1.0008000008000p60; | |
| constexpr auto b = 0x1.0000000081fffp07; // 0x0.00000000000008...p60; |
There was a problem hiding this comment.
I don't really know what exactly is going on here. Maybe if we convert this also to long double we may see what's going on.
gumb0
approved these changes
Apr 20, 2021
The i386 x87 FPU performs all arithmetic with 80-bit extended precision by default and in the end rounds it to the required type. This causes issues for f64 types because results may be different than when doing computation with 64-bit precision. f32 is not affected. These tests presents examples of results which are different in such case.
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.
The i386 x87 FPU performs all arithmetic with 80-bit extended precision
by default and in the end rounds it to the required type. This causes
issues for f64 types because results may be different than when doing
computation with 64-bit precision. f32 is not affected.
These tests presents examples of results which are different in such
case.