Benchmark {Float,Double,Float80}.description #15234
Merged
Conversation
This file contains hidden or 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
This just tests how fast we can format the standard floating-point types. It also includes a test that uses the result, to ensure that future optimized _creation_ of the string doesn't incidentally pessimize _use_ of the results.
This just tests how fast we can format the standard floating-point types. It also includes a test that uses the result, to ensure that future optimized _creation_ of the string doesn't incidentally pessimize _use_ of the results.
|
In particular, I'd appreciate a check whether the |
|
@swift-ci Please smoke benchmark |
milseman
approved these changes
Mar 14, 2018
| for i in 0..<count { | ||
| let raw = UInt32(i) * step | ||
| let f = Float(bitPattern: raw) | ||
| s = f.description |
There was a problem hiding this comment.
You can write this as: blackHole(f.description)
Blackhole is from TestUtils.swift, which has a few no-op functions that the optimizer cannot see into (separate module). This way you can avoid trying to trick the optimizer and have trivial uses of values.
There was a problem hiding this comment.
Thanks for pointing that out. That's much cleaner.
| for i in 0..<count { | ||
| let raw = UInt32(i) * step | ||
| let f = Float(bitPattern: raw) | ||
| s = "result was \(f)" |
There was a problem hiding this comment.
Is your intention to avoid having the final string be small? If so, blackHole("and the actual result was \(f)") is more future proof, as not even a 5-bit encoding can store this.
| for _ in 0..<N { | ||
| for i in 0..<count { | ||
| let fraction = UInt64(i) * step | ||
| let exponent = UInt(i) % 32768 |
There was a problem hiding this comment.
CC @stephentyrone as I don't know how to benchmark floats (or the magic numbers).
Build comment file:Optimized (O)Regression (15)
Improvement (26)
No Changes (346)
Added (6)
Unoptimized (Onone)Regression (9)
Improvement (14)
No Changes (364)
Added (6)
Hardware Overview |
stephentyrone
requested changes
Mar 14, 2018
|
@stephentyrone - Did you want to test values with short text forms or values close to 1? Generally speaking, manually-typed values in code tend to be both. The easiest way to generate such numbers would be something like: |
|
Values with modest exponents is what I really have in mind; it's common for input values to have short-forms, but not especially common for results that get printed, so I'm less concerned about that. I'd be OK with even your very simple example for now, as long as we have some coverage for this case. |
…swift into tbkka-float-description-benchmark
|
@swift-ci Please smoke benchmark |
|
@stephentyrone - I now have |
…swift into tbkka-float-description-benchmark
Build comment file:Build failed before running benchmark. |
|
@swift-ci Please smoke benchmark |
|
@swift-ci please smoke test |
Build comment file:Build failed before running benchmark. |
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.
This just tests how fast we can produce
.descriptionfor the standard floating-point types. It also includes a benchmark that actually does something with the result, to check whether future optimizations to the creation of the string accidentally pessimize use of the results.