This is an experimental branch that might lead to some faster performance when you expect long significand (more than 19 digits) #27
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
lemire
changed the title
This was referenced Jan 5, 2021
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Outdated suggestions cannot be applied.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.
If you have binary floating-point numbers, you ought to serialize them using no more than 17 digits. Anymore and you are being wasteful. Our code is optimized for such a scenario. However, if someone has data with lots and lots of digits per floating-point number, then our performance is somewhat less. This PR is better if you expect to be processing long significand.
Configuration: AMD Rome Processor (GCC 10)
The slowdown is caused by lower IPC (instruction per cycle) by up to 10% coupled with an increased in the number of instructions (about 5%). It does not cause more branch misses. Depending on the benchmark, between 10 to 20 new instructions per float are needed with this branch in the general case.
See #51 for a different trade-off.