This repository was archived by the owner on Apr 28, 2025. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 102
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
07973f0
to
a8069e5
Compare
Additionally, make use of this version to implement `floor` and `floorf`. Similar to `ceil`, musl'f `ceilf` routine seems to work better for all float widths than the `ceil` algorithm. Trying with the `ceil` (`f64`) algorithm produced the following regressions: icount::icount_bench_floor_group::icount_bench_floor logspace:setup_floor() Performance has regressed: Instructions (14064 > 13171) regressed by +6.78005% (>+5.00000) Baselines: softfloat|softfloat Instructions: 14064|13171 (+6.78005%) [+1.06780x] L1 Hits: 16821|15802 (+6.44855%) [+1.06449x] L2 Hits: 0|0 (No change) RAM Hits: 8|9 (-11.1111%) [-1.12500x] Total read+write: 16829|15811 (+6.43856%) [+1.06439x] Estimated Cycles: 17101|16117 (+6.10535%) [+1.06105x] icount::icount_bench_floorf128_group::icount_bench_floorf128 logspace:setup_floorf128() Baselines: softfloat|softfloat Instructions: 166868|N/A (*********) L1 Hits: 221429|N/A (*********) L2 Hits: 1|N/A (*********) RAM Hits: 34|N/A (*********) Total read+write: 221464|N/A (*********) Estimated Cycles: 222624|N/A (*********) icount::icount_bench_floorf16_group::icount_bench_floorf16 logspace:setup_floorf16() Baselines: softfloat|softfloat Instructions: 143029|N/A (*********) L1 Hits: 176517|N/A (*********) L2 Hits: 1|N/A (*********) RAM Hits: 13|N/A (*********) Total read+write: 176531|N/A (*********) Estimated Cycles: 176977|N/A (*********) icount::icount_bench_floorf_group::icount_bench_floorf logspace:setup_floorf() Performance has regressed: Instructions (14732 > 10441) regressed by +41.0976% (>+5.00000) Baselines: softfloat|softfloat Instructions: 14732|10441 (+41.0976%) [+1.41098x] L1 Hits: 17616|13027 (+35.2268%) [+1.35227x] L2 Hits: 0|0 (No change) RAM Hits: 8|6 (+33.3333%) [+1.33333x] Total read+write: 17624|13033 (+35.2260%) [+1.35226x] Estimated Cycles: 17896|13237 (+35.1968%) [+1.35197x]
Use the generic algorithms to provide implementations for these routines.
447ad60
to
2bc39ff
Compare
icount shows improvements:
|
2bc39ff
to
31f0802
Compare
tgross35
added a commit
that referenced
this pull request
Apr 18, 2025
Add `floorf16` and `floorf128`
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Add a generic version of
floor
. Additionally, make use of this version to implementfloor
andfloorf
.Similar to
ceil
, musl'fceilf
routine seems to work better for all float widths than theceil
algorithm. Trying with theceil
(f64
) algorithm produced the following regressions:Add
floorf16
andfloorf128
Use the generic algorithms to provide implementations for these routines.