Added -use for NONTEMPORAL #317
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
I added an option to change non-temporal memory access from the command line. The default is off even though on Radeon VII it is about a 0.5% gaain. I suspect an A100 or recent AMD consumer cards with large caches will see bigger gains without nontemporal access. We could change the default setting to depend on the cache size reported by clinfo. In either case, we should make -tune auto-config this option.
|
Thanks! we should make it ON by default (but guard it with LLVM builtin checks: __has_builtin() ) |
|
FYI, someone on the forum tried it on Intel Battlemage. Nontemporal is 4x
slower on Battlemage.
I see about 5us benefit on Radeon VII.
I don't really care which is the default. Data simply points to the
importance of benching this option as part of -tune.
…On Fri, Dec 27, 2024 at 4:20 PM Mihai Preda ***@***.***> wrote:
Thanks! we should make it ON by default (but guard it with LLVM builtin
checks: __has_builtin() )
—
Reply to this email directly, view it on GitHub
<#317 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AFJ4KM6MQFYPXIDTQJ6C3JT2HXAC5AVCNFSM6AAAAABUHW4ZFOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRUGAZTSNBVGA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
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Outdated suggestions cannot be applied.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 added an option to change non-temporal memory access from the command line. The default is off even though on Radeon VII it is about a 0.5% gaain. I suspect an A100 or recent AMD consumer cards with large caches will see bigger gains without nontemporal access. We could change the default setting to depend on the cache size reported by clinfo.
In either case, we should make -tune auto-config this option.