aarch64: Use LDP/STP if FEAT_LSE2 is available #11
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
taiki-e
added
the
O-arm
taiki-e
force-pushed
the
lse2
branch
3 times, most recently
from
568ac2c to
6ddf9be
Compare
taiki-e
changed the title
taiki-e
commented
Apr 30, 2022
| @@ -26,6 +26,7 @@ test_task: | |||
| - RUSTFLAGS="$RUSTFLAGS -C target-feature=+lse" RUSTDOCFLAGS="$RUSTDOCFLAGS -C target-feature=+lse" cargo test -vv --workspace --exclude asm-test --exclude bench --all-features --release -- -Z unstable-options --report-time | |||
| # Use -Z build-std because the prebuilt libtest seems to be incompatible with LTO, causing miscompilation: https://gist.github.com/taiki-e/9713f8e02e8f9f852ccee8d6f089ec24 | |||
| - RUSTFLAGS="$RUSTFLAGS -C target-feature=+lse" RUSTDOCFLAGS="$RUSTDOCFLAGS -C target-feature=+lse" CARGO_PROFILE_RELEASE_CODEGEN_UNITS=1 CARGO_PROFILE_RELEASE_LTO=fat cargo -Z build-std test -vv --workspace --exclude asm-test --exclude bench --all-features --release --tests --target $TARGET -- -Z unstable-options --report-time | |||
| # TODO: lse2 is not available on Graviton2 (armv8.2-a) | |||
There was a problem hiding this comment.
When Cirrus CI moved to Graviton3 (armv8.5-a? IIRC EDIT: armv8.4-a) we can test this with real aarsh64 machines on CI.
| @@ -62,3 +63,4 @@ valgrind_task: | |||
| - RUSTFLAGS="$RUSTFLAGS -C target-feature=+lse" RUSTDOCFLAGS="$RUSTDOCFLAGS -C target-feature=+lse" cargo test -vv --workspace --exclude asm-test --exclude bench --all-features --release -Z doctest-xcompile -- -Z unstable-options --report-time | |||
| # Use -Z build-std because the prebuilt libtest seems to be incompatible with LTO, causing miscompilation: https://gist.github.com/taiki-e/9713f8e02e8f9f852ccee8d6f089ec24 | |||
| - RUSTFLAGS="$RUSTFLAGS -C target-feature=+lse" RUSTDOCFLAGS="$RUSTDOCFLAGS -C target-feature=+lse" CARGO_PROFILE_RELEASE_CODEGEN_UNITS=1 CARGO_PROFILE_RELEASE_LTO=fat cargo -Z build-std test -vv --workspace --exclude asm-test --exclude bench --all-features --release --tests --target $TARGET -- -Z unstable-options --report-time | |||
| # TODO: lse2 is not available on Graviton2 (armv8.2-a) | |||
There was a problem hiding this comment.
I have not confirmed if valgrind can handle CPU features not supported by the host.
|
Multithreaded tests using QEMU with cpu=a64fx failed due to torn reads, but a64fx is armv8.2-a so this is fine. However, considering that QEMU with cpu=max also failed, it is also possible that QEMU has not yet properly implemented FEAT_LSE2. (Singlethread tests using QEMU succeed) |
taiki-e
commented
Apr 30, 2022
Comment on lines
+177
to
+186
| // As of rustc nightly-2022-04-30, target_feature "lse2" is not available on rustc side: | ||
| // https://github.com/rust-lang/rust/blob/d201c812d40932509b2b5307c0b20c1ce78d21da/compiler/rustc_codegen_ssa/src/target_features.rs#L45 |
There was a problem hiding this comment.
LLVM (and std_detect) support it, so using it is not a problem, but we need to parse rustflags in the build script.
|
cc @ibraheemdev |
ibraheemdev
approved these changes
May 1, 2022
taiki-e
force-pushed
the
lse2
branch
6 times, most recently
from
7cdbb66 to
96f80c6
Compare
taiki-e
added a commit
to taiki-e/atomic-maybe-uninit
that referenced
this pull request
May 30, 2022
taiki-e
added a commit
to taiki-e/atomic-maybe-uninit
that referenced
this pull request
May 30, 2022
taiki-e
added a commit
to taiki-e/atomic-maybe-uninit
that referenced
this pull request
May 30, 2022
taiki-e
added a commit
to taiki-e/atomic-maybe-uninit
that referenced
this pull request
May 30, 2022
taiki-e
added a commit
to taiki-e/atomic-maybe-uninit
that referenced
this pull request
May 30, 2022
taiki-e
added a commit
to taiki-e/atomic-maybe-uninit
that referenced
this pull request
May 30, 2022
taiki-e
added
the
O-aarch64
taiki-e
removed
the
O-arm
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.
aarch64 part of #10