-
Notifications
You must be signed in to change notification settings - Fork 12.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
travis: Disable LLVM assertions on OSX #44610
Conversation
Our OSX builders are routinely and significantly over hour 2 hour "soft limit" for testing PRs. I *think* that a big portion of this time comes from the fact that LLVM and debug assertions are enabled. In an effort to speed up these builders and reduce cycle time this commit disables LLVM assertions on OSX for all builders. My thinking is that we'll let this bake for a bit after merged to see what the effect is on timing on Travis. If it doesn't actually help much we can turn them back on, and if it doesn't help enough we can disable Rust debug assertions as well.
(rust_highfive has picked a reviewer for you, use r? to override) |
I've conferred with @nikomatsakis and @eddyb on IRC about this from the compiler side of things, and figure it's worth at least a test run to see what's up. |
@bors r+ |
📌 Commit 97fee3e has been approved by |
@bors p=10 (that way we see the results sooner) |
travis: Disable LLVM assertions on OSX Our OSX builders are routinely and significantly over hour 2 hour "soft limit" for testing PRs. I *think* that a big portion of this time comes from the fact that LLVM and debug assertions are enabled. In an effort to speed up these builders and reduce cycle time this commit disables LLVM assertions on OSX for all builders. My thinking is that we'll let this bake for a bit after merged to see what the effect is on timing on Travis. If it doesn't actually help much we can turn them back on, and if it doesn't help enough we can disable Rust debug assertions as well.
💔 Test failed - status-travis |
⌛ Testing commit 97fee3e with merge 7ab10ea2e11b7ab056afd483552e43a94b8c2671... |
@bors: retry prioritizing rollups |
⌛ Testing commit 97fee3e with merge 76f64f82d6973449b3dd24e04285683ce8d5c634... |
@bors: retry |
…ark-Simulacrum travis: Disable LLVM assertions on OSX Our OSX builders are routinely and significantly over hour 2 hour "soft limit" for testing PRs. I *think* that a big portion of this time comes from the fact that LLVM and debug assertions are enabled. In an effort to speed up these builders and reduce cycle time this commit disables LLVM assertions on OSX for all builders. My thinking is that we'll let this bake for a bit after merged to see what the effect is on timing on Travis. If it doesn't actually help much we can turn them back on, and if it doesn't help enough we can disable Rust debug assertions as well.
⌛ Testing commit 97fee3e with merge d248841178b67be9de0571b3a63d7355986b8942... |
@bors: retry |
travis: Disable LLVM assertions on OSX Our OSX builders are routinely and significantly over hour 2 hour "soft limit" for testing PRs. I *think* that a big portion of this time comes from the fact that LLVM and debug assertions are enabled. In an effort to speed up these builders and reduce cycle time this commit disables LLVM assertions on OSX for all builders. My thinking is that we'll let this bake for a bit after merged to see what the effect is on timing on Travis. If it doesn't actually help much we can turn them back on, and if it doesn't help enough we can disable Rust debug assertions as well.
☀️ Test successful - status-appveyor, status-travis |
Doesn't seem effective. The two |
This PR itself had to populate the sccache caches as it recompiled LLVM, I think it's more useful to look at the build just before and the build just after which shows:
Almost 20 minute improvements! I'll submit another PR for debug assertions in rustc. |
This commit disables debug assertions for OSX in an effort to improve cycle time on OSX. It looks like rust-lang#44610 didn't shave off quite as much time as desired so let's see how much this helps.
…rk-Simulacrum ci: Disable rustc debug assertions on OSX This commit disables debug assertions for OSX in an effort to improve cycle time on OSX. It looks like #44610 didn't shave off quite as much time as desired so let's see how much this helps.
…acrum Remove apple-alt dist build. This removes the dist-x86_64-apple-alt build to reduce CI usage because I suspect nobody is using it. This builder is almost identical to the `dist-x86_64-apple` with the small difference that the latter adds `rust.lto=thin`, and I do not think that difference was intentional. The reason they are identical is because llvm assertions were disabled in rust-lang#44610, but I did not see any discussion about the consequence that this made the alt build identical to the normal build. Perhaps because rust-lang#44610 was intended to be temporary?
Our OSX builders are routinely and significantly over hour 2 hour "soft limit"
for testing PRs. I think that a big portion of this time comes from the fact
that LLVM and debug assertions are enabled. In an effort to speed up these
builders and reduce cycle time this commit disables LLVM assertions on OSX for
all builders.
My thinking is that we'll let this bake for a bit after merged to see what the
effect is on timing on Travis. If it doesn't actually help much we can turn them
back on, and if it doesn't help enough we can disable Rust debug assertions as
well.