[WIP] polymorphisation: re-enable

[davidtwco]

Do not merge this PR!
Before re-enabling polymorphisation, we're going to perform a crater run to make sure that there are no more regressions.

r? @lcnr

[davidtwco]

@bors try

[bors]

⌛ Trying commit 9dbbc9b7348a0d8385685610ca95a42f96c2cae5 with merge 38eae7d349c20cfc48c6beefd91af9499a4470cc...

[bors]

☀️ Try build successful - checks-actions, checks-azure
Build commit: 38eae7d349c20cfc48c6beefd91af9499a4470cc (38eae7d349c20cfc48c6beefd91af9499a4470cc)

[davidtwco]

@craterbot run mode=build-and-test

[craterbot]

👌 Experiment pr-75185 created and queued.
🤖 Automatically detected try build 38eae7d349c20cfc48c6beefd91af9499a4470cc
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[lcnr]

Should we rebase this and requeue the crater run, considering the recent changes?

[davidtwco]

Should we rebase this and requeue the crater run, considering the recent changes?

I'm tempted to, we'd only go back one spot in the queue.

[lcnr]

yeah, we can then also reuse this PR for perf ✨

[davidtwco]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion

[bors]

⌛ Trying commit 151b4c67ff89fd6ca485fddb114cb262ca4d96c5 with merge 0f97a070e9c9b216b3c18ed674269987207357b4...

[rust-log-analyzer]

The job dist-x86_64-linux of your PR failed (pretty log, raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

   Compiling itertools v0.9.0
   Compiling unicode-normalization v0.1.12
   Compiling getopts v0.2.21
   Compiling rustc_lexer v0.1.0 (/checkout/src/librustc_lexer)
error[E0391]: cycle detected when determining which generic parameters are unused by `Pack::SHIFT`
   --> /cargo/registry/src/jackfan.us.kg-1ecc6299db9ec823/sharded-slab-0.0.9/src/lib.rs:605:5
    |
605 |     const SHIFT: usize = Self::Prev::SHIFT + Self::Prev::LEN;
    |
    |
    = note: ...which again requires determining which generic parameters are unused by `Pack::SHIFT`, completing the cycle
note: cycle used when determining which generic parameters are unused by `page::slot::<impl at /cargo/registry/src/jackfan.us.kg-1ecc6299db9ec823/sharded-slab-0.0.9/src/page/slot.rs:51:1: 68:2>::LEN`
   --> /cargo/registry/src/jackfan.us.kg-1ecc6299db9ec823/sharded-slab-0.0.9/src/page/slot.rs:54:5
    |
54  |     const LEN: usize = (cfg::WIDTH - C::RESERVED_BITS) - Self::SHIFT;

   Compiling rustc_serialize v0.0.0 (/checkout/src/librustc_serialize)
error: aborting due to previous error


For more information about this error, try `rustc --explain E0391`.
error: could not compile `sharded-slab`.
To learn more, run the command again with --verbose.
warning: build failed, waiting for other jobs to finish...
error: build failed
error: build failed
command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "build" "--target" "x86_64-unknown-linux-gnu" "-Zbinary-dep-depinfo" "-j" "16" "--release" "--locked" "--color" "always" "--features" "jemalloc llvm" "--manifest-path" "/checkout/src/rustc/Cargo.toml" "--message-format" "json-render-diagnostics"
failed to run: /checkout/obj/build/bootstrap/debug/bootstrap dist --host x86_64-unknown-linux-gnu --target x86_64-unknown-linux-gnu
Build completed unsuccessfully in 0:13:19
== clock drift check ==
  local time: Sat Aug  8 18:53:23 UTC 2020
  local time: Sat Aug  8 18:53:23 UTC 2020
  network time: Sat, 08 Aug 2020 18:53:23 GMT
== end clock drift check ==
##[error]Process completed with exit code 1.
Terminate orphan process: pid (5528) (python)

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @rust-lang/infra. (Feature Requests)

[bors]

💔 Test failed - checks-actions

[lcnr]

@craterbot abort

[craterbot]

🗑️ Experiment pr-75185 deleted!

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[lcnr]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion

[bors]

⌛ Trying commit 86656408006a330ceaf5ef3106777f4ef954dee4 with merge 725c03a00dafa941e19ba48fbecd171d4af1ac02...

[bors]

☀️ Try build successful - checks-actions, checks-azure
Build commit: 725c03a00dafa941e19ba48fbecd171d4af1ac02 (725c03a00dafa941e19ba48fbecd171d4af1ac02)

[rust-timer]

Queued 725c03a00dafa941e19ba48fbecd171d4af1ac02 with parent 4b9ac51, future comparison URL.

[rust-timer]

Finished benchmarking try commit (725c03a00dafa941e19ba48fbecd171d4af1ac02): comparison url.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. Please note that if the perf results are neutral, you should likely undo the rollup=never given below by specifying rollup- to bors.

Importantly, though, if the results of this run are non-neutral do not roll this PR up -- it will mask other regressions or improvements in the roll up.

@bors rollup=never

[davidtwco]

Incremental has gotten really bad but everything else is mostly unaffected. I assume that no longer storing a "no unused params" result is worse when polymorphization is enabled.

I'd expect no real gains - we know that there aren't many polymorphization opportunities (especially with the script benchmark removed) in the benchmark suite.

[Mark-Simulacrum]

I'd expect no real gains - we know that there aren't many polymorphization opportunities (especially with the script benchmark removed) in the benchmark suite.

To be clear, this is with current polymorphization, right? We know of at least one case that I identified last night that could be a win (drop_in_place).

Regardless -- I am personally hesitant to see it enabled by default with no wins and only losses on the nightly/beta/stable compilers. Maybe we can just enable it in rustc's test suite? (At least on some builders, so we're still testing both?) Or do we expect nightly coverage to be really beneficial?

[davidtwco]

Regardless -- I am personally hesitant to see it enabled by default with no wins and only losses on the nightly/beta/stable compilers. Maybe we can just enable it in rustc's test suite? (At least on some builders, so we're still testing both?) Or do we expect nightly coverage to be really beneficial?

I'm not that interested in actually enabling polymorphization in this PR - it primarily exists to do a crater run.

I intend to wait until we support more cases and have a better understanding of how incr-comp is interacting with polymorphization.

To be clear, this is with current polymorphization, right? We know of at least one case that I identified last night that could be a win (drop_in_place).

Yeah - absolutely. Hopefully as polymorphization becomes more capable we'll see that in the benchmark suite.

[Dylan-DPC]

@davidtwco any updates on this? do we need to still keep it open?

[davidtwco]

@davidtwco any updates on this? do we need to still keep it open?

No updates, we're still working on fixing some of the bugs with polymorphization then we'll look at resurrecting this.

[JohnCSimon]

@davidtwco
Ping from triage: I'm closing this due to inactivity, Please reopen when you are ready to continue with this.
Note: if you do please open the PR BEFORE you push to it, else you won't be able to reopen - this is a quirk of github.

@rustbot label: +S-inactive