[do not merge] Measure upper limit for performance of 32 bit Span

[petrochenkov]

What would happen if span interner accesses were cheap and didn't require any synchronization?

r? @ghost

[petrochenkov]

@bors try

[bors]

⌛ Trying commit 6930ecbb0c927fb458bfde7fa6bb82be26ddef18 with merge 79b69238624df985b4415688efea25470ba8a824...

[petrochenkov]

TODO: Exclude rustdoc from bors try for this PR.

[nnethercote]

This will be interesting. #59693 (comment) suggests that roughly half the speedup of 64-bit spans comes from fewer TLS accesses, and the other half comes from fewer hash table accesses.

[petrochenkov]

@bors try

[bors]

⌛ Trying commit 7c2ab8d with merge dc7e5bd...

[petrochenkov]

Oh, come on, why x.py dist tests rustbook, I disabled it.

[petrochenkov]

@bors try

[bors]

⌛ Trying commit ed9912de2f7fc10c1f3ec969d9c835c833c374b8 with merge 3691212ca5933a51857de07b5a62084dabe81b82...

[petrochenkov]

@bors try

[bors]

⌛ Trying commit af6e2d7 with merge 026939b...

[petrochenkov]

@bors try

[bors]

⌛ Trying commit 4f208af with merge aff596d...

[rust-highfive]

The job mingw-check of your PR failed on Travis (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.

travis_time:end:2f0f8140:start=1554970395325370316,finish=1554970397671287397,duration=2345917081
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---
[00:03:16] configure: build.locked-deps    := True
[00:03:16] configure: llvm.ccache          := sccache
[00:03:16] configure: build.cargo-native-static := True
[00:03:16] configure: dist.missing-tools   := True
[00:03:16] configure: build.configure-args := ['--enable-sccache', '--disable-manage-submodu ...
[00:03:16] configure: writing `config.toml` in current directory
[00:03:16] configure: 
[00:03:16] configure: run `python /checkout/x.py --help`
[00:03:16] configure: 

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 @TimNN. (Feature Requests)

[petrochenkov]

cc @Mark-Simulacrum
Any hints on how to "disable running any rustdoc" in rustbuild on bors try?
So far bors try was behaving entirely differently from what I see on local attemts to run dist with or without docker/run.sh.

[bors]

💔 Test failed - checks-travis

[rust-highfive]

The job dist-x86_64-linux of your PR failed on Travis (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.

travis_fold:end:services

travis_fold:start:git.checkout
travis_time:start:0e9b073d
$ git clone --depth=2 --branch=try https://github.com/rust-lang/rust.git rust-lang/rust
---

[01:44:09] travis_time:end:stage2-cargo-miri:start=1554976672979278033,finish=1554976673467240019,duration=487961986

[01:44:11] Dist docs (x86_64-unknown-linux-gnu)
[01:44:11] thread 'main' panicked at 'fs::read_dir(src) failed with No such file or directory (os error 2)', src/bootstrap/lib.rs:1228:18
[01:44:11] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap dist --host x86_64-unknown-linux-gnu --target x86_64-unknown-linux-gnu
[01:44:11] Build completed unsuccessfully in 1:38:17
travis_time:end:2dff1b77:start=1554970423575529284,finish=1554976676247734959,duration=6252672205675
The command "stamp sh -x -c "$RUN_SCRIPT"" exited with 1.
---
travis_time:end:05a5f348:start=1554976678854571109,finish=1554976678877608078,duration=23036969
travis_fold:end:after_failure.3
travis_fold:start:after_failure.4
travis_time:start:00aa6bf8
$ ln -s . checkout && for CORE in obj/cores/core.*; do EXE=$(echo $CORE | sed 's|obj/cores/core\.[0-9]*\.!checkout!\(.*\)|\1|;y|!|/|'); if [ -f "$EXE" ]; then printf travis_fold":start:crashlog\n\033[31;1m%s\033[0m\n" "$CORE"; gdb --batch -q -c "$CORE" "$EXE" -iex 'set auto-load off' -iex 'dir src/' -iex 'set sysroot .' -ex bt -ex q; echo travis_fold":"end:crashlog; fi; done || true
travis_fold:end:after_failure.4
travis_fold:start:after_failure.5
travis_time:start:0d29359a
travis_time:start:0d29359a
$ cat ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers || true
cat: ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers: No such file or directory
travis_fold:end:after_failure.5
travis_fold:start:after_failure.6
travis_time:start:057f9960
$ dmesg | grep -i kill

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 @TimNN. (Feature Requests)

[petrochenkov]

@bors try

[bors]

⌛ Trying commit 955ca4c with merge d806815...

[rust-highfive]

The job mingw-check of your PR failed on Travis (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.

travis_time:end:0004aec0:start=1555009905052951606,finish=1555009907088588091,duration=2035636485
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---
[00:04:41]     Checking rustc_data_structures v0.0.0 (/checkout/src/librustc_data_structures)
[00:04:43]     Checking arena v0.0.0 (/checkout/src/libarena)
[00:04:43]     Checking syntax_pos v0.0.0 (/checkout/src/libsyntax_pos)
[00:04:45]     Checking rustc_errors v0.0.0 (/checkout/src/librustc_errors)
[00:04:51] error[E0599]: no method named `as_ptr` found for type `lock_api::mutex::Mutex<parking_lot::raw_mutex::RawMutex, syntax_pos::span_encoding::SpanInterner>` in the current scope
[00:04:51]    --> src/libsyntax/lib.rs:105:64
[00:04:51]     |
[00:04:51] 105 |                     globals.syntax_pos_globals.span_interner.0.as_ptr();
[00:04:51] 
[00:04:57] error: aborting due to previous error
[00:04:57] 
[00:04:57] For more information about this error, try `rustc --explain E0599`.
---
travis_time:end:013531b4:start=1555010228105362502,finish=1555010228109808459,duration=4445957
travis_fold:end:after_failure.3
travis_fold:start:after_failure.4
travis_time:start:0d01932e
$ ln -s . checkout && for CORE in obj/cores/core.*; do EXE=$(echo $CORE | sed 's|obj/cores/core\.[0-9]*\.!checkout!\(.*\)|\1|;y|!|/|'); if [ -f "$EXE" ]; then printf travis_fold":start:crashlog\n\033[31;1m%s\033[0m\n" "$CORE"; gdb --batch -q -c "$CORE" "$EXE" -iex 'set auto-load off' -iex 'dir src/' -iex 'set sysroot .' -ex bt -ex q; echo travis_fold":"end:crashlog; fi; done || true
travis_fold:end:after_failure.4
travis_fold:start:after_failure.5
travis_time:start:000f4746
travis_time:start:000f4746
$ cat ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers || true
cat: ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers: No such file or directory
travis_fold:end:after_failure.5
travis_fold:start:after_failure.6
travis_time:start:09077760
$ dmesg | grep -i kill

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 @TimNN. (Feature Requests)

[bors]

☀️ Try build successful - checks-travis
Build commit: d806815

[petrochenkov]

@rust-timer build d806815

[rust-timer]

Success: Queued d806815 with parent 3de0106, comparison URL.

[rust-timer]

Finished benchmarking try commit d806815

[nnethercote]

The comparison URL is empty, showing an error message about "could not find commit..." :(

[Mark-Simulacrum]

Yeah, that's a known bug. The parent commit is still building.

[matprec]

Parent commit is finished

[nnethercote]

This gets roughly 75% of the Span64 improvements (#59693).

[petrochenkov]

@nnethercote
On the largest crates that's true, but I wouldn't say that in general.
Here's the diff between this PR (the first column) and that PR (the second column).
https://perf.rust-lang.org/compare.html?start=d80681574370cb24650e053d4f30bf0e28f8b923&end=cfa77658d2ba937d22801e3a5222fcfc7f63ca47

[petrochenkov]

I'd say that if we had perfectly inline-able TLS accesses or explicitly passed interner, then 64 bit spans would rather be a regression for most of crates beside script-servo.