[DO NOT MERGE] start building Enzyme in CI

[ZuseZ4]

My goal is to put this in CI on April 26, to have a week to land some of the outstanding PRs (removed # in front of it to avoid spamming them every time I do a try build in this PR):
139700 [Ready]
139308 [Ready, blocked on 139700]
140104 [In the queue]
140030 [Landed]
140049 [Almost ready]
The autodiff flags PR should land first, but otherwise they don't overlap and are mostly ready, so it shouldn't be too hard to land them. In the meantime, I'll experiment here with some builders.

A fix for 140137 (Apple CI) would be nice, but I'm also happy to start with the working linux (x86-64 + aarch) as the most common target, and expand from there.

r? @oli-obk

Tracking:

• 124509

Linux
succeeded try with plugins & llvm-enzyme enabled: dist-x86_64-linux
succeeded try with plugins & llvm-enzyme enabled: dist-aarch64-linux

MacOS
failed try with plugins & llvm-enzyme enabled: dist-x86_64-apple (#140064 (comment))
failed try with plugins & llvm-enzyme enabled: dist-apple-various (#140064 (comment))
failed try with plugins & llvm-enzyme enabled: dist-aarch64-apple (#140064 (comment))

Windows
failed try with plugins & llvm-enzyme enabled: dist-x86_64-mingw (#140064 (comment))

try-job: dist-x86_64-linux
try-job: dist-aarch64-linux

[ZuseZ4]

@bors try

[bors]

⌛ Trying commit 1302345 with merge 85630b7...

[ZuseZ4]

@bors try @rust-timer queue

[bors]

⌛ Trying commit ca4cd19 with merge 22fa494d09e618a25269d354df51bccf741b5351...

[bors]

💔 Test failed - checks-actions

[ZuseZ4]

@jieyouxu I tried running both tests/ui/feature-gates/feature-gate-autodiff-use.rs and tests/ui/autodiff/visibility.rs with llvm.enzyme being either true or false in bootstrap.toml. I couldn't reproduce it that way. I also tried to copy the --cfg flags I saw in the log. Do you have recommendations on how to reproduce it locally (so I can update my rustc dev docs)?

[ZuseZ4]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 4b353ac with merge e690318f2c437976796a1aff8927ff8793b4d1b7...

[bors]

☀️ Try build successful - checks-actions
Build commit: e690318 (e690318f2c437976796a1aff8927ff8793b4d1b7)

[rust-timer]

Finished benchmarking commit (e690318): comparison URL.

Overall result: ❌ regressions - please read the text below

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	2.6%	[2.4%, 2.8%]	4
Regressions ❌ (secondary)	1.5%	[1.0%, 2.3%]	11
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.6%	[2.4%, 2.8%]	4

Max RSS (memory usage)

Results (primary 0.8%, secondary 1.9%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.9%	[0.4%, 1.8%]	27
Regressions ❌ (secondary)	3.2%	[1.5%, 7.4%]	9
Improvements ✅ (primary)	-0.7%	[-0.7%, -0.7%]	1
Improvements ✅ (secondary)	-1.8%	[-2.1%, -1.6%]	3
All ❌✅ (primary)	0.8%	[-0.7%, 1.8%]	28

Cycles

Results (primary 0.8%, secondary 2.3%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.8%	[0.4%, 2.8%]	10
Regressions ❌ (secondary)	2.3%	[1.8%, 3.0%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.8%	[0.4%, 2.8%]	10

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 775.102s -> 775.899s (0.10%)
Artifact size: 365.03 MiB -> 365.22 MiB (0.05%)

[bors]

⌛ Trying commit 9075e8e with merge 9a42842055443e5e9e4e3231cfef6c97c88c7270...

[bors]

💔 Test failed - checks-actions

[ZuseZ4]

@bors try

(It's a bit sad that all other runners stop when one fails)

[bors]

⌛ Trying commit 9075e8e with merge ad71ba0425b65e757d8d9b9e57c956608c6303c3...

[bors]

💔 Test failed - checks-actions

[ZuseZ4]

@bors try

[bors]

⌛ Trying commit 9075e8e with merge 8412a310bf3aa4d9b4801f29d856813a3f4db6a5...

[bors]

💔 Test failed - checks-actions

[ZuseZ4]

@bors try

[bors]

⌛ Trying commit 9075e8e with merge 6a4ba31...

[bors]

☀️ Try build successful - checks-actions
Build commit: 6a4ba31 (6a4ba3120960f17494a199fa78c9e84313eab700)

[ZuseZ4]

@bors try

[bors]

⌛ Trying commit a1d9041 with merge c59c2d4192bd80666bce3d4d5f5e0148a4ddfd8d...

[bors]

💔 Test failed - checks-actions

[ZuseZ4]

I guess linking against LLVM doesn't work here on mingw, but I don't really want to deal with linker issues right now, so I'll postpone windows support. The more relevant MSVC builders wouldn't work due to an Enzyme bug anyway.

[ZuseZ4]

Now testing Enzyme only, without enabling LLVM's plugin interface, since I made it optional in a previous enzyme PR. That might save us a few bytes.
@bors try

[bors]

⌛ Trying commit 142ec6f with merge f3fdf02...