release-1.7: Backports for 1.7-RC1 #41781
Merged
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
(cherry picked from commit 131b75b)
(cherry picked from commit 7005b7d)
Fixes dynamic arch detection on certain ARMv8 cores. See OpenMathLib/OpenBLAS#3060 (cherry picked from commit cd6e67f)
I am experimenting with showing some other information besides just the types here as well and I think this would be generally useful. This passes that information as an `IOContext` as to not break any downstream uses of this code. (cherry picked from commit c2b4b38)
Avoids an assert when this is triggered (cherry picked from commit 2d81742)
(cherry picked from commit 01d439e)
This adds `lbt_find_backing_library()`, which is a useful debugging routine to allow advanced users/package authors to query LBT to determine which backing BLAS library will service a particular BLAS call. It also exposes the "footgun API", which allows users to directly set/get forwarding on a per-function basis. Because this has the ability to generate truly bizarre setups, we do not advertise this capability broadly (simply using `lbt_forward()` should be enough for most usecases) however it's nice to have wrapped. (cherry picked from commit b40ae6b)
(cherry picked from commit 9ae49bc)
These get invalidated by loading Static.jl, specifically the method
```
Base.convert(::Type{T}, ::StaticInt{N}) where {T<:Number,N} = convert(T, N)
```
(cherry picked from commit 232ee11)
ChainRulesCore defines `==(a, b::AbstractThunk)` and its converse, and these end up invaliding parts of the REPL (including `eval_user_input`) via inference failures in `modules_to_be_loaded`. Co-authored by: Jameson Nash <vtjnash@gmail.com> (cherry picked from commit 7a6336d)
Zero and copy only the used portion of the union state buffer. (cherry picked from commit 0258553)
(cherry picked from commit 310bf16)
(cherry picked from commit 6e1bae4)
(cherry picked from commit aa421ff)
Lacking inlining, but now expressing the direct invoke: this gets us within about 2x of a primitive atomicrmw add. (cherry picked from commit 85518c8)
* inference: propagate variable changes to all exception frames Fix #42022 * Update test/compiler/inference.jl * Update test/compiler/inference.jl Co-authored-by: Shuhei Kadowaki <40514306+aviatesk@users.noreply.github.com> * fixup! inference: propagate variable changes to all exception frames Co-authored-by: Shuhei Kadowaki <40514306+aviatesk@users.noreply.github.com> (cherry picked from commit e83b317)
(cherry picked from commit e1669b6)
KristofferC
force-pushed
the
backports-release-1.7
branch
from
886b922
to
c2c5a83
Compare
|
Something went wrong when running your job: Unfortunately, the logs could not be uploaded. |
|
(Q: did you mean to compare to master?) |
|
No.. Fail :( Although, it is surprising that: |
|
@nanosoldier |
|
Something went wrong when running your job: Unfortunately, the logs could not be uploaded. |
|
I've uploaded without the data files: https://github.com/JuliaCI/NanosoldierReports/blob/master/benchmark/by_hash/0c1a860_vs_11e64d1/report.md But note also the very large number of errors reported in the "primary" log files: https://raw.githubusercontent.com/JuliaCI/NanosoldierReports/master/benchmark/by_hash/0c1a860_vs_11e64d1/logs/0c1a860ce5fb2f440bfd23be5a98343c28d39aef_primary.err due to c2c5a83 being backported incorrectly or without 795935f (#41697) |
|
Scary that all the tests passed... |
KristofferC
force-pushed
the
backports-release-1.7
branch
from
c2c5a83
to
c85012a
Compare
Separated from compiler-plugin prototyping. cherry-picked from 799136d
(this PR is the final output of my demo at [our workshop](https://github.com/aviatesk/juliacon2021-workshop-pkgdev)) This PR eliminated much of runtime dispatches within our type inference routine, that are reported by the following JET analysis: ```julia using JETTest const CC = Core.Compiler function function_filter(@nospecialize(ft)) ft === typeof(CC.isprimitivetype) && return false ft === typeof(CC.ismutabletype) && return false ft === typeof(CC.isbitstype) && return false ft === typeof(CC.widenconst) && return false ft === typeof(CC.widenconditional) && return false ft === typeof(CC.widenwrappedconditional) && return false ft === typeof(CC.maybe_extract_const_bool) && return false ft === typeof(CC.ignorelimited) && return false return true end function frame_filter((; linfo) = sv) meth = linfo.def isa(meth, Method) || return true return occursin("compiler/", string(meth.file)) end report_dispatch(CC.typeinf, (CC.NativeInterpreter, CC.InferenceState); function_filter, frame_filter) ``` > on master ``` ═════ 137 possible errors found ═════ ... ``` > on this PR ``` ═════ 51 possible errors found ═════ ... ``` And it seems like this PR makes JIT slightly faster: > on master ```julia ~/julia/julia master ❯ ./usr/bin/julia -e '@time using Plots; @time plot(rand(10,3));' 3.659865 seconds (7.19 M allocations: 497.982 MiB, 3.94% gc time, 0.39% compilation time) 2.696410 seconds (3.62 M allocations: 202.905 MiB, 7.49% gc time, 56.39% compilation time) ``` > on this PR ```julia ~/julia/julia avi/jetdemo* 7s ❯ ./usr/bin/julia -e '@time using Plots; @time plot(rand(10,3));' 3.396974 seconds (7.16 M allocations: 491.442 MiB, 4.80% gc time, 0.28% compilation time) 2.591130 seconds (3.48 M allocations: 196.026 MiB, 7.29% gc time, 56.72% compilation time) ``` cherry-picked from 795935f
release: backport complicated compiler changes
|
Let's try with non-broken inference then. Third time's the charm? @nanosoldier |
|
Something went wrong when running your job: Unfortunately, the logs could not be uploaded. |
KristofferC
changed the title
|
Awesome! What are your thoughts on the likelihood of getting #42125 in? In a sense maybe it doesn't matter since it's not a supported name, but it's not crazy to think that a renaming post-rc1 is also a bit weird. |
|
Are those names part of the public API/do those names appear in the documentation? If not, it wouldn't be breaking to change the names in 1.8, right? |
|
Not part of the public API, so we can break them without consequence. |
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.
platform_linuxjobs, after building Julia, make sure that the working directory is clean #41796pull_request_targetevent instead of thepull_requestevent #41814tester_jobs #41822macosaarch64andmusl64statuses #41824tester_linuxppc64lestatus #41831tester_linuxarmv7lstatus #41832permissionskey in the workflow file for the "Statuses" workflow #41851llvmpasses: use the latest rootfs image #41864modules_to_be_loaded#41878base-buildkite-docsrepository #41956jl_cumulative_compile_time_nsglobal (and reentrant). #41733LimitedAccuracyhandlings #42034abstract_iteration#41839ConditionalinPartialStruct#42091cmdlineargstests on Buildkite #42118Non-merged PRs with backport label:
package_linuxandtester_linuxfor more Linux architectures #41707usingthread-safe #41602