Skip to content
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

Activate NewPM support #46824

Merged
merged 3 commits into from
May 4, 2023
Merged

Activate NewPM support #46824

merged 3 commits into from
May 4, 2023

Conversation

pchintalapudi
Copy link
Member

The NewPM changes have gone in under a #define statement due to verification errors; this PR will be used to track how many of those errors still exist.

@pchintalapudi pchintalapudi force-pushed the pc/newpm_activate branch 2 times, most recently from ff413d4 to 09de8e9 Compare September 19, 2022 21:26
@pchintalapudi
Copy link
Member Author

@nanosoldier runtests(ALL, vs = ":master", buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"], vs_buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"])

@pchintalapudi
Copy link
Member Author

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@pchintalapudi
Copy link
Member Author

@nanosoldier runtests(ALL, vs = "%self", configuration = (buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"],))

@nanosoldier
Copy link
Collaborator

Your package evaluation job has completed - possible new issues were detected. A full report can be found here.

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@pchintalapudi
Copy link
Member Author

@nanosoldier runtests(ALL, vs = ":master", configuration = (buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"], vs_buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"]))

@vchuravy
Copy link
Member

@nanosoldier runtests(ALL, vs = ":master", configuration = (;buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"]))

@vchuravy
Copy link
Member

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@pchintalapudi
Copy link
Member Author

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@pchintalapudi
Copy link
Member Author

@nanosoldier runtests(ALL, vs = ":master", configuration = (buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"],))

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@nanosoldier
Copy link
Collaborator

Your package evaluation job has completed - possible new issues were detected. A full report can be found here.

@pchintalapudi
Copy link
Member Author

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@pchintalapudi pchintalapudi force-pushed the pc/newpm_activate branch 2 times, most recently from 4405053 to 2f6a2a9 Compare March 8, 2023 18:31
@giordano
Copy link
Contributor

@oscardssmith was this PR closed on purpose?

@oscardssmith
Copy link
Member

As I understand it, #47038 supersedes this.

@giordano
Copy link
Contributor

It'd be helpful to motivate why PRs are closed though

@giordano giordano deleted the pc/newpm_activate branch April 17, 2023 19:28
@vchuravy vchuravy restored the pc/newpm_activate branch April 17, 2023 19:38
@vchuravy vchuravy reopened this Apr 17, 2023
@vchuravy vchuravy force-pushed the pc/newpm_activate branch from 39c2fb6 to 036230c Compare April 17, 2023 19:39
@gbaraldi
Copy link
Member

It segfaults when receiving a specific target instead of erroring out, I guess it needs some separate cmdline arg handling

@vchuravy vchuravy force-pushed the pc/newpm_activate branch from 036230c to 5c8208b Compare April 29, 2023 23:19
@vchuravy vchuravy changed the base branch from master to vc/upgrade_llvm15 April 29, 2023 23:19
@vchuravy
Copy link
Member

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@vchuravy
Copy link
Member

@nanosoldier runtests(configuration = (buildflags=["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"],), vs_configuration = (buildflags = ["LLVM_ASSERTIONS=1", "FORCE_ASSERTIONS=1"],))

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@pchintalapudi
Copy link
Member Author

Top 10 newpm regressions:

["union", "array", ("perf_sum4", "ComplexF64", 0)]: 2.42 (5%)
["union", "array", ("perf_sum4", "ComplexF64", 1)]: 2.41 (5%)
["tuple", "reduction", ("sum", "(16, 16)")]: 2.24 (5%)
["union", "array", ("map", "*", "Int8", "(false, false)")]: 2.15 (5%)
["union", "array", ("map", "identity", "Int8", 0)]: 2.13 (5%)
["union", "array", ("broadcast", "identity", "Bool", 1)]: 2.09 (5%)
["union", "array", ("broadcast", "identity", "Int8", 1)]: 2.08 (5%)
["union", "array", ("skipmissing", "sum", "ComplexF64", 0)]: 2.07 (5%)
["union", "array", ("broadcast", "*", "Float64", "(false, false)")]: 2.04 (5%)
["union", "array", ("collect", "all", "Int8", 0)]: 2.03 (5%)

vs with legacy pm on LLVM 15:

["broadcast", "mix_scalar_tuple", (5, "tup_tup")]: 4.24 (5%)
["broadcast", "dotop", ("Float64", "(1000, 1000)", 2)]: 3.88 (5%)
["union", "array", ("broadcast", "identity", "Bool", 1)]: 2.53 (5%)
["union", "array", ("broadcast", "identity", "Int8", 1)]: 2.49 (5%)
["union", "array", ("broadcast", "*", "Int8", "(false, false)")]: 2.44 (5%)
["union", "array", ("broadcast", "*", "Float32", "(false, false)")]: 2.35 (5%)
["union", "array", ("broadcast", "abs", "Float32", 0)]: 2.33 (5%)
["union", "array", ("broadcast", "*", "Float64", "(false, false)")]: 2.2 (5%)
["union", "array", ("broadcast", "*", "Int64", "(false, false)")]: 2.2 (5%)
["union", "array", ("broadcast", "abs", "Bool", 1)]: 2.2 (5%) 

Looks like newpm is overall has fewer regressions compared to master, beating LLVM 15 legacypm on the two broadcast benchmarks and being roughly equal otherwise.

@nanosoldier
Copy link
Collaborator

Your package evaluation job has completed - possible new issues were detected.
A full report can be found here.

@vchuravy vchuravy force-pushed the pc/newpm_activate branch from 5c8208b to 512f688 Compare May 3, 2023 02:30
@vchuravy
Copy link
Member

vchuravy commented May 3, 2023

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@vchuravy
Copy link
Member

vchuravy commented May 3, 2023

@nanosoldier runbenchmarks(!"scalar", vs=":vc/upgrade_llvm15")

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@pchintalapudi pchintalapudi force-pushed the pc/newpm_activate branch 2 times, most recently from 3cc1d4d to 86fadcd Compare May 3, 2023 18:32
@vchuravy
Copy link
Member

vchuravy commented May 3, 2023

@nanosoldier runbenchmarks(!"scalar", vs=":master")

@vchuravy
Copy link
Member

vchuravy commented May 3, 2023

@nanosoldier runbenchmarks(!"scalar", vs=":vc/upgrade_llvm15")

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@nanosoldier
Copy link
Collaborator

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

@vchuravy vchuravy marked this pull request as ready for review May 4, 2023 02:47
@vchuravy vchuravy merged commit 32ef1a7 into vc/upgrade_llvm15 May 4, 2023
@vchuravy vchuravy deleted the pc/newpm_activate branch May 4, 2023 02:47
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants