Skip to content
This repository has been archived by the owner on Jan 20, 2022. It is now read-only.

Support for data-parallelism for adjacent find #22

Closed
wants to merge 24 commits into from
Closed

Support for data-parallelism for adjacent find #22

wants to merge 24 commits into from

Conversation

hkaiser
Copy link
Member

@hkaiser hkaiser commented Dec 26, 2021 •
edited

Supersedes STEllAR-GROUP/hpx#5651, please see there for remaining issues and comments.

@hkaiser hkaiser added this to the 0.1.0 milestone Dec 26, 2021
This was referenced Dec 26, 2021
Closed
Open
hkaiser and others added 10 commits December 26, 2021 14:58
@hkaiser
Merge pull request #19 from STEllAR-GROUP/noexcept_assert
da9c37c 
Removing HPX_NOEXCEPT_WITH_ASSERT
@hkaiser
Fixing problems created by moving PR to new repository
e920b0c
@msimberg
Merge #18
10b3d44 
18: Add `for_each` benchmark to performance test CI r=msimberg a=msimberg

- Moves the generic `for_each` benchmarking functionality to a separate header.
- Adds a CI "report" version of the `for_each` benchmark (always tests `parallel_executor`, `fork_join_executor`, `scheduler_executor`, and a plain for loop).
- Adds the `scheduler_executor` to the `foreach_scaling` benchmark.
- Explicitly sets the `hpx.thread_queue.init_threads_count` to a higher value in the perf tests to avoid warmup problems (I may have to come back to the default value of this option).

To do:
- [ ] Update the reference for the new benchmark (it's currently a dummy result from my laptop).

Co-authored-by: Mikael Simberg <mikael.simberg@iki.fi>
@srinivasyadav18
Add CPO to adjacent find
2696498
@srinivasyadav18
Adapt datapar adjacent find
817bdc2
@srinivasyadav18
Split adjacent find unit tests into header and source
25127d9
@srinivasyadav18
Add unit tests for datapar adjacentfind
77ef189
@srinivasyadav18
Replace HPX_INLINE_CONSTEXPR_VARIABLE with inline constexpr
30ec3a1
@srinivasyadav18
Fallback using sequential policies when iterator is not datapar compa…
b214a34 
…tible
@hkaiser @srinivasyadav18
Fixing problems created by moving PR to new repository
8b08c93
@srinivasyadav18
Minor fixes
146c914 
Replace std::distance with parallel::detail::distance
Replace include/datapar.hpp with parallel/datapar.hpp
@srinivasyadav18
Do not forward pred_projected
a654774
@srinivasyadav18
Merge branch 'hpx_5651' of https://github.com/STEllAR-GROUP/hpx-local …
2bc28b1 
…into hpx_5651
@StellarBot
Copy link

Performance test report

HPXLocal Performance

Comparison

BENCHMARKFORK_JOIN_EXECUTORPARALLEL_EXECUTORSCHEDULER_EXECUTOR
For Each(=)(=)-

Info

PropertyBeforeAfter
HPXLocal Commit01e4980ace71c6
HPXLocal Datetime2021-12-21T15:01:46+00:002022-01-14T05:06:53+00:00
Clusternamedaintdaint
Hostnamenid00932nid01452
Datetime2021-12-21T16:09:15.238666+01:002022-01-14T06:13:48.074126+01:00
Envfile
Compiler/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0

Comparison

BENCHMARKNO-EXECUTOR
Future Overhead - Create Thread Hierarchical - Latch+

Info

PropertyBeforeAfter
HPXLocal Commitf499a22ace71c6
HPXLocal Datetime2021-12-10T13:50:04+00:002022-01-14T05:06:53+00:00
Clusternamedaintdaint
Hostnamenid00243nid01452
Datetime2021-12-10T15:19:42.442217+01:002022-01-14T06:14:04.716659+01:00
Envfile
Compiler/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0

Comparison

BENCHMARKFORK_JOIN_EXECUTORPARALLEL_EXECUTORSCHEDULER_EXECUTOR
Stream Benchmark - Add(=)(=)(=)
Stream Benchmark - Scale(=)(=)(=)
Stream Benchmark - Triad(=)(=)(=)
Stream Benchmark - Copy(=)(=)(=)

Info

PropertyBeforeAfter
HPXLocal Commitf64fbd0ace71c6
HPXLocal Datetime2021-11-12T11:29:27+00:002022-01-14T05:06:53+00:00
Clusternamedaintdaint
Hostnamenid00007nid01452
Datetime2021-11-12T12:57:50.824026+01:002022-01-14T06:14:19.265872+01:00
Envfile
Compiler/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0

Explanation of Symbols

SymbolMEANING
=No performance change (confidence interval within ±1%)
(=)Probably no performance change (confidence interval within ±2%)
(+)/(-)Very small performance improvement/degradation (≤1%)
+/-Small performance improvement/degradation (≤5%)
++/--Large performance improvement/degradation (≤10%)
+++/---Very large performance improvement/degradation (>10%)
?Probably no change, but quite large uncertainty (confidence interval with ±5%)
??Unclear result, very large uncertainty (±10%)
???Something unexpected…

@StellarBot
Copy link

Performance test report

HPXLocal Performance

Comparison

BENCHMARKFORK_JOIN_EXECUTORPARALLEL_EXECUTORSCHEDULER_EXECUTOR
For Each(=)(=)=

Info

PropertyBeforeAfter
HPXLocal Datetime2021-12-21T15:01:46+00:002022-01-14T05:48:10+00:00
HPXLocal Commit01e4980961f3ce
Datetime2021-12-21T16:09:15.238666+01:002022-01-14T06:54:29.496786+01:00
Hostnamenid00932nid01191
Envfile
Compiler/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0
Clusternamedaintdaint

Comparison

BENCHMARKNO-EXECUTOR
Future Overhead - Create Thread Hierarchical - Latch+

Info

PropertyBeforeAfter
HPXLocal Datetime2021-12-10T13:50:04+00:002022-01-14T05:48:10+00:00
HPXLocal Commitf499a22961f3ce
Datetime2021-12-10T15:19:42.442217+01:002022-01-14T06:54:46.122073+01:00
Hostnamenid00243nid01191
Envfile
Compiler/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0
Clusternamedaintdaint

Comparison

BENCHMARKFORK_JOIN_EXECUTORPARALLEL_EXECUTORSCHEDULER_EXECUTOR
Stream Benchmark - Add(=)(=)(=)
Stream Benchmark - Scale(=)-(=)
Stream Benchmark - Triad(=)(=)(=)
Stream Benchmark - Copy(=)(=)(=)

Info

PropertyBeforeAfter
HPXLocal Datetime2021-11-12T11:29:27+00:002022-01-14T05:48:10+00:00
HPXLocal Commitf64fbd0961f3ce
Datetime2021-11-12T12:57:50.824026+01:002022-01-14T06:55:00.655816+01:00
Hostnamenid00007nid01191
Envfile
Compiler/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0/apps/daint/SSL/HPX/packages/llvm-11.0.0/bin/clang++ 11.0.0
Clusternamedaintdaint

Explanation of Symbols

SymbolMEANING
=No performance change (confidence interval within ±1%)
(=)Probably no performance change (confidence interval within ±2%)
(+)/(-)Very small performance improvement/degradation (≤1%)
+/-Small performance improvement/degradation (≤5%)
++/--Large performance improvement/degradation (≤10%)
+++/---Very large performance improvement/degradation (>10%)
?Probably no change, but quite large uncertainty (confidence interval with ±5%)
??Unclear result, very large uncertainty (±10%)
???Something unexpected…

@msimberg msimberg mentioned this pull request Jan 17, 2022
Closed
@hkaiser hkaiser mentioned this pull request Jan 20, 2022
Merged
@hkaiser
Copy link
Member Author

hkaiser commented Jan 20, 2022

This is superseded by STEllAR-GROUP/hpx#5741

@hkaiser hkaiser closed this Jan 20, 2022
@hkaiser hkaiser deleted the hpx_5651 branch January 20, 2022 02:02
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@aurianer aurianer Awaiting requested review from aurianer aurianer is a code owner

@biddisco biddisco Awaiting requested review from biddisco biddisco is a code owner

@msimberg msimberg Awaiting requested review from msimberg msimberg is a code owner

Assignees
No one assigned
Labels
category: algorithms type: compatibility issue type: enhancement type: optimization
Projects
None yet
Milestone
0.1.0
Development

Successfully merging this pull request may close these issues.

None yet
4 participants
@hkaiser @StellarBot @msimberg @srinivasyadav18