Implement remove and remove_if algorithms for datapar execution by BhoomishGupta · Pull Request #6970 · TheHPXProject/hpx

BhoomishGupta · 2026-03-06T23:14:35Z

Partially Fixes #2333

Proposed Changes

-Added datapar/remove.hpp implementing SIMD-accelerated sequential_remove_if and sequential_remove via tag_invoke overloads for vectorpack execution policies
-Added detail/remove.hpp with sequential_remove_if_t and sequential_remove_t tag-fallback CPOs used by both scalar and datapar paths
-Added datapar unit tests remove_datapar.cpp and remove_if_datapar.cpp covering simd, par_simd, simd(task), and par_simd(task) policies with both random-access and forward iterators; registered both in CMakeLists.txt

Any background context you want to provide?

Checklist

Not all points below apply to all pull requests.

I have added a new feature and have added tests to go along with it.
I have fixed a bug and have added a regression test.
I have added a test using random numbers; I have made sure it uses a seed, and that random numbers generated are valid inputs for the tests.

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

StellarBot · 2026-03-06T23:15:03Z

Can one of the admins verify this patch?

hkaiser · 2026-03-07T16:06:11Z

@BhoomishGupta From what I can see, your new implementations do not take advantage of data parallelisation but simply do a per-element processing once the first matching element is found. Do you see a way of taking advantage of data parallelism for the remainder of the sequence as well?

BhoomishGupta · 2026-03-10T01:01:57Z

@BhoomishGupta From what I can see, your new implementations do not take advantage of data parallelisation but simply do a per-element processing once the first matching element is found. Do you see a way of taking advantage of data parallelism for the remainder of the sequence as well?

Thanks for the feedback! I will look into it.

Just as a quick reminder, my university exams are still going on right now, so my availability is quite limited. I will dive into this and work on an improved approach right after my exams finish on March 14th.

Thanks for your patience!

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

hkaiser · 2026-03-21T15:49:38Z

libs/core/executors/include/hpx/executors/execution_policy_mappings.hpp

+                "must not be a simd (vectorpack) execution policy");
+            return std::forward<ExPolicy>(policy);
+        }
+    } to_non_simd{};


These mapping CPOs are already defined here: https://github.com/STEllAR-GROUP/hpx/blob/master/libs/core/executors/include/hpx/executors/datapar/execution_policy_mappings.hpp

I moved these mapping CPOs from https://github.com/STEllAR-GROUP/hpx/blob/master/libs/core/executors/include/hpx/executors/datapar/execution_policy_mappings.hpp to libs/core/executors/include/hpx/executors/execution_policy_mappings.hpp

These CPOs were originally defined in the datapar header, but I moved them to the common execution policy mappings header so they are available even when datapar/SIMD is not enabled. This is needed for code paths that reference the mappings independently of datapar support.
The datapar header now just includes the common header (guarded by HPX_HAVE_DATAPAR) for compatibility.

Why would this be needed if datapar is not enabled?

Yeah, I reviewed this. And you are right.
I originally fell back to the non_simd policy (and consequently had to move the execution_policy_mappings.hpp header to make it globally visible) because the f1 loop was failing to compile under the SIMD backends.

However, doing this forced the parallel chunks to evaluate the predicate sequentially, which indeed defeats the purpose of the SIMD implementation.

The root cause of the original compilation failure was that the flags array was declared as bool[]. In the std::simd backends, boolean vector arrays instantiate as proxy mask objects rather than standard packed primitives, which broke the vector_pack traits when it tried to iterate and assign. Also, the f1 lambda was explicitly taking (zip_iterator it) instead of (auto it), which prevented the datapar loop_n from passing in the vectorized chunks.

To resolve this properly, I am thinking of doing this:

I will change the flags array allocation from bool to std::uint8_t so it reliably packs.

I will change the f1 lambda parameter to auto it and revert its loop policy back to the native std::decay_t so that the heavy predicate evaluation actually vectorizes.

I will completely revert the execution_policy_mappings.hpp move, as to_non_simd will no longer be leaking unconditionally into the generic algorithm.

hkaiser · 2026-03-21T15:50:59Z

libs/core/algorithms/include/hpx/parallel/algorithms/remove.hpp

                              std::size_t part_size) -> void {
                    // MSVC complains if pred or proj is captured by ref below
-                    util::loop_n<std::decay_t<ExPolicy>>(part_begin, part_size,
+                    util::loop_n<inner_policy_type>(part_begin, part_size,


Wouldn't invoking the loop with a non-simd policy defeat the whole purpose of introducing SIMD functionalities to remove_if?

You're absolutely right. Invoking the loop with inner_policy_type (non-SIMD) was a workaround because of a compilation error with the SIMD backend. I've since identified that the root cause was the bool[] flags array and the explicit zip_iterator parameter in the lambda. I'm reverting this to use the original execution policy to ensure the predicate evaluation is properly vectorized, as detailed in my other comment

libs/core/algorithms/include/hpx/parallel/datapar/remove.hpp

hkaiser · 2026-03-21T15:58:49Z

libs/core/algorithms/include/hpx/parallel/datapar/remove.hpp

+                else if (!hpx::parallel::traits::all_of(msk))
+                {
+                    //mixed
+                    for (std::size_t i = 0; i < size; ++i)


For std::simd (std::experimental::simd) you could utilize https://en.cppreference.com/w/cpp/experimental/simd/find_first_set.html to possibly reduce the amount of iterations needed.

For the other datapar backends (EVE, VC, arm simd) you may have to emulate this or simply always return 0 and simd_width respectively)

Thanks for the heads up! It looks like HPX already has a generic implementation for this across all backends via hpx::parallel::traits::find_first_of(msk) in libs/core/execution/include/hpx/execution/traits/vector_pack_find.hpp

I'll update the loop to use it.

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

codacy-production · 2026-04-01T00:12:27Z

Not up to standards ⛔

🔴 Issues 1 medium

Alerts:
⚠ 1 issue (≤ 0 issues of at least minor severity)

Results:
1 new issue

Category Results

ErrorProne 1 medium

View in Codacy

🟢 Metrics 0 complexity · 0 duplication

Metric Results

Complexity 0

Duplication 0

View in Codacy

_{TIP This summary will be updated as you push new changes. Give us feedback}

hkaiser · 2026-04-01T01:19:26Z

@BhoomishGupta Would you be able to follow up on the compilation errors reported by the CIs?

BhoomishGupta · 2026-04-01T22:40:50Z

@BhoomishGupta Would you be able to follow up on the compilation errors reported by the CIs?

Working on it!

…xecution Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

hkaiser · 2026-04-02T21:54:23Z

@BhoomishGupta Please pay attention to the CIs. ARM is failing to compile. Also, the remove test is now failing to compile even on CIs that do not enable SIMD functionalities.

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta · 2026-04-03T08:44:23Z

@BhoomishGupta Please pay attention to the CIs. ARM is failing to compile. Also, the remove test is now failing to compile even on CIs that do not enable SIMD functionalities.

My mistake! I missed an include. Just pushed a fix. Watching the CIs now

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

hkaiser · 2026-04-05T14:10:45Z

libs/core/algorithms/include/hpx/parallel/algorithms/remove.hpp


 #include <hpx/config.hpp>
+#include <hpx/algorithms/traits/projected.hpp>
+#include <hpx/execution/traits/vector_pack_conditionals.hpp>


Please alsways use only the generted module includes (except inside a module itself):

Suggested change

#include <hpx/execution/traits/vector_pack_conditionals.hpp>

#include <hpx/modules/execution.hpp>

Sure, I will keep to generated module includes in non-module code paths from now on.

hkaiser · 2026-04-05T14:14:03Z

libs/core/algorithms/include/hpx/parallel/algorithms/remove.hpp

+                    typename std::iterator_traits<Iter>::value_type>;
+                using flag_t =
+                    std::conditional_t<use_vectorpack_predicate, value_t, bool>;
+                using zip_iterator = hpx::util::zip_iterator<Iter, flag_t*>;


Please refactor the code similarily to what we have done for he other algorithms. The code related to datapar should be kept fully separate from the non-datapar code paths. If datapar is disabled, e.g., the is_vectorpack_execution_policy you use here are not available.

Agreed. I will refactor the code such that datapar-specific assumptions are out of generic algorithm paths so non-datapar builds remain valid.

hkaiser · 2026-04-05T14:14:51Z

libs/core/algorithms/include/hpx/parallel/algorithms/remove.hpp

                    {
                        // Self-assignment must be detected.
-                        util::loop_n<execution_policy_type>(
+                        util::loop_n<hpx::execution::sequenced_policy>(


Wouldn't this lose information? Why is this change needed?

This change does not lose algorithm/data information.
That loop has a loop-carried dependency on dest plus self-assignment checks, so running it with vectorpack/unsequenced execution is not safe without a prefix-sum style compaction step. Using sequenced_policy here preserves correctness and ordering; the SIMD work remains in the predicate-evaluation phase f1. So the change drops vectorization in f2.

libs/core/algorithms/include/hpx/parallel/algorithms/remove.hpp

hkaiser · 2026-04-05T14:17:05Z

libs/core/algorithms/include/hpx/parallel/container_algorithms/remove.hpp

+                    hpx::is_invocable_v<Pred,
+                        typename std::iterator_traits<FwdIter>::value_type
+                    >
+                )


Can you please explain the rationale of this change?

The rationale is to keep the range execution-policy overload aligned with current dispatch behavior in remove_if.
If we required only the policy-aware projected callable check, scalar-only predicates would be rejected at the CPO boundary, even though the implementation can still run them through the scalar predicate-evaluation path when vectorized invocation is not available.
Adding the scalar invocable check preserves backward compatibility and keeps iterator/range policy-overload behavior consistent.
This changes only whether predicate evaluation can be vectorized for a given call; it does not change remove_if semantics or output correctness.

If you prefer stricter behavior for par_simd and related policies, I can tighten this to require only the policy-aware projected callable check and fail early otherwise.

libs/core/execution/include/hpx/execution/traits/detail/eve/vector_pack_load_store.hpp

hkaiser · 2026-04-08T13:36:52Z

@BhoomishGupta ping?

BhoomishGupta · 2026-04-08T15:10:08Z

@BhoomishGupta ping?

Sorry for the delay. Resolving the comments asap

Implement remove and remove_if algorithms for datapar execution

9635098

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta requested a review from hkaiser as a code owner March 6, 2026 23:14

Merge branch 'master' into remove_datapar

2c68127

hkaiser added the category: algorithms label Mar 7, 2026

BhoomishGupta and others added 2 commits March 20, 2026 08:24

Data Parallelism in datapar_remove_if implementation

68820ef

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

Merge branch 'master' into remove_datapar

8276327

hkaiser reviewed Mar 21, 2026

View reviewed changes

BhoomishGupta force-pushed the remove_datapar branch from e1e93a0 to 152d3a2 Compare March 31, 2026 04:23

improve simd remove algorithm implementation

6fbb976

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta force-pushed the remove_datapar branch from 152d3a2 to 6fbb976 Compare March 31, 2026 05:12

BhoomishGupta added 2 commits March 31, 2026 11:17

Add missing type_traits include

0e02c99

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

Merge remote-tracking branch 'upstream/master' into remove_datapar

8bc4e40

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta force-pushed the remove_datapar branch from cb1a147 to 8bc4e40 Compare March 31, 2026 05:47

Added missing scope declarations and formatting

81a92c4

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

Enhance predicate callable checks for remove algorithms in parallel e…

a202414

…xecution Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

Included a module and changed variable name

720597c

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta and others added 2 commits April 3, 2026 15:00

Merge branch 'master' into remove_datapar

91fc510

Remove conflicting duplicate sequential_remove_if definition

1ac308a

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta force-pushed the remove_datapar branch from f7d0dd4 to 1ac308a Compare April 3, 2026 10:08

Fix vector_pack_store::unaligned for EVE backend

87c8b42

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta force-pushed the remove_datapar branch 2 times, most recently from 7bd012e to 7d1fae7 Compare April 3, 2026 12:52

Fix EVE SIMD backend traits and unaligned store fallbacks

0968a1e

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta force-pushed the remove_datapar branch from 7d1fae7 to 0968a1e Compare April 3, 2026 15:53

Improved mask handling

f01732b

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

BhoomishGupta force-pushed the remove_datapar branch from fea749e to f01732b Compare April 4, 2026 11:04

removed if defined portion

f6ae30b

Signed-off-by: BhoomishGupta <bhoomishguptabits@gmail.com>

hkaiser reviewed Apr 5, 2026

View reviewed changes

	#include <hpx/execution/traits/vector_pack_conditionals.hpp>
	#include <hpx/modules/execution.hpp>

Uh oh!

Conversation

BhoomishGupta commented Mar 6, 2026

Proposed Changes

Any background context you want to provide?

Checklist

Uh oh!

StellarBot commented Mar 6, 2026

Uh oh!

hkaiser commented Mar 7, 2026

Uh oh!

BhoomishGupta commented Mar 10, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

BhoomishGupta Mar 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

BhoomishGupta Mar 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codacy-production bot commented Apr 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Not up to standards ⛔

Uh oh!

hkaiser commented Apr 1, 2026

Uh oh!

BhoomishGupta commented Apr 1, 2026

Uh oh!

hkaiser commented Apr 2, 2026

Uh oh!

BhoomishGupta commented Apr 3, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

hkaiser commented Apr 8, 2026

Uh oh!

BhoomishGupta commented Apr 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

BhoomishGupta Mar 24, 2026 •

edited

Loading

BhoomishGupta Mar 30, 2026 •

edited

Loading

codacy-production bot commented Apr 1, 2026 •

edited

Loading