ARM64-SVE: Additions with GetActiveElementCount() should use incp

Open
#108,400 4 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
3/5
Estimated time
1-2 days
Newbie friendliness
38/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Stale
Tech stack
csharp
Domain
compilers

Research direction

Search the ARM64-SVE compiler lowering for Sve.GetActiveElementCount and the generated cntp+add pattern shown in the issue. Compare the result with the requested incp instruction, account for the related predicate-conversion work in #108241, and verify the generated assembly matches the example.

Written by the indexing model from the issue text.

Description

arch-arm64 area-CodeGen-coreclr arm-sve Priority:2 
      i += Sve.GetActiveElementCount(pred, pred);

becomes:

            ptrue   p2.b
            cmpne   p1.b, p2/z, z16.b, #0
            ptrue   p0.b
            cmpne   p0.b, p0/z, z16.b, #0
            cntp    x0, p1, p0.b
            add     x0, x0, x1

Firstly, the two cntp should use a the same predicate instead of doing two conversions! That is covered already under #108241.

Secondly, the cntp+add should be optimized to a single incp:

            ptrue   p2.b
            cmpne   p1.b, p2/z, z16.b, #0
            incp    x0, p1
Dominant language
C#
Stars
18.3k
Forks
5.6k
PR metrics pending

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from dotnet/runtime

All issues in dotnet/runtime

Similar issues

More C# issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.