Horton

Re: [PATCH] SVE popcount support

John Naylor <johncnaylorls@gmail.com>

View thread
From: John Naylor <johncnaylorls@gmail.com>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: "Chiranmoy.Bhattacharya@fujitsu.com" <Chiranmoy.Bhattacharya@fujitsu.com>, "Malladi, Rama" <ramamalladi@hotmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, "Ragesh.Hajela@fujitsu.com" <Ragesh.Hajela@fujitsu.com>, Salvatore Dipietro <dipiets@amazon.com>, "Devanga.Susmitha@fujitsu.com" <Devanga.Susmitha@fujitsu.com>
Date: 2025-03-27T08:31:27Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Optimize popcount functions with ARM SVE intrinsics.

  2. Optimize popcount functions with ARM Neon intrinsics.

  3. Rename TRY_POPCNT_FAST to TRY_POPCNT_X86_64.

On Thu, Mar 27, 2025 at 10:38 AM Nathan Bossart
<nathandbossart@gmail.com> wrote:
> I also noticed a silly mistake in 0003 that would cause us to potentially
> skip part of the tail.  That should be fixed now.

I'm not sure whether that meant it could return the wrong answer, or
just make more work for paths further down.
If the former, then our test coverage is not adequate.

Aside from that, I only found one more thing that may be important: I
tried copying the configure/meson checks into godbolt.org, and both
gcc and clang don't like it, so unless there is something weird about
their setup (or my use of it) it's possible some other hosts won't
like it either.:

```
<source>:29:10: error: call to 'svwhilelt_b8' is ambiguous
                pred = svwhilelt_b8(0, sizeof(buf));
                       ^~~~~~~~~~~~
/opt/compiler-explorer/clang-16.0.0/lib/clang/16/include/arm_sve.h:15526:10:
note: candidate function
svbool_t svwhilelt_b8(uint64_t, uint64_t);
         ^
/opt/compiler-explorer/clang-16.0.0/lib/clang/16/include/arm_sve.h:15534:10:
note: candidate function
svbool_t svwhilelt_b8(int32_t, int32_t);
         ^

<source>: In function 'autoconf_popcount_test':
<source>:29:24: error: call to 'svwhilelt_b8' is ambiguous; argument 1
has type 'int32_t' but argument 2 has type 'uint64_t'
   29 |                 pred = svwhilelt_b8(0, sizeof(buf));
      |                        ^~~~~~~~~~~~
Compiler returned: 1
```

...Changing it to

pred = svwhilelt_b8((uint64_t)0, sizeof(buf));"

clears it up.

-- 
John Naylor
Amazon Web Services