Re: Popcount optimization using AVX512

Nathan Bossart <nathandbossart@gmail.com>

From: Nathan Bossart <nathandbossart@gmail.com>
To: "Amonson, Paul D" <paul.d.amonson@intel.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, David Rowley <dgrowleyml@gmail.com>, Andres Freund <andres@anarazel.de>, Alvaro Herrera <alvherre@alvh.no-ip.org>, "Shankaran, Akash" <akash.shankaran@intel.com>, Noah Misch <noah@leadboat.com>, Matthias van de Meent <boekewurm+postgres@gmail.com>, "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>
Date: 2024-03-29T16:22:11Z
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. Fix __attribute__((target(...))) usage.

  2. Use __attribute__((target(...))) for AVX-512 support.

  3. Fix code for probing availability of AVX-512.

  4. Optimize visibilitymap_count() with AVX-512 instructions.

  5. Optimize pg_popcount() with AVX-512 instructions.

  6. Inline pg_popcount() for small buffers.

  7. Avoid function call overhead of pg_popcount() in syslogger.c.

  8. Refactor code for setting pg_popcount* function pointers.

  9. Inline pg_popcount{32,64} into pg_popcount().

  10. Remove MSVC scripts

  11. Use ARMv8 CRC instructions where available.

  12. Use Intel SSE 4.2 CRC instructions where available.

On Fri, Mar 29, 2024 at 10:59:40AM -0500, Nathan Bossart wrote:
> It might be nice if we conditionally built pg_popcount_avx512.o in autoconf
> builds, too, but AFAICT we still need to wrap most of that code with
> macros, so I'm not sure it's worth the trouble.  I'll take another look at
> this...

If we assumed that TRY_POPCNT_FAST would be set and either
HAVE__GET_CPUID_COUNT or HAVE__CPUIDEX would be set whenever
USE_AVX512_POPCNT_WITH_RUNTIME_CHECK is set, we could probably remove the
surrounding macros and just compile pg_popcount_avx512.c conditionally
based on USE_AVX512_POPCNT_WITH_RUNTIME_CHECK.  However, the surrounding
code seems to be pretty cautious about these assumptions (e.g., the CPUID
macros are checked before setting TRY_POPCNT_FAST), so this would stray
from the nearby precedent a bit.

A counterexample is the CRC32C code.  AFAICT we assume the presence of
CPUID in that code (and #error otherwise).  I imagine its probably safe to
assume the compiler understands CPUID if it understands AVX512 intrinsics,
but that is still mostly a guess.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com