public inbox for [email protected]  
help / color / mirror / Atom feed
From: Nathan Bossart <[email protected]>
To: Amonson, Paul D <[email protected]>
Cc: Tom Lane <[email protected]>
Cc: David Rowley <[email protected]>
Cc: Andres Freund <[email protected]>
Cc: Alvaro Herrera <[email protected]>
Cc: Shankaran, Akash <[email protected]>
Cc: Noah Misch <[email protected]>
Cc: Matthias van de Meent <[email protected]>
Cc: [email protected] <[email protected]>
Subject: Re: Popcount optimization using AVX512
Date: Fri, 29 Mar 2024 10:35:15 -0500
Message-ID: <20240329153515.GA1046039@nathanxps13> (raw)
In-Reply-To: <BL1PR11MB5304E51336123CE6F041A920DC3B2@BL1PR11MB5304.namprd11.prod.outlook.com>
References: <BL1PR11MB530478ACABE588A836338E74DC322@BL1PR11MB5304.namprd11.prod.outlook.com>
	<BL1PR11MB53040D8E18518B2BDA4BB83ADC362@BL1PR11MB5304.namprd11.prod.outlook.com>
	<[email protected]>
	<BL1PR11MB53043014C10F2D30B4231F9ADC362@BL1PR11MB5304.namprd11.prod.outlook.com>
	<BL1PR11MB53049F1796E834DBE0E05864DC362@BL1PR11MB5304.namprd11.prod.outlook.com>
	<20240325200551.GB2924132@nathanxps13>
	<20240327220010.GA231058@nathanxps13>
	<BL1PR11MB53047289F146DB43C9B2FC71DC342@BL1PR11MB5304.namprd11.prod.outlook.com>
	<20240328213854.GA916732@nathanxps13>
	<BL1PR11MB5304E51336123CE6F041A920DC3B2@BL1PR11MB5304.namprd11.prod.outlook.com>

On Thu, Mar 28, 2024 at 10:03:04PM +0000, Amonson, Paul D wrote:
>> * I think we need to verify there isn't a huge performance regression for
>>   smaller arrays.  IIUC those will still require an AVX512 instruction or
>>   two as well as a function call, which might add some noticeable overhead.
> 
> Not considering your changes, I had already tested small buffers. At less
> than 512 bytes there was no measurable regression (there was one extra
> condition check) and for 512+ bytes it moved from no regression to some
> gains between 512 and 4096 bytes. Assuming you introduced no extra
> function calls, it should be the same.

Cool.  I think we should run the benchmarks again to be safe, though.

>> I forgot to mention that I also want to understand whether we can
>> actually assume availability of XGETBV when CPUID says we support
>> AVX512:
> 
> You cannot assume as there are edge cases where AVX-512 was found on
> system one during compile but it's not actually available in a kernel on
> a second system at runtime despite the CPU actually having the hardware
> feature.

Yeah, I understand that much, but I want to know how portable the XGETBV
instruction is.  Unless I can assume that all x86_64 systems and compilers
support that instruction, we might need an additional configure check
and/or CPUID check.  It looks like MSVC has had support for the _xgetbv
intrinsic for quite a while, but I'm still researching the other cases.

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






view thread (22+ messages)  latest in thread

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: [email protected]
  Cc: [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected]
  Subject: Re: Popcount optimization using AVX512
  In-Reply-To: <20240329153515.GA1046039@nathanxps13>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox