Re: Using POPCNT and other advanced bit manipulation instructions

From: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com>, Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk>, David Rowley <david(dot)rowley(at)2ndquadrant(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Using POPCNT and other advanced bit manipulation instructions
Date: 2019-02-14 18:47:13
Message-ID: 20190214184713.GA6080@alvherre.pgsql
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

On 2019-Feb-14, Tom Lane wrote:

> Some further thoughts here ...
>
> Does the "lzcnt" runtime probe actually do anything useful?
> On the x86_64 compilers I tried (gcc 8.2.1 and 4.4.7), __builtin_clz
> and __builtin_ctz compile to sequences involving bsrq and bsfq
> regardless of -mpopcnt. It's fairly hard to see how lzcnt would
> buy anything over those sequences even if there were zero overhead
> involved in using it.

Hah, I just realized you have to add -mlzcnt in order for these builtins
to use the lzcnt instructions. It goes from something like

bsrq %rax, %rax
xorq $63, %rax

to
lzcntq %rax, %rax

Significant?

I have this patch now, written before I realized the above; I'll augment
it to cater to this (adding -mlzcnt and a new set of functions --
perhaps a new file "lzcnt.c" or maybe put the lot in pg_popcount.c and
rename it?) and resubmit after an errand I have to run now.

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Attachment Content-Type Size
popcount.patch text/x-diff 11.9 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2019-02-14 18:54:32 Re: Using POPCNT and other advanced bit manipulation instructions
Previous Message Andres Freund 2019-02-14 18:44:02 Re: \describe*