[prev in list] [next in list] [prev in thread] [next in thread]
List: sbcl-devel
Subject: Re: [Sbcl-devel] [Sbcl-commits] master: Introduce more fine-grained integer simd pack types.
From: Charles Zhang via Sbcl-devel <sbcl-devel () lists ! sourceforge ! net>
Date: 2022-04-10 0:50:14
Message-ID: 1432632642.580188.1649551814042 () mail ! yahoo ! com
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
Fixed, by teaching the cross compiler new tricks. (It was as easy as you claimed on \
IRC, minus figuring out where the cars and cdrs are)
On Sat, Apr 9, 2022 at 4:54, Christophe Rhodes<csr21@cantab.net> wrote: apache--- \
via Sbcl-commits <sbcl-commits@lists.sourceforge.net> writes:
> The branch "master" has been updated in SBCL:
> via a5789ec9466c949b9312fec03d5b11548d84a37f (commit)
> from 91e588fd7d0f4976615795aca7c60ded34cbd996 (commit)
>
> - Log -----------------------------------------------------------------
> commit a5789ec9466c949b9312fec03d5b11548d84a37f
> Author: Marco Heisig <marco.heisig@fau.de>
> Date: Mon Apr 4 12:44:11 2022 +0200
>
> Introduce more fine-grained integer simd pack types.
>
> Split the type (simd-pack[-256] integer) into (simd-pack[-256] (A B)),
> where A is one of {signed-byte,unsigned-byte} and where B is one of
> {8,16,32,64}. The benefits of doing so are more precise static type
> checking, and a more helpful printed representation of integer SIMD packs.
>
> Additionally, introduce a simd-pack[-256]-any primitive type that can be
> used to implement VOPs whose implementation is independent of the actual
> SIMD pack type, such as %simd-pack-low or %simd-pack-256-0. This way, we
> can get rid of the clumsy sb-vm:simd-pack[-256]-dispatch macros that lead
> to unnecessary code bloat.
Something in here causes build failures on x86-64 from clisp and ccl
build hosts, when compiling x86-64/simd-pack.lisp -- probably something
to do with inlining? The failed averrance is in RECONSTRUCT-LEXENV.
Is the culprit obvious?
Christophe
_______________________________________________
Sbcl-devel mailing list
Sbcl-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel
[Attachment #5 (text/html)]
<div id="ymail_android_signature">Fixed, by teaching the cross compiler new tricks. \
(It was as easy as you claimed on IRC, minus figuring out where the cars and cdrs \
are)</div> <br> <blockquote style="margin: 0 0 20px 0;"> <div \
style="font-family:Roboto, sans-serif; color:#6D00F6;"> <div>On Sat, Apr 9, 2022 at \
4:54, Christophe Rhodes</div><div><csr21@cantab.net> wrote:</div> </div> <div \
style="padding: 10px 0 0 20px; margin: 10px 0 0 0; border-left: 1px solid #6D00F6;"> \
<div dir="ltr">apache--- via Sbcl-commits <<a \
ymailto="mailto:sbcl-commits@lists.sourceforge.net" \
href="mailto:sbcl-commits@lists.sourceforge.net">sbcl-commits@lists.sourceforge.net</a>> \
writes:<br></div><div dir="ltr"><br></div><div dir="ltr">> The branch "master" has \
been updated in SBCL:<br></div><div dir="ltr">> \
via a5789ec9466c949b9312fec03d5b11548d84a37f (commit)<br></div><div \
dir="ltr">> from \
91e588fd7d0f4976615795aca7c60ded34cbd996 (commit)<br></div><div \
dir="ltr">><br></div><div dir="ltr">> - Log \
-----------------------------------------------------------------<br></div><div \
dir="ltr">> commit a5789ec9466c949b9312fec03d5b11548d84a37f<br></div><div \
dir="ltr">> Author: Marco Heisig <<a ymailto="mailto:marco.heisig@fau.de" \
href="mailto:marco.heisig@fau.de">marco.heisig@fau.de</a>><br></div><div \
dir="ltr">> Date: Mon Apr 4 12:44:11 2022 +0200<br></div><div \
dir="ltr">><br></div><div dir="ltr">> Introduce more fine-grained \
integer simd pack types.<br></div><div dir="ltr">> <br></div><div \
dir="ltr">> Split the type (simd-pack[-256] integer) into \
(simd-pack[-256] (A B)),<br></div><div dir="ltr">> where A is one of \
{signed-byte,unsigned-byte} and where B is one of<br></div><div dir="ltr">> \
{8,16,32,64}. The benefits of doing so are more precise static \
type<br></div><div dir="ltr">> checking, and a more helpful printed \
representation of integer SIMD packs.<br></div><div dir="ltr">> \
<br></div><div dir="ltr">> Additionally, introduce a \
simd-pack[-256]-any primitive type that can be<br></div><div dir="ltr">> \
used to implement VOPs whose implementation is independent of the \
actual<br></div><div dir="ltr">> SIMD pack type, such as \
%simd-pack-low or %simd-pack-256-0. This way, we<br></div><div \
dir="ltr">> can get rid of the clumsy sb-vm:simd-pack[-256]-dispatch \
macros that lead<br></div><div dir="ltr">> to unnecessary code \
bloat.<br></div><div dir="ltr"><br></div><div dir="ltr">Something in here causes \
build failures on x86-64 from clisp and ccl<br></div><div dir="ltr">build hosts, when \
compiling x86-64/simd-pack.lisp -- probably something<br></div><div dir="ltr">to do \
with inlining? The failed averrance is in RECONSTRUCT-LEXENV.<br></div><div \
dir="ltr">Is the culprit obvious?<br></div><div dir="ltr"><br></div><div \
dir="ltr">Christophe<br></div><div dir="ltr"><br></div><div dir="ltr"><br></div><div \
dir="ltr">_______________________________________________<br></div><div \
dir="ltr">Sbcl-devel mailing list<br></div><div dir="ltr"><a \
ymailto="mailto:Sbcl-devel@lists.sourceforge.net" \
href="mailto:Sbcl-devel@lists.sourceforge.net">Sbcl-devel@lists.sourceforge.net</a><br></div><div \
dir="ltr"><a href="https://lists.sourceforge.net/lists/listinfo/sbcl-devel" \
target="_blank">https://lists.sourceforge.net/lists/listinfo/sbcl-devel</a><br></div> \
</div> </blockquote>
_______________________________________________
Sbcl-devel mailing list
Sbcl-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic