From: Bill Schmidt <wschmidt@linux.ibm.com>
To: Segher Boessenkool <segher@kernel.crashing.org>
Cc: gcc-patches@gcc.gnu.org, dje.gcc@gmail.com
Subject: Re: [PATCH 00/57] Replace the Power target-specific built-in machinery
Date: Fri, 21 May 2021 07:53:14 -0500 [thread overview]
Message-ID: <388f72f5-2ea5-cf24-facc-1fa83f8ec508@linux.ibm.com> (raw)
In-Reply-To: <20210520215742.GU10366@gate.crashing.org>
On 5/20/21 4:57 PM, Segher Boessenkool wrote:
> Is there some (semi-)automatic way to compare the results of the old
> and new systems?
Yes...very "semi". There's a patch in the series that updates the
information printed from -mdebug=builtin. I use this to print the
builtins generated by the old and new versions, and run them through a
Python script to look for anything missing or added, any mismatches in
the type system, mismatches in const/pure/etc., that sort of thing. I
found quite a lot of mistakes this way and got them fixed ahead of
time. In the script I maintain a table of expected differences, where
there were just errors before, etc., and there's a short doc string with
each of them. So I can go back and look at that if we run into any
discrepancies in the future. The script doesn't have long-term value
beyond that; once we delete the old version in a future patch, it won't
have anything to compare against.
>> Patch 0057: Fix one last late-breaking change
>>
>> Keeping the code up-to-date with upstream has been fun. When I
>> rebased to create the patch set, I found one new issue where a
>> small change had been made to the overload handling for the
>> vec_insert builtins. This patch reflects that change into the
>> new handling. My version of git is having trouble with
>> interactive rebasing, so it was easier to just add the extra patch.
> What breaks by keeping this fix after the other patches?
Nothing truly breaks. There is a VSX-efficient implementation of
vec_insert and a less-efficient implementation. Late in GCC 11,
somebody recognized the VSX-efficient implementation could be used in
more situations. Without this patch, we still use the less-efficient
implementation in the new version.
>
>> (3) A number of built-ins used "long" for DImode types, which would
>> break these for 32-bit. I changed those arguments and return values
>> to "long long" to avoid such problems, when those built-ins were not
>> restricted to 64-bit mode already. There aren't many such cases.
> You can do this for 64-bit-only builtins as well -- the actual argument
> type is never visible (to the user), and everything becomes modes early.
>
Yep -- in fact, the old system generally used a (V2)DImode type for such
things, and I've just translated all of those to (vector) long long in
the new prototypes.
Thanks for the ongoing review!
Bill
prev parent reply other threads:[~2021-05-21 12:53 UTC|newest]
Thread overview: 99+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-27 15:32 Bill Schmidt
2021-04-27 15:32 ` [PATCH 01/57] Allow targets to specify build dependencies for out_object_file Bill Schmidt
2021-04-27 15:57 ` Jakub Jelinek
2021-04-27 16:14 ` Bill Schmidt
2021-04-27 16:47 ` Jakub Jelinek
2021-04-27 17:44 ` Bill Schmidt
2021-04-27 15:32 ` [PATCH 02/57] Support scanning of build-time GC roots in gengtype Bill Schmidt
2021-05-11 16:01 ` Bill Schmidt
2021-05-20 22:24 ` Segher Boessenkool
2021-06-04 19:03 ` Bill Schmidt
2021-06-07 10:39 ` Richard Sandiford
2021-06-07 12:35 ` Bill Schmidt
2021-06-07 13:36 ` Richard Biener
2021-06-07 15:38 ` Bill Schmidt
2021-06-07 17:45 ` Richard Biener
2021-06-07 17:48 ` Bill Schmidt
2021-06-08 20:45 ` Bill Schmidt
2021-06-09 10:53 ` Richard Biener
2021-06-09 10:54 ` Richard Biener
2021-06-09 12:53 ` Bill Schmidt
2021-05-20 22:19 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 03/57] rs6000: Initial create of rs6000-gen-builtins.c Bill Schmidt
2021-05-20 22:32 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 04/57] rs6000: Add initial input files Bill Schmidt
2021-05-20 22:46 ` Segher Boessenkool
2021-05-21 12:58 ` Bill Schmidt
2021-04-27 15:32 ` [PATCH 05/57] rs6000: Add file support and functions for diagnostic support Bill Schmidt
2021-05-20 23:03 ` Segher Boessenkool
2021-05-21 13:06 ` Bill Schmidt
2021-04-27 15:32 ` [PATCH 06/57] rs6000: Add helper functions for parsing Bill Schmidt
2021-05-21 18:51 ` Segher Boessenkool
2021-05-21 20:56 ` Bill Schmidt
2021-05-21 23:43 ` Segher Boessenkool
2021-06-01 15:50 ` Bill Schmidt
2021-05-23 22:37 ` Bernhard Reutner-Fischer
2021-05-24 21:35 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 07/57] rs6000: Add functions for matching types, part 1 of 3 Bill Schmidt
2021-05-21 20:50 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 08/57] rs6000: Add functions for matching types, part 2 " Bill Schmidt
2021-05-21 21:36 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 09/57] rs6000: Add functions for matching types, part 3 " Bill Schmidt
2021-05-21 21:46 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 10/57] rs6000: Red-black tree implementation for balanced tree search Bill Schmidt
2021-05-21 22:29 ` Segher Boessenkool
2021-04-27 15:32 ` [PATCH 11/57] rs6000: Main function with stubs for parsing and output Bill Schmidt
2021-04-27 15:32 ` [PATCH 12/57] rs6000: Parsing built-in input file, part 1 of 3 Bill Schmidt
2021-04-27 15:32 ` [PATCH 13/57] rs6000: Parsing built-in input file, part 2 " Bill Schmidt
2021-04-27 15:32 ` [PATCH 14/57] rs6000: Parsing built-in input file, part 3 " Bill Schmidt
2021-04-27 15:32 ` [PATCH 15/57] rs6000: Parsing of overload input file Bill Schmidt
2021-04-27 15:32 ` [PATCH 16/57] rs6000: Build and store function type identifiers Bill Schmidt
2021-04-27 15:32 ` [PATCH 17/57] rs6000: Write output to the builtin definition include file Bill Schmidt
2021-04-27 15:32 ` [PATCH 18/57] rs6000: Write output to the builtins header file Bill Schmidt
2021-04-27 15:32 ` [PATCH 19/57] rs6000: Write output to the builtins init file, part 1 of 3 Bill Schmidt
2021-04-27 15:32 ` [PATCH 20/57] rs6000: Write output to the builtins init file, part 2 " Bill Schmidt
2021-04-27 15:32 ` [PATCH 21/57] rs6000: Write output to the builtins init file, part 3 " Bill Schmidt
2021-04-27 15:32 ` [PATCH 22/57] rs6000: Write static initializations for built-in table Bill Schmidt
2021-04-27 15:32 ` [PATCH 23/57] rs6000: Write static initializations for overload tables Bill Schmidt
2021-04-27 15:32 ` [PATCH 24/57] rs6000: Incorporate new builtins code into the build machinery Bill Schmidt
2021-04-27 15:33 ` [PATCH 25/57] rs6000: Add gengtype handling to " Bill Schmidt
2021-04-27 15:33 ` [PATCH 26/57] rs6000: Add the rest of the [altivec] stanza to the builtins file Bill Schmidt
2021-04-27 15:33 ` [PATCH 27/57] rs6000: Add VSX builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 28/57] rs6000: Add available-everywhere and ancient builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 29/57] rs6000: Add power7 and power7-64 builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 30/57] rs6000: Add power8-vector builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 31/57] rs6000: Add Power9 builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 32/57] rs6000: Add more type nodes to support builtin processing Bill Schmidt
2021-04-27 15:33 ` [PATCH 33/57] rs6000: Add Power10 builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 34/57] rs6000: Add MMA builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 35/57] rs6000: Add miscellaneous builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 36/57] rs6000: Add Cell builtins Bill Schmidt
2021-04-27 15:33 ` [PATCH 37/57] rs6000: Add remaining overloads Bill Schmidt
2021-04-27 15:33 ` [PATCH 38/57] rs6000: Execute the automatic built-in initialization code Bill Schmidt
2021-04-27 15:33 ` [PATCH 39/57] rs6000: Darwin builtin support Bill Schmidt
2021-04-30 20:05 ` Iain Sandoe
2021-04-27 15:33 ` [PATCH 40/57] rs6000: Add sanity to V2DI_type_node definitions Bill Schmidt
2021-04-27 15:33 ` [PATCH 41/57] rs6000: Always initialize vector_pair and vector_quad nodes Bill Schmidt
2021-04-27 15:33 ` [PATCH 42/57] rs6000: Handle overloads during program parsing Bill Schmidt
2021-04-27 15:33 ` [PATCH 43/57] rs6000: Handle gimple folding of target built-ins Bill Schmidt
2021-04-27 15:33 ` [PATCH 44/57] rs6000: Support for vectorizing built-in functions Bill Schmidt
2021-04-27 15:33 ` [PATCH 45/57] rs6000: Builtin expansion, part 1 Bill Schmidt
2021-04-27 15:33 ` [PATCH 46/57] rs6000: Builtin expansion, part 2 Bill Schmidt
2021-04-27 15:33 ` [PATCH 47/57] rs6000: Builtin expansion, part 3 Bill Schmidt
2021-04-27 15:33 ` [PATCH 48/57] rs6000: Builtin expansion, part 4 Bill Schmidt
2021-04-27 15:33 ` [PATCH 49/57] rs6000: Builtin expansion, part 5 Bill Schmidt
2021-04-27 15:33 ` [PATCH 50/57] rs6000: Builtin expansion, part 6 Bill Schmidt
2021-04-27 15:33 ` [PATCH 51/57] rs6000: Update rs6000_builtin_decl Bill Schmidt
2021-04-27 15:33 ` [PATCH 52/57] rs6000: Miscellaneous uses of rs6000_builtin_decls_x Bill Schmidt
2021-04-27 15:33 ` [PATCH 53/57] rs6000: Debug support Bill Schmidt
2021-04-27 15:33 ` [PATCH 54/57] rs6000: Update altivec.h for automated interfaces Bill Schmidt
2021-04-27 15:33 ` [PATCH 55/57] rs6000: Test case adjustments Bill Schmidt
2021-04-27 15:33 ` [PATCH 56/57] rs6000: Enable the new builtin support Bill Schmidt
2021-04-27 15:33 ` [PATCH 57/57] rs6000: Adjust to late-breaking change Bill Schmidt
2021-04-30 12:38 ` [PATCH "58/57"] rs6000: Avoid problems with undefined decimal float types Bill Schmidt
2021-04-30 12:42 ` [PATCH "59/57"] rs6000: Fix builtins that should have been available everywhere Bill Schmidt
2021-04-30 18:55 ` [PATCH "60/57"] rs6000: Fix AltiVec builtin marked as VSX Bill Schmidt
2021-05-11 15:57 ` [PATCH 00/57] Replace the Power target-specific built-in machinery Bill Schmidt
2021-05-11 23:20 ` Segher Boessenkool
2021-05-20 21:57 ` Segher Boessenkool
2021-05-21 12:53 ` Bill Schmidt [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=388f72f5-2ea5-cf24-facc-1fa83f8ec508@linux.ibm.com \
--to=wschmidt@linux.ibm.com \
--cc=dje.gcc@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=segher@kernel.crashing.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).