From: Brian Inglis <Brian.Inglis@Shaw.ca>
To: cygwin-patches@cygwin.com
Subject: Re: [PATCH] Cygwin: Make gcc-specific code in <sys/cpuset.h> compiler-agnostic
Date: Sat, 8 Jul 2023 13:22:41 -0600 [thread overview]
Message-ID: <1f7d3254-234e-378f-a852-63ca5d7ca01f@Shaw.ca> (raw)
In-Reply-To: <073cd700-c727-ee29-017e-df8d86a1db59@Shaw.ca>
On 2023-07-07 12:54, Brian Inglis wrote:
> On 2023-07-07 03:44, Corinna Vinschen wrote:
>> Hi Mark,
>>
>> On Jul 7 00:41, Mark Geisert wrote:
>>> The current version of <sys/cpuset.h> cannot be compiled by Clang due to
>>> the use of __builtin* functions. Their presence here was a dubious
>>> optimization anyway, so their usage has been converted to standard
>>> library functions. A popcnt (population count of 1 bits in a word)
>>> function is provided here because there isn't one in the standard library
>>> or elsewhere in the Cygwin DLL.
>>
>> And clang really doesn't provide it? That's unfortunate.
>>
>> Do you really think it's not worth to use it if it's available?
>>
>> You could workaround it like this:
>>
>>> +/* Modern CPUs have popcnt* instructions but the need here is not worth
>>> + * worrying about builtins or inline assembler for different compilers. */
>>> +static inline int
>>> +__maskpopcnt (__cpu_mask mask)
>>> +{
>> #if (__GNUC__ >= 4)
>> return __builtin_popcountl (mask);
Missed the difference in spelling, but clang supports the same builtin functions
__builtin_popcount{,l,ll} et. al. or provides *optimized* inline functions if
not directly available as an instruction on the architecture.
>> #else
>>> + int res = 0;
>>> + unsigned long ulmask = (unsigned long) mask;
>>> +
>>> + while (ulmask != 0)
>>> + {
>>> + if (ulmask & 1)
>>> + ++res;
>>> + ulmask >>= 1;
>>> + }
>>> + return res;
>> #endif
>>> +}
>>> +
>>
>> But, if you think that's not worth it, I'll push your patch as is.
>
> In the cygwin list clang iostream discussion, I pointed out the clang
> configuration using gcc default paths and skipping the provided equivalent clang
> libraries including package libclang8 for clang intrinsics and builtins in:
>
> /usr/lib/clang/8.0.1/include/
>
> which appear very similar to those provided by gcc-core in:
>
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/
>
> e.g.
>
> $ l /usr/lib/{gcc/x86_64-pc-cygwin/11,clang/8.0.1}/include/*popcnt*
> /usr/lib/clang/8.0.1/include/avx512vpopcntdqintrin.h
> /usr/lib/clang/8.0.1/include/avx512vpopcntdqvlintrin.h
> /usr/lib/clang/8.0.1/include/popcntintrin.h
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/avx512vpopcntdqintrin.h
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/avx512vpopcntdqvlintrin.h
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/popcntintrin.h
>
> and from comparing features, the gcc test will work just fine with the clang
> compiler builtins or the clang intrinsics:
>
> $ grep 'popc\(ou\)\?nt'
> /usr/lib/{gcc/x86_64-pc-cygwin/11,clang/8.0.1}/include/popcnt*
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/popcntintrin.h:#pragma GCC
> target("popcnt")
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/popcntintrin.h:_mm_popcnt_u32 (unsigned
> int __X)
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/popcntintrin.h: return
> __builtin_popcount (__X);
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/popcntintrin.h:_mm_popcnt_u64 (unsigned
> long long __X)
> /usr/lib/gcc/x86_64-pc-cygwin/11/include/popcntintrin.h: return
> __builtin_popcountll (__X);
> /usr/lib/clang/8.0.1/include/popcntintrin.h:/*===---- popcntintrin.h - POPCNT
> intrinsics -------------------------------===
> /usr/lib/clang/8.0.1/include/popcntintrin.h:#define __DEFAULT_FN_ATTRS
> __attribute__((__always_inline__, __nodebug__, __target__("popcnt")))
> /usr/lib/clang/8.0.1/include/popcntintrin.h:_mm_popcnt_u32(unsigned int __A)
> /usr/lib/clang/8.0.1/include/popcntintrin.h: return __builtin_popcount(__A);
> /usr/lib/clang/8.0.1/include/popcntintrin.h:_popcnt32(int __A)
> /usr/lib/clang/8.0.1/include/popcntintrin.h: return __builtin_popcount(__A);
> /usr/lib/clang/8.0.1/include/popcntintrin.h:_mm_popcnt_u64(unsigned long long __A)
> /usr/lib/clang/8.0.1/include/popcntintrin.h: return __builtin_popcountll(__A);
> /usr/lib/clang/8.0.1/include/popcntintrin.h:_popcnt64(long long __A)
> /usr/lib/clang/8.0.1/include/popcntintrin.h: return __builtin_popcountll(__A);
--
Take care. Thanks, Brian Inglis Calgary, Alberta, Canada
La perfection est atteinte Perfection is achieved
non pas lorsqu'il n'y a plus rien à ajouter not when there is no more to add
mais lorsqu'il n'y a plus rien à retirer but when there is no more to cut
-- Antoine de Saint-Exupéry
next prev parent reply other threads:[~2023-07-08 19:22 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-07 7:41 Mark Geisert
2023-07-07 9:44 ` Corinna Vinschen
2023-07-07 10:13 ` Mark Geisert
2023-07-07 10:45 ` Corinna Vinschen
2023-07-07 15:46 ` Jon Turney
2023-07-07 15:49 ` Corinna Vinschen
2023-07-07 18:54 ` Brian Inglis
2023-07-08 19:22 ` Brian Inglis [this message]
2023-07-08 20:59 ` Mark Geisert
2023-07-08 21:53 ` Mark Geisert
2023-07-09 4:27 ` Brian Inglis
2023-07-09 7:58 ` Mark Geisert
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=1f7d3254-234e-378f-a852-63ca5d7ca01f@Shaw.ca \
--to=brian.inglis@shaw.ca \
--cc=cygwin-patches@cygwin.com \
/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).