This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] PPC64: First in the series of patches implementing POWER8 vector math.
- From: Tulio Magno Quites Machado Filho <tuliom at ascii dot art dot br>
- To: Steve Ellcey <sellcey at marvell dot com>, "libc-alpha\@sourceware.org" <libc-alpha at sourceware dot org>, "tnggil\@protonmail.com" <tnggil at protonmail dot com>, andrew dot n dot senkevich at gmail dot com
- Cc: William J. Schmidt <wschmidt at linux dot ibm dot com>, Segher Boessenkool <segher at kernel dot crashing dot org>
- Date: Mon, 18 Feb 2019 15:32:40 -0300
- Subject: Re: [PATCH] PPC64: First in the series of patches implementing POWER8 vector math.
- References: <7z03WSXwPIOYABvW1ZYzHobMOv90wBWKo-e0YptRFovFUpRnrSEGN-zv08kOe_fPJMFnHx1khbcfcfooubTStEAtMX96Q2U4cRVZOm_TZl8=@protonmail.com> <6e6061809e8124e44b8f830bd311ba696c19c8a9.camel@marvell.com>
Steve Ellcey <sellcey@marvell.com> writes:
> I am curious, have patches been sent or will patches be sent to GCC to
> generate calls to vector functions. I do not see any of the
> TARGET_SIMD_CLONE* macros defined for power8 in the GCC tree so I don't
> see how it would ever generate calls to the vector functions.
That still has to be implemented.
> I am also not clear on how we decide on the mangling of the vector
> function names, particularly the 'b'. I know that x86 uses 'b', 'c',
> 'd', or 'e' depending on the FP vectors available. Aarch64 is using
> 'n' in its name mangling (and maybe something else for SVE). Is it OK
> for Power8 to use the same letter as x86? I don't know if this is
> covered in some standard or how the letters were chosen.
It was proposed to the X86-64 System V Application Binary Interface, but
it was refused. [1]
Another question: in the C++ ABI, "_ZGV" is reserved for guard variables.
How is this name collision being treated?
Andrew, could you help answer these questions?
[1] https://groups.google.com/d/msg/x86-64-abi/LmppCfN1rZ4/TydP1Gxr4cIJ
--
Tulio Magno