Re: [PATCH v1 11/11] tests/tcg: add vectorised sha512 versions

[Matheus K. Ferst]

On 14/02/2022 12:14, Alex Bennée wrote:
> "Matheus K. Ferst" <matheus.ferst@eldorado.org.br> writes:
>
> > On 11/02/2022 13:03, Alex Bennée wrote:
> > > This builds vectorised versions of sha512 to exercise the vector code:
> > >     - aarch64 (AdvSimd)
> > >     - i386 (SSE)
> > >     - s390x (MVX)
> > >     - ppc64 (vector)
> > > Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
> > > Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> > > Message-Id: <20220202191242.652607-5-alex.bennee@linaro.org>
> > > ---
> > > v2
> > >     - use -msse4.1 -O3 instead of -pentium4 for i386 build
> > > ---
> > >    tests/tcg/multiarch/sha512.c      | 2 +-
> > >    tests/tcg/aarch64/Makefile.target | 7 +++++++
> > >    tests/tcg/arm/Makefile.target     | 8 ++++++++
> > >    tests/tcg/i386/Makefile.target    | 6 ++++++
> > >    tests/tcg/ppc64le/Makefile.target | 5 ++++-
> > >    tests/tcg/s390x/Makefile.target   | 9 +++++++++
> > >    tests/tcg/x86_64/Makefile.target  | 7 +++++++
> > >    7 files changed, 42 insertions(+), 2 deletions(-)
> >
> > <snip>
> >
> > > diff --git a/tests/tcg/ppc64le/Makefile.target 
> > > b/tests/tcg/ppc64le/Makefile.target
> > > index 480ff0898d..4f1d03dfcf 100644
> > > --- a/tests/tcg/ppc64le/Makefile.target
> > > +++ b/tests/tcg/ppc64le/Makefile.target
> > > @@ -5,10 +5,13 @@
> > >    VPATH += $(SRC_PATH)/tests/tcg/ppc64le
> > >    ifneq ($(DOCKER_IMAGE)$(CROSS_CC_HAS_POWER8_VECTOR),)
> > > -PPC64LE_TESTS=bcdsub non_signalling_xscv
> > > +PPC64LE_TESTS=bcdsub non_signalling_xscv sha512-vector
> > >    endif
> > >    $(PPC64LE_TESTS): CFLAGS += -mpower8-vector
> >
> > Since this test does not target a specific instruction, maybe it
> > should use -mvsx/-maltivec to allow the compiler to use newer
> > instructions.
>
> I wasn't sure which vector instructions are supported by the TCG front
> ends so if the above flags won't trip up the TCG I can add them to the
> cflags.

AFAICT, we should have all vector instruction until POWER9. POWER10 is 
WIP, but current versions of GCC/Clang are not emitting any of the 
missing instructions, even with -mcpu=power10

Thanks,
Matheus K. Ferst
Instituto de Pesquisas ELDORADO <http://www.eldorado.org.br/>
Analista de Software
Aviso Legal - Disclaimer <https://www.eldorado.org.br/disclaimer.html>