This is the mail archive of the
newlib@sources.redhat.com
mailing list for the newlib project.
R: General information on Hitachi SH math performance
- To: "tm" <tm at kloo dot net>
- Subject: R: General information on Hitachi SH math performance
- From: "Giulix" <giuliano dot to at tin dot it>
- Date: Sat, 18 Aug 2001 12:24:40 +0200
- Cc: <newlib at sources dot redhat dot com>,"Silverio (casa)" <sdiquig at goldvi dot it>
- References: <009501c124b0$ff08dee0$7aa72dd5@343> <3B7DD37B.8D875235@kloo.net>
> You didn't mention which SH family part you are using, but you may want
> to check
> if you are using the correct libgcc.a for your version of the SH
> processor.
>
> There are different versions built for the SH1, SH2, SH3, SH3E, SH4,
> SH4-single,
> and Sh4-single-only ABIs.
I'm using the SH7045F Hitachi microprocessor witch is a SH2 series. So I use
the -m2 options in gcc compiler.
I think the performance difference is because the math library (especially
float library) is wrote in C source coce in gcc and in assembly in other
commercial compilers.
Whath do you think about ?
Giulix