[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [f-cpu] F-CPU invited at the Libre Software Meeting



Hi there,

from my experience the software stuff always is at least
one or two generations behind hardware things. My concerns
about f-cpu are also related to development of the compiler
tool kit. Is there a document where the compiler writers
can get an idea about the rules for generation of good and
efficient code?

JG


On Mon, 25 Feb 2002, Yann Guidon wrote:
> Andreas Romeyke wrote:
> > On Sun, 24 Feb 2002, nicO wrote:
> > > I imagine that use SIMD code are quite difficult so it's intersting to
> > > see if it miss some instructions to use all the power of the SIMD stuff.
> > > i know that one of the most problematic things are the instruction that
> > > pack the data, maybe we miss some.
> 
> which ? the existing ones are rather complete AFAIK.
> 
> > Why not use a temporary bytecode generated with some higher level
> > functions (between assembler and libc, I think on a level as forth
> > is).
> > 
> > First the higher level functions will be assembled with easier
> > machine depended code, later they will be substituted by optimized
> > SIMD-Code...
> > 
> > Any hints?
> 
> i am not sure to understand.
> 
> however, in the race for performance, one of worst the problems is to
> get rid of of the intermediate levels of representation, because we
> lose data and semantics during each conversion. What matters most is
> "what does the program do" rather than "how does it do it" because we
> can find better ways.
> 
> I have the feeling that GCC will make really slow programs.
> while the superpipeline can reach a somewhat higher frequency,
> an inadequate compiler makes the system work really slow.
> i fear that this constatation can be used as an argument
> against the project.
> 
> > Bye Andreas
> WHYGEE
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> PS: anybody knows whether he comes to the LSM ?

*************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/