Re: GCJ Native Proposal
On Tue, Mar 15, 2005 at 04:06:45PM +0100, Wolfgang Baer wrote:
> Michael Koch wrote:
> [...]
> >You are right, its not always a gain. Tom Tromey told me that he is
> >aware of one case where the native library is slower then interpreting
> >the jar.
> >Doing (c) and fixing JIT runtimes can be good, but it ist a hard work
> >too. The Kaffe people put much efforts into this. JIT works find on i386
> >but not at all on powerpc for them. Same mixture for all the other
> >Debian archs. When Kaffe supports the BC-ABI too this will be huge gain
> >after all as it is surely be faster then non-JITed bytecode. The idea is
> >make the code reuseable to other VMs can adopt it too.
> >We have now 3 possibilitiest:
> >1) compile all to native
> >2) compile a selected group of jars to native
> >3) dont compile to native at all and decide later
>
> Are there any tests done in the past for the gain against jit or
> interpretative mode ?
>
> Otherwise I think some testcases would be very good here:
>
> How much gain does it bring for architectures (like i386) - where
> already a good jit implementation is available ?
>
> How much gain does it bring for achitectures without a jit
> implementation ?
>
> If there is quite a gain for architectures without a jit atm and maybe its also a long way until they have one - it would be worth to provide something in my opinion. Especially as these arches (non i386) are the
> one which are left alone by the commercial vm implementors.
>
> This way we could also give recommendations to the users later. E.g.
> if you are on non-jit arch install the recommended -jbi (or whatever
> name will be choosen) packages.
There are floating around some older comparisions. E.g. this one here:
http://www.klomp.org/mark/free-vm-benchmarks/. Afaik they are all made
on i386 as that is the most common arch.
Michael
--
Java Trap: http://www.gnu.org/philosophy/java-trap.html
Reply to: