Gregory, Interesting investigations. Looks like classlib porting to 64 bit is needed.
Btw, here is another vector for hacking :) Since the JIT is not ported to em64t yet, but works on ia32, it could be possible to build 32-bit version of DRLVM using compatibility layer in linux-x86_64 (gcc -m32). It should give reasonable performance. -- Ivan 12.05.06, Gregory Shimansky<[EMAIL PROTECTED]> написал(а):
On Thursday 11 May 2006 20:44 Rana Dasgupta wrote: > > >>Yes, there are some other places which would have to be slightly > > >>corrected to make this buildable on x86_64. > > >>But, may be it makes sense to ensure that everybody can build & run it > > >>successfully on the supported (e.g. IA32) configuration first? > > I agree. I think that it is premature to start patching the submission > for 64 bit, it's just too confusing. The submission was what, VM? As I've written VM doesn't have inherent limitations not to work on x86_64 platforms in interpreter mode. I've just actually made Hello World application to run. Running eclipse causes segfault so far somewhere in native libraries of classlib, not in VM. > > Since port library comes only > > in IA32 version, the file hysignal.c fails to compile, specifically > > functions > > > infoForGPR, infoForControl and infoForModule. ... > > Architecture/platform specific code is a genuine issue. We will need to > fix files like hysignal.c ( and others similar ) when we do the 64 bit > port. At the moment I've changed the function bodies to assert(0 && "unimplemented") and these assertions didn't fail so far. -- Gregory Shimansky, Intel Middleware Products Division --------------------------------------------------------------------- Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- Ivan Intel Middleware Products Division