harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mikhail Fursov" <mike.fur...@gmail.com>
Subject Re: [drlvm][jit] Code review of the following issue - [jira] Commented: (HARMONY-3246) [drlvm][jit][opt] Jitted code can not be executed on non-SSE2 platforms
Date Thu, 19 Apr 2007 05:44:11 GMT
Nathan,
Instead of working on intermediate 2.a item we discussed 2.b
"b. Start discussion: Use CPUID to detect if SSE2 is available and make JIT
turn this pass on automatically if needed. "
I'm preparing e a patch to make EM/JIT ready to work on old hardware without
any additional options specified.

Bad news is that with the patch applied only OPT compiler will be ready to
work with x87 instructions. The unavailability of JET won't affect
performance but will hit client applications startup time.

On 4/19/07, Nathan Beyer <ndbeyer@apache.org> wrote:
>
> On 4/18/07, Rana Dasgupta <rdasgupt@gmail.com> wrote:
> > On 4/18/07, Nathan Beyer <ndbeyer@apache.org> wrote:
> > >
> > > No, no no! We need to be able to run on a non-SSE2 CPU without special
> > > command-line properties.
> > >
> >
> > That is what I heard Mikhail say...add a p5 pass to every jit
> > configuration  ( default off ) and switch it on automatically only if
> > the jit detects a non SSE2 cpu. No additional command line properties
> > needed.
> >
>
> Then what does this mean?
>
> > 2) a. After the commit: Make it turned off by default and use
> > additional cmd-line -XX:jit.arg.codegen.p5=on when run on PC without
> SSE2
>
> Did I read that wrong?
>
> I feel like you guys are speaking in a different language here. Maybe
> I'm just behind on a few things, but I got lost in all of the EM, OPT,
> JET, conf, k5, p5 references. Is there lexicon reference somewhere, so
> I can match this stuff up to the DRLVM documentation on the web site?
> Is the web site the most up-to-date reference for this [1]?
>
> -Nathan
>
> [1] http://harmony.apache.org/subcomponents/drlvm/
>



-- 
Mikhail Fursov

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message