Rachel -
I also am running an SMP kernel with a RedHat 5.2 installation. I have found the same
observations
that you did regarding native and green threads. Unfortunately, green thread with no
JIT is really a
pain.
My observation about native threads and JIT on a multiprocessor system is that
sometimes calls get
scheduled on the second processor and do not return to the correct location (i.e. to
the original
calling function), they get stuck returning on the other processor and the original
call crashes with
a null status, or times out.
I submitted a more detailed report to the Blackdown jitterbug site. I think it was
#591.
Mike
-----Original Message-----
From: Rachel Greenham [SMTP:[EMAIL PROTECTED]]
Sent: Thursday, June 24, 1999 2:52 PM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Java 1.2 pre2 and SMP?
Cees de Groot wrote:
>
> Just to add a data point: I've experienced the same. Also using SuSE 6.1
> (SMP 2.2.10 kernel, I don't know which libc but I think not glibc2.1) on
> a dual PII box. Green threads don't work at all, native threads seem
> to be less stable than the same on a single processor box (which is
> a mobile Pentium/MMX in my case, with the exact same SuSE installation).
Update as time goes on:
1.2 native threads & JIT on SMP is unstable after all.
1.2 green threads & JIT on SMP doesn't start at all.
1.2 green threads & no JIT on SMP seems to be fine.
1.2 native threads & JIT on single-processor system still going strong!
--
Rachel
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
