At 10:27 PM -0800 1999/12/30, Mike Smith wrote:
> The difference has already been explained as a different instruction mix.
> This should be obvious to anyone that has been in the industry for as
> long as you have.
It seems to me that you guys are all talking past each other:
1. He's mentioned that he runs the same instruction mix
(i.e., two seti@home clients) under both Linux and FreeBSD
2. He's mentioned that he's run FreeBSD in both uniprocessor
and SMP modes
3. He's mentioned that he's run FreeBSD SMP with 3.3 without problems
4. He's mentioned that the overclocking is a recent
introduction to the issue and the system was overheating before then
5. He's also mentioned that this is a chipset which we now
know is not directly supported by the measurement interface, and
therefore the temperature multiples might (or might not) be off
Now, I don't recall seeing which version of the Linux kernel he
was talking about -- I'd be guessing that might have some effect. If
it was a 2.0.x kernel, then obviously he's not getting more than one
processor under Linux anyway. Depending on which 2.1.x or 2.2.x
kernel he runs, he should be getting SMP under Linux (if he's built
the kernel to support SMP), although he might not have very good SMP
(2.2.13+, from all reports I've heard). I'd be real surprised if
they found some way to HLT the processor under Linux in SMP, but I
wouldn't rule it out as impossible.
I do still find it suspicious that the temperature difference is
14 or 26 degrees, and that it rises much more quickly than would
appear to be possible, given the thermal mass, etc.... I also find
it suspicious that this works fine under FreeBSD in SMP in version
3.3 (even with all the overclocking, etc...) and that this only
breaks in 3.4.
Of course, the fact that everyone is talking past each other is
not surprising, seeing as many of the correspondents are spread
around the globe.
If I might make a suggestion based on my own recent introduction
to this mailing list -- it really helps if the new poster provides as
much detail as possible and doesn't make any assumptions about how
much they know about the problem based on their decade-plus (or more)
experience, and it really helps if the other people involved give the
guy at least half a break while he tries to explain what weirdo
whacked out thing he's seeing.
Myself, I think I've permanently pissed-off some of the important
contributors to the FreeBSD project (and perhaps even some of the
core members) by coming in too strong. I hope that one day I'll
manage to be able to erase that deficit either by patching things up
with the parties in question, or by making enough of my own positive
contributions that the past can be forgiven, if not forgotten.
Ghu help me if I'm actually the voice of reason in this case, but
I felt like something needed to be said before this got too far out
of hand.
--
These are my opinions -- not to be taken as official Skynet policy
____________________________________________________________________
|o| Brad Knowles, <[EMAIL PROTECTED]> Belgacom Skynet NV/SA |o|
|o| Systems Architect, News & FTP Admin Rue Col. Bourg, 124 |o|
|o| Phone/Fax: +32-2-706.11.11/12.49 B-1140 Brussels |o|
|o| http://www.skynet.be Belgium |o|
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Unix is like a wigwam -- no Gates, no Windows, and an Apache inside.
Unix is very user-friendly. It's just picky who its friends are.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
