Date: Tue, 9 Nov 1999 15:56:44 +0000 (GMT) From: Mark Powell <M.S.Powell@salford.ac.uk> To: Charles Randall <crandall@matchlogic.com> Cc: freebsd-questions@freebsd.org Subject: RE: Stress testing a machine with "make world" Message-ID: <Pine.BSF.4.05.9911091553150.21666-100000@localhost> In-Reply-To: <64003B21ECCAD11185C500805F31EC0304621B98@houston.matchlogic.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 9 Nov 1999, Charles Randall wrote:
> Out of curiosity, how much RAM and swap do you have?
128MB RAM - 556MB swap
make world (Didn't swap at all. Completed okay)
make -j60 world (Saw 50-60MB in swap, but was only a cursory check.
Paniced.)
Later I put the RAM upto a total of 256MB.
make -j200 world (Saw it swapping, but can't remember how much.
Completed okay)
> Were there any indication that the system was running low on swap?
None at all. Too much swap for that?
> Was there a message in the syslog?
Nope. Re: swap space, you mean?
> Charles
>
> -----Original Message-----
> From: Mark Powell [mailto:M.S.Powell@salford.ac.uk]
> Sent: Tuesday, November 09, 1999 4:35 AM
> To: freebsd-questions@freebsd.org
> Subject: Stress testing a machine with "make world"
>
>
> We all know "make world" is a good acid test of whether a machine has
> decent motherboard, RAM and disk subsystem. However, how far does one go?
> And what is a good stress test of a machine with lots of RAM?
> I've had a machine forced upon me, which is based on a Micro-Star 6119
> motherboard. Never heard of them before, so I decided to test it out.
> Handles a 3.3R "make world" fine, but will panic the kernel on a
> "make -j60 world".
>
> -----
> IdlePTD 2711552
> initial pcb at 2310b8
> panicstr: page fault
> panic messages:
> ---
> Fatal trap 12: page fault while in kernel mode
> fault virtual address = 0x29a
> fault code = supervisor read, page not present
> instruction pointer = 0x8:0xc0174bef
> stack pointer = 0x10:0xc78b5f30
> frame pointer = 0x10:0xc7be8174
> code segment = base 0x0, limit 0xfffff, type 0x1b
> = DPL 0, pres 1, def32 1, gran 1
> processor eflags = interrupt enabled, resume, IOPL = 0
> current process = 2 (pagedaemon)
> interrupt mask =
> trap number = 12
> panic: page fault
>
> syncing disks... 133 133 133 119 93 45 7 1 done
>
> ....
>
> ---
> #0 boot (howto=Cannot access memory at address 0x20.
> ) at ../../kern/kern_shutdown.c:285
> 285 dumppcb.pcb_cr3 = rcr3();
> ---
>
> These sort of errors are bad hardware?
> This always happens when the machine has 128MB of RAM. If the RAM is
> increased to 256MB, I can't get the "make world" to fail, even putting the
> number of concurrent jobs up into the hundreds. I presume this is due to
> it swapping less and not encountering the problems with the hardware? How
> can one test such a machine without resorting to a "make -j100 world" on
> each and even DIMM in turn?
> Cheers.
>
> Mark Powell - UNIX System Administrator - Clifford Whitworth Building
> A.I.S., University of Salford, Salford, Manchester, UK.
> Tel: +44 161 295 5936 Fax: +44 161 295 5888 www.pgp.com for PGP key
> M.S.Powell@ais.salfrd.ac.uk (spell salford correctly to reply to me)
>
>
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-questions" in the body of the message
>
Mark Powell - UNIX System Administrator - Clifford Whitworth Building
A.I.S., University of Salford, Salford, Manchester, UK.
Tel: +44 161 295 5936 Fax: +44 161 295 5888 www.pgp.com for PGP key
M.S.Powell@ais.salfrd.ac.uk (spell salford correctly to reply to me)
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.05.9911091553150.21666-100000>