* Re: Probing for physical number of cpus.
2005-10-28 11:18 Probing for physical number of cpus Roy Dragseth
@ 2005-10-28 12:43 ` Andreas Schwab
2005-11-03 8:10 ` Roy Dragseth
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Andreas Schwab @ 2005-10-28 12:43 UTC (permalink / raw)
To: linux-ia64
Roy Dragseth <Roy.Dragseth@cc.uit.no> writes:
> Do anyone know how to figure out the physical number of cpus in an ia64
> system? No, /proc/cpuinfo wont do because the kernel is booted with
> maxcpus=1.
Read and parse the ACPI tables off of /dev/mem, starting with the address
from /sys/firmware/efi/systab.
Andreas.
--
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
"And now for something completely different."
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: Probing for physical number of cpus.
2005-10-28 11:18 Probing for physical number of cpus Roy Dragseth
2005-10-28 12:43 ` Andreas Schwab
@ 2005-11-03 8:10 ` Roy Dragseth
2005-11-03 21:06 ` Ashok Raj
2005-11-07 10:20 ` Roy Dragseth
3 siblings, 0 replies; 5+ messages in thread
From: Roy Dragseth @ 2005-11-03 8:10 UTC (permalink / raw)
To: linux-ia64
On Friday 28 October 2005 14:43, Andreas Schwab wrote:
> Roy Dragseth <Roy.Dragseth@cc.uit.no> writes:
> > Do anyone know how to figure out the physical number of cpus in an ia64
> > system? No, /proc/cpuinfo wont do because the kernel is booted with
> > maxcpus=1.
>
> Read and parse the ACPI tables off of /dev/mem, starting with the address
> from /sys/firmware/efi/systab.
Yikes! Thanks for the answer. I've been toying around with pmtools/acpidump
which gives me a text dump, but I'm still a bit confused (to say the least).
r.
--
The Computer Center, University of Tromsø, N-9037 TROMSØ Norway.
phone:+47 77 64 41 07, fax:+47 77 64 41 00
Roy Dragseth, High Performance Computing System Administrator
Direct call: +47 77 64 62 56. email: royd@cc.uit.no
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: Probing for physical number of cpus.
2005-10-28 11:18 Probing for physical number of cpus Roy Dragseth
2005-10-28 12:43 ` Andreas Schwab
2005-11-03 8:10 ` Roy Dragseth
@ 2005-11-03 21:06 ` Ashok Raj
2005-11-07 10:20 ` Roy Dragseth
3 siblings, 0 replies; 5+ messages in thread
From: Ashok Raj @ 2005-11-03 21:06 UTC (permalink / raw)
To: linux-ia64
On Thu, Nov 03, 2005 at 12:10:34AM -0800, Roy Dragseth wrote:
>
> On Friday 28 October 2005 14:43, Andreas Schwab wrote:
> > Roy Dragseth <Roy.Dragseth@cc.uit.no> writes:
> > > Do anyone know how to figure out the physical number of cpus in an
> ia64
> > > system? No, /proc/cpuinfo wont do because the kernel is booted
> with
> > > maxcpus=1.
> >
> > Read and parse the ACPI tables off of /dev/mem, starting with the
> address
> > from /sys/firmware/efi/systab.
>
> Yikes! Thanks for the answer. I've been toying around with
> pmtools/acpidump
> which gives me a text dump, but I'm still a bit confused (to say the
> least).
>
Actually even of you boot with maxcpus=1, we still create sysfs entries for
each cpus physically present in the system.
you should be able to look at /sys/devices/system/cpu/cpuX entries
if you compiled with CONFIG_HOTPLUG_CPU enabled, then you could
echo 1 > online file in the appropriate cpu directory to add more
cpus.
Let me know if you dont see this behaviour.
Cheers,
ashok
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Probing for physical number of cpus.
2005-10-28 11:18 Probing for physical number of cpus Roy Dragseth
` (2 preceding siblings ...)
2005-11-03 21:06 ` Ashok Raj
@ 2005-11-07 10:20 ` Roy Dragseth
3 siblings, 0 replies; 5+ messages in thread
From: Roy Dragseth @ 2005-11-07 10:20 UTC (permalink / raw)
To: linux-ia64
On Thursday 03 November 2005 22:06, Ashok Raj wrote:
> Actually even of you boot with maxcpus=1, we still create sysfs entries for
> each cpus physically present in the system.
>
> you should be able to look at /sys/devices/system/cpu/cpuX entries
>
> if you compiled with CONFIG_HOTPLUG_CPU enabled, then you could
> echo 1 > online file in the appropriate cpu directory to add more
> cpus.
>
> Let me know if you dont see this behaviour.
Hi, this is more in line of what I was hoping for ;-)
However I don't see this behaviour on my system. I'm running RH EL 4 with
this setup:
[root@testing ~]# ls /sys/devices/system/cpu/
cpu0
[root@testing ~]# uname -a
Linux testing.cluster.none 2.6.9-22.EL #1 SMP Wed Oct 5 15:20:11 EEST 2005
ia64 ia64 ia64 GNU/Linux
[root@testing ~]# cat /proc/cmdline
BOOT_IMAGE=scsi0:\EFI\redhat\vmlinuz-2.6.9-22.EL console=ttyS0 maxcpus=1 rhgb
quiet root=LABEL=/1 ro
Upgrading the kernel will not help because I need to be able to probe this
under the install kernel.
Some background:
This is for the Rocks Cluster Distribution (www.rocksclusters.org), where the
compute node installation is done using RH kickstart files that is
automatically generated on, and fetched from, the frontend based on the
information that is sent from the compute node when it asks for the
kickstart file. For instance, the number of cpus on the node is used to
configure the queueing system.
Any further hints is greatly appreciated.
Regards,
r.
--
The Computer Center, University of Tromsø, N-9037 TROMSØ Norway.
phone:+47 77 64 41 07, fax:+47 77 64 41 00
Roy Dragseth, High Performance Computing System Administrator
Direct call: +47 77 64 62 56. email: royd@cc.uit.no
^ permalink raw reply [flat|nested] 5+ messages in thread