Messages in this thread | | | Subject | Re: SMP scalability: 8 -> 32 CPUs | From | Jes Sorensen <> | Date | 03 Dec 1998 12:01:56 +0100 |
| |
>>>>> "Stephen" == Stephen C Tweedie <sct@redhat.com> writes:
Stephen> Hi, On Wed, 2 Dec 1998 02:09:32 +0200 (GMT+0200), Liran Stephen> Zvibel <liranz@math.tau.ac.il> said:
>> If I'm not mistaken, on my last trip to SGI, they told me that they >> could put up to 2048 processors in their machines.
Stephen> SGI are not doing SMP, though. They divide the system into Stephen> up to 64 partitions or regions, and run a separate instance Stephen> of the kernel on each one. The death of one partition does Stephen> not kill the entire system.
Stephen> The fact that they have hardware support for shared memory Stephen> allows it to look somewhat like SMP to the user, and their Stephen> kernel has very very nice support for single system image Stephen> behaviour, but it is still closer to a tightly-bound cluster Stephen> than to an SMP system.
I am not sure I understand you right here, on the O2000 you can either run the whole box as a single image or I _think_ partition it and run multiple instances of IRIX on it (I think IRIX 6.5 allows this at least, maybe I'm wrong).
The O2000s we have here run as a single image and looks like one big SMP to the users.
Jes
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |