Re: [PATCH] GFP_THISNODE for the slab allocator

[David Rientjes <rientjes@google.com>]

On Sun, 17 Sep 2006, Paul Jackson wrote:

> David,
> 
> Could you run the following on your fake numa booted box, and
> report the results:
> 
> 	find /sys/devices -name distance | xargs head
> 

With NUMA emulation, the distance from a node to itself is 10 and the 
distance to all other fake nodes is 20.

So for numa=fake=4,
root@numa:/$ cat /sys/devices/system/node/node*/distance
10 20 20 20
20 10 20 20
20 20 10 20
20 20 20 10

> You've been looking at this fake NUMA code recently, David.
> 
> Perhaps you can recommend some other way from within the
> mm/page_alloc.c code to efficiently (just a couple cache lines)
> answer the question:
> 
>     Given two node numbers, are they really just two fake nodes
>     on the same hardware node, or are they really on two distinct
>     hardware nodes?
> 

The cpumap for all fake nodes are always 00000000 except for node0 which 
reports the true hardware configuration.

Using the previous example,
root@numa:/$ cat /sys/devices/system/node/node*/cpumap
00000003
00000000
00000000
00000000

(Note: 00000003 because numa.* is a dual-core machine).

		David

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>