vic & affinity

[Robert Olson]

At 01:23 PM 11/24/2003, Ivan R. Judson wrote:

>Oh yeah; when I did the performance test to compare openmash vic with ag-vic
>I used this to "pin" each to a cpu of a dual cpu box. They come out +/-5% of
>each other on average. No stability problems for 5+ days for either of them.
>
>I'm not sure I'd still propose moving to OM vic; but I would like to keep
>the headless vic part of it around, unless we find an alternative. I'm not
>particularly happy with the state of openmash stuff at the moment.

How so? I thought things were good over there.

If we do have the performance in om, something that we still really need 
(that came up as really obvious at SC) is the layout-authoring capability 
... and the new autoplace hook infrastructre in OM would make that much 
easier to do.

--bob