1 |
Hi Duncan & Ryan, |
2 |
|
3 |
Thanks for the hints so far. I'm going to check out Tyan's support for Linux |
4 |
before getting more astonished (by the possibilities Duncan mentioned :) |
5 |
|
6 |
I think the 4-way plugin and the possibility to substitute single to dual core |
7 |
procs are one of S4881's strengths but I'm not sure that next year or later |
8 |
(when one gets the money) it will still be possible to order the plugin. |
9 |
|
10 |
One more promising fact is that I haven't seen the half-of-four gigs waste of |
11 |
RAM with Tyan S2882 (as opposed to MSI MasterX-Far mobs which, |
12 |
in addition, wouldn't let me install >4G). As I'v learnt from one of |
13 |
Duncan's earlier messages, it was most probably due to address space |
14 |
allocation for various buses by the BIOS. On S2882, I could also mix up RAM |
15 |
modules of various sizes and clock rates (I suppose the mob syncronized to |
16 |
the slowest) on any of the two ranks... |
17 |
|
18 |
Anyways, I hope memory can be shared on S4881 (or -D, S4882) as well |
19 |
as on S2882, meaning that any thread can utilize any part of (or all) the |
20 |
physical memory mounted on the mob. They want to use all the 64G memory |
21 |
(only "all" because (i) no mob supports more and (ii) resources are limited). |
22 |
|
23 |
Thanks again, |
24 |
Tamas |
25 |
-- |
26 |
gentoo-amd64@g.o mailing list |