1 |
Luke-Jr posted <200507261540.06591.luke-jr@×××××××.org>, excerpted below, |
2 |
on Tue, 26 Jul 2005 15:40:05 +0000: |
3 |
|
4 |
> On Monday 25 July 2005 22:38, Olivier Crete wrote: |
5 |
>> On Mon, 2005-25-07 at 22:24 +0000, Luke-Jr wrote: |
6 |
>> > On Saturday 23 July 2005 18:44, Brian Litzinger wrote: |
7 |
>> > > > On Sat, 2005-07-23 at 16:48 +0200, Simon Strandman wrote: |
8 |
>> > > > > Memory to memory copy rate = 1291.600098 MBytes / sec. Block size = |
9 |
>> > > > > Memory to memory copy rate = 2389.321777 MBytes / sec. Block size = |
10 |
>> > > |
11 |
>> > > Memory to memory copy rate = 1302.701782 MBytes / sec. Block size = |
12 |
>> > > Memory to memory copy rate = 2051.979980 MBytes / sec. Block size = |
13 |
>> > |
14 |
>> > Before: Memory to memory copy rate = 557.960449 MBytes / sec. Block size |
15 |
>> > = After: Memory to memory copy rate = 1120.773804 MBytes / sec. Block |
16 |
>> > size = |
17 |
>> > |
18 |
>> > Anyone have a clue why I'm getting half what everyone else gets? o.O |
19 |
>> |
20 |
>> What kind of cpu/ram/motherboard do you have ? |
21 |
> |
22 |
> RAM: 2875MB/1002MB (286%) used (I didn't see swapping during the test, tho) |
23 |
> Motherboard: Asus K8V-SE |
24 |
> CPU: AMD Athlon(tm) 64 Processor 3200+ (2202.876 MHz) |
25 |
|
26 |
Perhaps it has something to do with the layout of your memory (and BIOS |
27 |
configuration), single-channel vs. dual-channel memory access? |
28 |
|
29 |
I'm getting the same lower (550-ish pre- haven't tried the patch yet) |
30 |
readings here. However, I know I only have a single 512 meg stick slotted |
31 |
for each of my two CPUs (dual Opteron), and I'm running NUMA mode, so the |
32 |
memory is only being accessed at single-channel speeds. I expect I'd |
33 |
double those numbers if I had paired sticks operating in per-node |
34 |
interleaved dual-channel mode. If I turned off NUMA and set inter-node |
35 |
interleaving as well, with 4 matched memory sticks, to get full |
36 |
quad-channel 128-bit interleaving, I expect the numbers would rise |
37 |
accordingly. (Of course, the latter would be at the expense of allowing |
38 |
parallel threads running on each CPU independent but parallel access to |
39 |
their own memory. I can get dual channel without foregoing that, since |
40 |
each node is dual-channel capable, but couldn't get quad-channel without |
41 |
foregoing that independent parallel access, since quad-channel is |
42 |
inter-node interleaved.) |
43 |
|
44 |
Of course, that's just supposition, here. If those with the 1100/2200 |
45 |
rates would confirm whether they are running in dual-channel memory mode, |
46 |
as I suspect, and you confirm that you are running single-channel mode, as |
47 |
I am, it'll pretty much confirm that supposition, however. |
48 |
|
49 |
-- |
50 |
Duncan - List replies preferred. No HTML msgs. |
51 |
"Every nonfree program has a lord, a master -- |
52 |
and if you use the program, he is your master." Richard Stallman in |
53 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
54 |
|
55 |
|
56 |
-- |
57 |
gentoo-amd64@g.o mailing list |