1 |
Bob Sanders <rsanders@×××.com> posted 20060814151338.GA102938@×××.com, |
2 |
excerpted below, on Mon, 14 Aug 2006 08:13:38 -0700: |
3 |
|
4 |
> Gfx, especially 3D, is about memory bandwidth. Move the memory out |
5 |
> of direct contact with the gpu chip - say via a socket, and it's |
6 |
> necessary to drop the frequency that the memory interface runs at. |
7 |
> It's basic electronics - add more capacitaince and inductance, and |
8 |
> the speed of the interface goes down. |
9 |
|
10 |
Two factors in counter-point. We're talking FB-DIMM tech time this comes |
11 |
out, and this will be using upgrades to the same memory controller on the |
12 |
CPU (only in this case GPU) AMD is already using to maintain its lead in |
13 |
multi-socket performance against Intel (even while Intel's Core-Duo has |
14 |
caught up at the low single-socket end). |
15 |
|
16 |
> Remember, Intel doesn't make real 3D Gfx chips. Unlike, Nvidia and |
17 |
> Ati, Intel does most of it's 3D processing in software. Thus opening up |
18 |
> the chips specs and driver has little impact on any IP outside of Intel as |
19 |
> it doesn't expose any IP that might belong to Micrsoft or SGI. |
20 |
|
21 |
Does <> will do. With aeroglass driving dx10 requirements, the low end of |
22 |
the gfx market is about to get very legacy, very fast. Their new chips |
23 |
do more in hardware because they must, due to the performance |
24 |
requirements. That's the reason Intel's latest free driver release made |
25 |
the headlines in the FLOSS world -- it's the first time we've had full |
26 |
sources to something that capable. |
27 |
|
28 |
> While ArsTechnica tends to be fairly reliable, I've found stories from the |
29 |
> Register to be very, very, wrong - the assumptions were way off base or the |
30 |
> reporter got the entire story wrong because they ignored documented history. |
31 |
> I'd be very wary of believing a story in the Register without cross checking |
32 |
> it with independent sources. Sometimes the Register will get one sentence |
33 |
> in the story correct. Sometimes, the entire story is made up from that one |
34 |
> sentence. |
35 |
|
36 |
True to a point. However, they /do/ get credit for being one of the first |
37 |
to report the AMD/ATI merger discussion, way back when it was low-cred and |
38 |
most were saying it was crazy because AMD had always said it did better |
39 |
NOT doing the chipsets -- it did them early on in a cycle when third party |
40 |
chipsets weren't yet widely available, but didn't upgrade them and |
41 |
eventually phased them out as third party chipset solutions came online. |
42 |
(This was the case for both the K7 and K8 product cycles.) |
43 |
|
44 |
Anyway, I'm not saying it'll happen, only that this is the best |
45 |
opportunity for it to happen we have, and if for whatever reason AMD/ATI |
46 |
fail to open their drivers, there will be a lot of folks for whom AMD is |
47 |
no longer a viable option (unless other solutions appear). AMD certainly |
48 |
realizes this, and I expect they'll try to open them -- at least a |
49 |
functional subset (possibly retaining some stuff closed, but as I said, a |
50 |
functional 3D subset, enough to keep them in the game, anyway). |
51 |
|
52 |
Meanwhile, I'm adjusting to the fact that it's quite possible my next |
53 |
platform will be Intel. However, as I mentioned earlier, that's a ways |
54 |
out yet, and in the computer technology sector, a lot can happen in 2-4 |
55 |
years... Hey, in that time, it's actually possible Vista will prove |
56 |
seriously problematic and the MS market share will have actually budged a |
57 |
bit (comparable to the IE/Firefox situation today, MS still has a big |
58 |
lead, but for the first time in history, their share has actually budged |
59 |
in the downward direction in real measurable terms). The signs are |
60 |
already there that it's ready for it, at least elsewhere than the US. How |
61 |
far behind the US will be, who knows, but it's obviously not front of the |
62 |
pack. The thing is... if those numbers actually start to budge, it'll |
63 |
change the dynamics substantially, and I'm not sure any of us grok just |
64 |
how different things might be in that case or where the tipping point is. |
65 |
|
66 |
-- |
67 |
Duncan - List replies preferred. No HTML msgs. |
68 |
"Every nonfree program has a lord, a master -- |
69 |
and if you use the program, he is your master." Richard Stallman |
70 |
|
71 |
-- |
72 |
gentoo-amd64@g.o mailing list |