Re: [gentoo-amd64] Re: Re: More ATi driver madness

[Bob Sanders]

Duncan, mused, then expounded:
>
> As for AMD, here's what I skipped over in the previous post.  They've
> already teamed up with various third parties to develop and sell physics
> and floating point processors slotted into additional CPU sockets, linked
> directly to the multi-core CPUs via Cohesive HyperTransport.  The next
> step is doing the same thing with full video processors, a decent portion
> of which are physics processors anyway.  There are however a couple issues
> with the idea, including the fact that getting video folks to  commit to an
> AMD specific platform when they already have PCI-E would have been rather
> difficult.  This is supposedly one of the big reasons they bought ATI --
> to give AMD the where-with-all to follow up on that idea.

Sorry, the conclusion is very likely wrong.  True AMD wants several things
that ATI brings to the table.  But, there is little chance of Gfx being
integrated fully into the cpu or even into a socket via a HyperTransport
link.  It's just too costly and the performance, for all but UMA (Unified
Memory Archittecture) memory interface would suffer greatly.  It's also
one of the reasons AMD is adding a PCIe interface onto their cpus - the
1207 pin Socket F (is that the correct socket?) models, later next year.

Gfx, especially 3D, is about memory bandwidth.  Move the memory out
of direct contact with the gpu chip - say via a socket, and it's
necessary to drop the frequency that the memory interface runs at.
It's basic electronics - add more capacitaince and inductance, and
the speed of the interface goes down. 

True, it's possible to put low-end, UMA based, graphics into a socket,
but that's not really a long-term goal.  Rather, just an interesting
Engineering exercise.  Plus, in laptops, where it would be best, the
socket cost - both dollars and size, negates any benefit of moving
away from it existing on the southbridge where it is today.


> Now, in ordered
> to fully populate the ecosystem, they'll /have/ to open things up.  Keep
> in mind that Intel is designing similar stuff with it's integrated
> solutions, and if AMD didn't get in the game, it wouldn't be long until
> they were no more relevant than Via in the x86 CPU market.

Both AMD and Intel are moving, but not in the direction of more complexity.
They are moving to simpler parallel execution units - like the parallel
Gfx pipes found on GPU chips today, only for generic computing.  All this
time, cpus have had a very diffcult time moving from serial processing
and the efforts for massivly parallel units have not faired well - Sun's
last UltrsSparc.  At the same time Gfx systems have become more and more
parallel, and with the push from the late 3DLabs, have become more 
like generic, programable, compute blocks, though still diffcult to
access from main menory.


>
> So... I think AMD/ATI /could/ open their video specs.  However, it's still
> an open question of whether they /will/, tho I think the chances are
> pretty good, as Intel will have them between a rock and a hard place if
> they don't.
>

Uh, no.  Remember, Intel doesn't make real 3D Gfx chips.  Unlike, Nvidia and
Ati, Intel does most of it's 3D processing in software.  Thus opening up
the chips specs and driver has little impact on any IP outside of Intel as
it doesn't expose any IP that might belong to Micrsoft or SGI.

Both ATI and Nvidia have IP in their hardware that came from outside
of those companies.  Opening up their drivers exposes this IP and it's
not theirs to give away.  And given that some IP was sold outright
to Microsoft, even if SGI were to open up the Gfx IP that is licensed, some
would still be held up in Redmond, as would any implementation in hardware
that insures decent performance with DirectX 9 and 10.  So outside of the
binary blobs that hide the IP, like Nvidia implements today, there will
be little change in drivers from Nvidia.  With luck AMD will get ATI
to at least get their Linux drivers into a mroe suitable form on par
with Nvidia.
 
> (My sources for much of this were articles at ArsTechnica and the
> Register, in turn quoting trade mags and industry analysts, plus AMD's own
> PR.)
>

While ArsTechnica tends to be fairly reliable, I've found stories from the
Register to be very, very, wrong - the assumptions were way off base or the
reporter got the entire story wrong because they ignored documented history.
I'd be very wary of believing a story in the Register without cross checking
it with independent sources.  Sometimes the Register will get one sentence
in the story correct.  Sometimes, the entire story is made up from that one
sentence.

Bob
-  
-- 
gentoo-amd64@g.o mailing list