On Mon, 12 Mar 2007, Bill Broadley wrote:
Robert G. Brown wrote:
It's worth a small editorial insertion here that I "like" hypertransport
for a variety of reasons -- perhaps because it is a packet-based
internal network that makes your computer's CPU-memory architecture
surprising like a cluster in miniature on the inside. And just like any
compute cluster, you need to tune your design choices towards your
application.
I agree completely. Sad that AMD doesn't play to their strength.
Hypertransport had the potential for significant differentiation from
Intel solutions. Currently it allows for dual sockets with somewhat
better memory performance that Intel's latest greatest dual FSB
woodcrest machines. The story on quad socket shows a clear and
substantial advantage for hypertransport, but alas most of the market
isn't in quad sockets.
Seems like if AMD didn't get so stingy with the coherent HT links (which
only the 8xx chips usually have spare ports) that the market might generate
some really interesting opteron based solutions.
Have you looked over the HT consortium site? It is truly a lovely spec
but they didn't quite make the most of their brief window of opportunity
to kill off the bus. Everybody went PCI-X instead of HT. I don't know
if this was due to Intel FUD, or a lack of energy from other consortium
members, or the difficulty of overcoming vendor persistence in sticking
with PCI or... most likely it was just the lack of Intel support.
Who would invest in an HT connection for their network card when it will
never run on 3/4 of all servers made by design? I've seen this before,
a couple of times, over the lifetime of the PC. Somebody comes up with
a couple of competing bus designs, sometimes there are even cards
released for both, but eventually Intel gets behind just one or the
other proves unscalable, and... the rest is history, at least for 4-5
years until the next one comes along.
Intel isn't anywhere nearly as bad as Microsoft, and there have "always"
been 2-3 makers of competing Intel compatible chips, since shortly after
the PC. AMD is just the survivor of this last-man-standing competition
and they do have the chops to do some really good things. The opteron
really gave them a solid chunk of the high performance market and a lot
of cred. But definitely -- they needed to carry HT a lot further,
because it was very much one of the major advantages of their entire
design philosophy that has never quite made it big.
rgb
After all if pathscale can manage hypertransport -> IB, how much harder
would shared memory be..... if of course coherent HT worked at a reasonable
price.
Or maybe if Nvidia can manage to build a low volume video card
with 768MB and a 80GB/sec memory system for $560 maybe some enterprising
company could build a opteron based motherboard with amazing memory
bandwidth..... if of course coherent HT worked at a reasonable
price.
Why shouldn't a video card be able to directly access all memory (and
vice versa)?
Then again maybe this is what AMD's torenza initiative is about
Seems like AMD stands to gain much more from the potential of
future coherent HT products then they do by losing $$$'s on lower
prices for the 8xx and 8xxx products vs the 2xx and 2xxxx products.
Not to mention if the world bought more quad socket machines that would
be a world that AMD would have a much larger lead in (because of
the onchip memory controller).
--
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:[EMAIL PROTECTED]
_______________________________________________
Beowulf mailing list, Beowulf@beowulf.org
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
