NVISION 08 - Jen-Hsun Talks Larrabee, Mobility, VIA and More
by Anand Shimpi & Larry Barber on August 26, 2008 3:00 PM EST- Posted in
- Trade Shows
The Larrabee Question
Let's talk about Larrabee.
NVIDIA pointed out that Larrabee x86 isn't binary compatible with other Intel x86 processors (since it doesn't support any of the SSEs) - so there's no advantage there.
Honestly, x86 today is a burden for Larrabee, not a boon as it is not the most desirable ISA from anything other than a compatibility standpoint. The difference between G2xx and Larrabee is in the programming model not the ISA. It's the threading model with G2xx that the developer complaints are really about.
NVIDIA says that it simply takes a new approach to development - focus on data in and data out, rather than conventional top to bottom function coding. The issue is that programmers don't like to change the way they work.
The real question is: when Larrabee ships, will its threaded programming model be significantly easier than G2xx. At this point it's simply too early to tell, Intel thinks it will be and many of the developers I've spoken to agree, but NVIDIA keeps arguing that Larrabee's programming model will be just as different as CUDA and that NVIDIA has the inherent advantage here because of the experience it has had in building GPUs for the past 15 years.
Would NVIDIA Integrate a CPU?
David Kirk summarized, quite well, his thoughts on whether NVIDIA would ever pursue putting a CPU on die next to one of its GPUs.
Kirk's view is that at the low end there's a place for a single-chip CPU/GPU, he views integration (rightly so) as a low cost play. "None of our customers ask us for less performance, why would we ever take away part of our GPU and put a CPU in it?"
NVIDIA currently competes in the low end of the GPU market with its sub $75 GPUs and IGP chipsets. The integrated CPU/GPU does stand the chance of eating into NVIDIA's largest quantity market, and it doesn't look like NVIDIA stands the chance to compete there - at least in x86 desktops/notebooks. Why would you pay more for a NVIDIA chipset with integrated graphics, if you already get integrated graphics on every single CPU you buy?
We've got a future where AMD/Intel ship these hybrid CPU/GPUs on the low end, GPUs like the RV770 and Larrabee at the high end, and NVIDIA is already being pushed out on the chipset side (neither Intel nor AMD wants to be the #2 manufacturer of chipsets for their own CPUs). In the worst case scenario, if NVIDIA gets squeezed by everything I just mentioned over the next few years, what is NVIDIA's strategy going forward? Jen-Hsun actually highlighted one possible direction...
Facebook
Twitter
RSS
18 Comments
View All Comments
SoulSlave - Wednesday, August 27, 2008 - link
Well, they simply don't have to deliver linear performance increase. Anything above what SLI has to offer, at a reasonable price point would suffice, I mean, who would prefer to be straped to a single GPU manufacturer when you could have any combination of GPUs you want, delivering better performance?biostud - Wednesday, August 27, 2008 - link
The only problem for nvidia regarding Lucid Hydra is the loss of a need for a nvidia chipset.For both AMD/ATi and nVidia Lucid Hydra still needs two videocards, and since this is the core business for both companies it doesn't matter that much if it's going to replace Crossfire and SLI.
rjc - Tuesday, August 26, 2008 - link
As have an affected notebook, have been following the bad chips news stories, there is a reasonably informative article here:http://www.tgdaily.com/content/view/39045/135/">http://www.tgdaily.com/content/view/39045/135/
If above holds i cant imagine the relationship between nvidia and notebook oems will be very good over the next 12 months. Whichs a pity for them as the notebook market seems to be growing quite well despite everything. I guess this is the idea behind the smartphones move - different business though volumes higher and margins tighter than on high end graphic cards.
whatthehey - Wednesday, August 27, 2008 - link
Strangely enough, all of the stories related to this inevitably seem to lead back to the Inquirer "reporting". You'll pardon me if I just can't take anything from Charlie Demerjian seriously. This is the same guy that hates NVIDIA and Intel, so I hardly expect his take on the situation to be reliable. There are certainly some problems with some of NVIDIA's chips, but nothing has indicated this extends much beyond a few select chips (8400M, 8600M and 8700M I think).What specific laptop do you have that has had problems, and what GPU does it use? I know I have a laptop with an 8800M GTX and I've had no problems whatsoever. (It's one of the Clevo units.)
rjc - Wednesday, August 27, 2008 - link
The article i linked claims from their sources the problem chips are the G84, G86 and G92. They couldn't confirm if the G94 also had problems.Again from the article nvidia has sold a total of 70 million of these chips. If say only 1%-2% need to be replaced then nvidia will be ok, say 10% or more would be kind of disastrous. As its early days people are trying to extrapolate from the current failures to get the eventual overall rate.
Re shooting the messenger is a natural thing to want to do. I think psychologists call it "correspondence bias" where you assume the messenger has an agenda ie is evil and thus their message can be ignored.
Personally my laptop is a asus with a 8600gt a v1s. It idles at over 60C, heavy workload can push it into the 70s ie the danger zone from the article, finally games send it into the mid 90s. According to notebook-review forums asus is getting regular returns on their g1s gaming laptop which has a 8600gt and trying to replace them on the quiet. I guess they want to try and avoid the bad publicity dell and hp recently got.
Frallan - Friday, August 29, 2008 - link
Any1 heard about the C90s??Same chip - but other cooling however Im starting to get some very disturbing signs from mine.
Also yes it it the Inq-well but that doesn't mean its wrong. Analog Just because you are paranoid it doesnt mean they are not coming for ya.
;0)
/F
iwodo - Tuesday, August 26, 2008 - link
ARM with Nvidia Graphics, aka the Tegra. And Desktop, x86 with VIA Nano.I am sure with some Restructuring Nvidia could surely sell x86 in some form. As long as it take control of VIA by other means without losing x86 license.
Interconnect - there are already Hypertransport. Nvidia has the Chipset experience. And Dual Core Nano isn't exactly low performance either.
If AMD could do it, why not Nvidia?
Griswold - Wednesday, August 27, 2008 - link
"As long as it take control of VIA by other means without losing x86 license."Voodoo. But since that doesnt exist, its not going to happen. Which answers your last question.
shin0bi272 - Tuesday, August 26, 2008 - link
Nvidia seems to be too concerned with other things at the moment to do anything about Lucid's Hydra. I also dont think that the hydra engine cares what the instruction is it just passes it to the right video card and takes requests from the cards and passes them to the north bridge. Its like the Xor chip on a hardware raid5 card. It just sends what ever data is sent to it to the right card or reads from the different cards and sends those requests to the system. Its a traffic cop. Well thats how I see it anyway... whether or not it actually works remains to be seen.computerfarmer - Tuesday, August 26, 2008 - link
It is nice to hear new things and different takes on the subject.One thing I would like to see, is the promised review on the AMD 790gx/SB750.