NVIDIA shows signs ... [2008 - 2017]

Status
Not open for further replies.
The problem they (x86) have is that their low end desktops still make up the vast majority of their sales. If this is what is being eaten by tablets then they still have to find something to use those low end chips on. Many of them are of course the worst chips of the wafer - the Celerons and Pentiums at 65W etc.

If you basically just say that market is disappearing, then what is going to happen? Is the whole market going to shift up in price so that Celerons are priced at i3 prices? Or are they just going to throw these lower end cpu's away? There has to be a knock-on effect somewhere.

AMD is in deep trouble under these circumstances because they just don't have the performance chips that are going to be left, and they are obligated to pay $1 billion per year to GF for wafers. Kaveri better be something special or they better have a really good plan B.

I wouldn't say it's disappearing, it's transitioning.

Right now, Intel has only a lukewarm response to Arm in tablets with Clovertrail. It's compute capabilities are competitive but it's graphics capabilities are horrible.

If Intel can successfully transition into tablet devices, then that replaces the low end PC netbook/notebook market that is shrinking. And likely regains share that it lost to tablet devices.

And at that point, if we assume that the majority of those devices run some version of Windows rather than Android, then it's basically still PC marketshare albeit in a tablet form.

AMD has the same opportunity there. I'm sure they'll be targeting the tablet landscape as well with their APUs using the Jaguar cores.

And to take that a little further. Is there really a need for a low end desktop PC if an x86 tablet can serve the exact same functionality when hooked up to a monitor, keyboard, and mouse? Granted for extensive storage, you'd have to hook it up to an external USB drive.

Sure, currently Windows tablets are a heck of a lot more than a low end desktop PC, but will that remain true if the volume of Windows tablets were to rise dramatically? Look at netbooks. The prices on those dropped quite significantly when their popularity took off and volumes increased.

Regards,
SB
 
JPR's 2013Q1 GPU report

http://jonpeddie.com/publications/market_watch

AMD lost 0.3%, quarter-to-quarter, Intel slipped 5.3%, and Nvidia increased by 3.6%.

...

AMD’s quarter-to-quarter total shipments of desktop heterogeneous GPU/CPUs, i.e., APUs jumped 30% from Q4 and declined 7.3% in notebooks. The company’s overall PC graphics shipments slipped 0.3%.

...

Nvidia’s quarter-to-quarter desktop discrete shipments were flat from last quarter; and, the company’s mobile discrete shipments increased 7.6%. The company’s overall PC graphics shipments increase 3.6%.
 
Do you have proof?
That's a rhetorical question, right?

I don't know anybody who pays the big bucks for market reports for either company and I'm pretty sure our jimbo doesn't either. So whatever 'source' he has is probably of similar quality as sources about, let's say, silicon yield.

But that doesn't matter anyway. It's more useful to put Peddie and Mercury reports next to each other and see if their results correlate over time. That's an exercise I leave to jimbo. For me, the fact that this outfit has put out this report for many many years is sufficient proof that enough people are willing pay for it.

(Let's also not forget that the public Peddie numbers don't differentiate between integrated and desktop.)
 
I know that AMD has used Mercury reports for the purposes of their slides.

amd_hd4770_01.jpg


I know that Nvidia prefers Mercury's methods as well, or has done in the past - Nvidia declares JPR report 'an irrelevant metric'

I know that people on financial boards tend to use Mercury over Peddie - http://www.investorvillage.com/smbd.asp?mb=476&mn=263502&pt=msg&mid=12790483

Can you find me any evidence of the reverse, ie AMD or Nvidia using Peddie's figures for anything?

Let's not forget Peddie's brilliant benchmarking either! - Why not make it fun by changing names and colours half way through? http://jonpeddie.com/reviews/comments/benchmarking-nvidias-gtx-480-fermi-aib

20100330-mttl-3.jpg


20100330-mttl-4.jpg
 
Last edited by a moderator:
The problem they (x86) have is that their low end desktops still make up the vast majority of their sales. If this is what is being eaten by tablets then they still have to find something to use those low end chips on. Many of them are of course the worst chips of the wafer - the Celerons and Pentiums at 65W etc.

If you basically just say that market is disappearing, then what is going to happen? Is the whole market going to shift up in price so that Celerons are priced at i3 prices? Or are they just going to throw these lower end cpu's away? There has to be a knock-on effect somewhere.

Intel has been selling such bottom of the barrel chips at low power, already - at 17 watts, with the sandy bridge Celeron 847, 867 etc. in the low 1.x GHz clock range. It's not suitable for a tablet but you can see it in nettops, netbooks (the cheap Acer Chromebook), ITX.

Eventually they'll sell 10 watt Ivy Bridge crap, then Haswell and Broadwell crap. Calling it "crap" is all relative, if you gave me a laptop or netbook with a Celeron 857 or similar I'd be pretty happy with it.
But with next-gen Atoms these low end chips might be less needed.
 
(Let's also not forget that the public Peddie numbers don't differentiate between integrated and desktop.)

I find the numbers for 2013Q1 interesting in that Nvidia's market share (discrete only) at 18% nearly equals AMD's market share of 20.2% which include APU's, Chipsets (with GPUs in them) and discrete.

AMD lost 5 percentage points over the last year from 25.2% down to 20.2% while Nvidia gained nearly 3 percentage points from 15.1% to 18%.
 
The public Peddie numbers make it hard to make direct conclusions about individual market segments. That's what Nvidia was complaining about in that link of yours. I assume the juicy numbers are in their for-pay report.

As for quality of data: these kind of reports have numbers from the field, which happens to be quite a bit easier to get right than forward looking speculations. All you need are the right connections with manufacturers, distributors, etc and do a lot of leg work to gather all the information. You'd need to be extraordinary incompetent get this kind of stuff wrong, especially if you've been doing it for 10+ years and people have been willing to pay for it.

A slide from 2008, an article that confirms that the Mercury and Peddie numbers point to the same conclusion, and a post with copyrighted data from a $x000 report on an amateur stock chatroom. That's all you were able to come up with? :D
 
I find the numbers for 2013Q1 interesting in that Nvidia's market share (discrete only) at 18% nearly equals AMD's market share of 20.2% which include APU's, Chipsets (with GPUs in them) and discrete.

AMD lost 5 percentage points over the last year from 25.2% down to 20.2% while Nvidia gained nearly 3 percentage points from 15.1% to 18%.
Yes. The Peddie numbers are especially worrisome wrt the position of AMD vs Intel. It doesn't seem as if their APU strategy is working very well.
 

That's not even close to what the original quote says

Since we're working on quality rather than frame rate that's probably one of the major reasons why we're out talking to people at all. And people don't get to really understand it then effectively we're kind of falling behind on the space race for frame rate; that's another major reason why we're talking.

To paraphrase, "We're talking about smoothness because everyone is focused on frame rate"
 
Frame rate is also extremely important and is one of the characteristics of good user experience. Quality and frame rate should be very closely connected.
High frame rate is a necessary but not sufficient condition.

Either way: 'Nvidia falling behind up on high frame rate' sounds like a link bait title. Looks like it worked with the target audience.
 
Last edited by a moderator:
NVIDIA is changing their business model.
They'll start by licensing out Kepler-architecture and give licensees the necessary designs, tools and support to scale it from smartphone to supercomputer level.

What are the chances of AMD licensing NVIDIA's GPUs? I think the discrete GPU market is shrinking rapidly so I wonder if there's need for both the AMD and NVIDIA to invest in designing new chips.
 
Status
Not open for further replies.
Back
Top