Chances for Nvidia Chipset again?

OH Christ, then Nvdia is Fxxked then.........
No, CPU , No Chipset business, Main Stream GPU is falling. Tegra is not getting anywhere.
The only thing that is left are their Workstation business. Which will then turn themselves into another SGI.

And the worst thing is Nvidia 's management team dont seem to be worry at all.
 
why not push Via nano processors, especially dual core ones with the ION brand?, on ITX.
they can brand the 320M as ION 3 and call it a done game.

people are buying those overpriced ION boards already, you can get buy with a few vendors, one or two models per vendor, mass produce and just sell it.
hell why not go radical and sell it as the trouble-free, graphics enabled linux computer for multimedia and storage needs in the living room.

they are the vendor that just fsck'ing works on the graphics side of things already, so they could have a hardware platform that just works. On ubuntu, as the one particular OS that people may know about.

Have Tegra be able to just run regular ubuntu as well, sell 10" and 11.6" tegra laptops, 13.3" and 15.6" nano laptops. Why not do it?

Bring it on. refreshing, trouble-free, and not the "people don't want a real OS" mantra from Google which I don't think will be any more successful in the open than Windows CE.
 
why not push Via nano processors, especially dual core ones with the ION brand?, on ITX.
they can brand the 320M as ION 3 and call it a done game.

people are buying those overpriced ION boards already, you can get buy with a few vendors, one or two models per vendor, mass produce and just sell it.
hell why not go radical and sell it as the trouble-free, graphics enabled linux computer for multimedia and storage needs in the living room.

they are the vendor that just fsck'ing works on the graphics side of things already, so they could have a hardware platform that just works. On ubuntu, as the one particular OS that people may know about.

Have Tegra be able to just run regular ubuntu as well, sell 10" and 11.6" tegra laptops, 13.3" and 15.6" nano laptops. Why not do it?

Because from what we know, Nano + 320M doesn't look very good compared to Ontario. And I have a feeling that once Ontario actually launches, Nano + 320M will look downright shitty.

And of course, there's Atom as the lower-power and cheaper solution, and Llano/Sandy-Bridge right above, as the pricier but much more powerful option.

And don't forget that Nano + 320M is really Nano + 320M + Southbridge. A 3-chip solution is much less than ideal in that market segment.
 
I said this as soon as AMD merged with ATI. Nvidia has no place to go with Intel on one side and AMD on the other, both eating towards the middle where Nvidia used to keep their lunch. Nvidia is between a rock and a hard place there, hence their attempts to try and create a new market for themselves via the HPC arena.
 
Yes, that and Tegra. It's the only way to offset the loss of the IGP business and the dramatic shrinkage of the value/mainstream discrete GPU market. NVIDIA needs a lot of revenue to pay for the substantial fixed costs they have, and Tegra is the way to do that.

Then there's Quadro/Tesla, which is meant to bring in some revenue, but mostly, profits.

Of course, this is all well and good, but it only works if Tegra and Tesla actually sell. And so far…
 
Yes, that and Tegra. It's the only way to offset the loss of the IGP business and the dramatic shrinkage of the value/mainstream discrete GPU market. NVIDIA needs a lot of revenue to pay for the substantial fixed costs they have, and Tegra is the way to do that.

Then there's Quadro/Tesla, which is meant to bring in some revenue, but mostly, profits.

Of course, this is all well and good, but it only works if Tegra and Tesla actually sell. And so far…
Yep, so far Tegra has only lost them money
 
Yes, that and Tegra. It's the only way to offset the loss of the IGP business and the dramatic shrinkage of the value/mainstream discrete GPU market. NVIDIA needs a lot of revenue to pay for the substantial fixed costs they have, and Tegra is the way to do that.

Then there's Quadro/Tesla, which is meant to bring in some revenue, but mostly, profits.

Of course, this is all well and good, but it only works if Tegra and Tesla actually sell. And so far…


Yep, there was a lot of hype about Tegra, but all those design wins have come to little. There was an article I read, may have been SemiAccurate that pointed out that design wins just mean you get picked to go forward and become one of the finalists for the real contract. As Tegra has blown out it's power budget in an industry that is highly sensitive to power requirements, it's not getting as far as actually being built into any commercial products.

Lots of expense and marketing spin, but no actual products.
 
well it's probably a failure for phones, game boys and ipods but if you check the news there are a number of androids tablets (may I call them tabletoids?), and even at last something about laptops.

(first one is some old prototype..)
http://www.engadget.com/2010/01/08/asus-pegatron-neo-with-tegra-2-hands-on/
http://www.i4u.com/36491/toshiba-ac100-android-netbook-launched

I'll admit I don't know about how works Android ; it certainly looks weird and crude ran as a desktop OS that way. I was ranting about Chrome OS.

I wonder if they'll do a thin-ish, 11.6" one with looong battery life and a sunlight readable dual mode screen. I feel dumb for owning and not using a 9" thick unreadable under the sun netbook. would wipe the OS for any gnu/linux variant.
 
Nvidia to build a X86 CPU ?

http://www.semiaccurate.com/2010/08/17/details-emerge-about-nvidias-x86-cpu/

Overlooking the usual charlie bias , do you guys think they have a shot ?

there are some interesting points in the comments , (charlie didn't mention them) I wonder if they are true ..

-Many x86 patents are expiring soon.

-FTC defines, "Compatible x86 Microprocessor" as, -in part iii- "that is substantially binary compatible with an Intel x86 Microprocessor without using non-native execution such as emulation...

Which means that if the chip is using emulation to run x86 code, then it is NOT a "Compatible x86 Microprocessor", so transmeta and nvidia are off the hook (at least in as much as this paragraph is concerned)


-
 
Because from what we know, Nano + 320M doesn't look very good compared to Ontario. And I have a feeling that once Ontario actually launches, Nano + 320M will look downright shitty.

And of course, there's Atom as the lower-power and cheaper solution, and Llano/Sandy-Bridge right above, as the pricier but much more powerful option.

And don't forget that Nano + 320M is really Nano + 320M + Southbridge. A 3-chip solution is much less than ideal in that market segment.

Well the only info we have on Ontario's performance till now is that "It will have 90% of today's mainstream performance at less than half the size". So assuming that mainstream is Athlon II, performance should be good. However im extremely doubtful they'll be able to achieve their power targets(sub 1 watt capable they claim). This is where Nano could do well, on a modern process it could offer reasonable performance at low power consumption(note than till now, the nano has been generally built on older processes. afaik its still on 65nm)

And 320M is a single chip NB+SB solution, just like the Geforce 9400M.
 
http://www.digitimes.com/news/a20100901PD215.html

nVidia is planning to integrate gfx chips to southbridges in future, and is hoping for motherboard manufacturers to get sandybridge platforms without Intel southbridge and use nVidia southbridge instead.
Cutting the Intel SB would save around $10-15, but the cost of nV SB is unknown
 
lol, I just realized SB could stand for both Sandy Bridge and South Bridge :LOL:

So the "SBs" in my post mean southbridges, of course.

On the details how this could be done, do they have lisence etc, I don't know, but apparently the FTC rulings should give them a chance to do this
 
Riiiight, I'm sure that's going to work out just fine.
Sounds crazy indeed. Unless they are going to outfit this with dedicated memory (and 64bit ddr3 might not be enough) there's no way it's going to compete with SB graphics. And using dedicated memory should drive costs up significantly, which makes the cost savings rather dubious.
 
Back
Top