Nvidia Tegra 3 Under Pressure After Tegra 2 Concerns

It doesn't help that Kal'el isn't even 28nm. It's supposed to use the old fab process we have had for 2 years. I know I'm waiting till Wayne comes out to get a tablet.

But can it play Crysis ?

You might not know...
There is always a chance.

"It doesn't help that Kal'el isn't even 28nm. It's supposed to use the old fab process we have had for 2 years. I know I'm waiting till Wayne comes out to get a tablet."

Really ??? mmmmmmmmmmmmmmmmmbulshit!!! =(

UBS's who exactly? Just because *some* unknown analyst sitting behind a desk playing paper football makes a random prediction, we're supposed to sell off (investors), or hold off buying a tablet (consumers)?

He's a nobody. And nobody will remember his name 5 minutes after Nvidia's Q4 earnings report beats his "expectations". Mean time, I will reap the profits when all the sheep who sold out and lowered the value of the stock on this guy's word try to buy back in when it shoots back up. Yeah, buy high and sell low, fools.

If you are really so desperate to make stupid comments in hopes to gain favor, try badmouthing Apple; you will have much more success.

Yes it can, it has very similar performance to Xbox360.

The real question, can it play Metro 2033. The answer........ No, and who cares.

can it play BF3?

The real question is: can you people posting "can it play.." go play something and stop posting retarded comments

Huh? What he says doesn't make sense to you? It's quite logical.

can it play....
gotcha!
anyway, nvidia's discreet market gain is good news for the high end gfx line. but the real money is in the low sub $90-$100 segment and nvidia has nothing there. amd rules that part. if nvidia hadn't screwed up fermi it would have been the leader.
nvidia's higher end discreet cards are quite powerful but they're not exactly well rounded. for example all of amd's 6790+ cards are 256 bit, all the way to 6990. but nvidia has up to 384 bit cards which trade blows with amd's 256 bit cards. that's not really a good sign of efficiency.
nvidia's new 28 nm gpus seem to be released much later than amd's. i wonder why tsmc could volume produce amd's 28 nm gpus but delay(have problems with) nvidia's.
nvidia's optimus technology in notebooks is much better than amd's switchable whatever thingie. they also have much better driver support.
amd's apus and intel's ever-improving igpus are taking away one of nvidia's biggest top earners imo. android's lack of sales doesn't help anything either.
bottom line is: even if tegra 3 is better than tegra 2 nvidia will have a problem on their hands (no the top few percent high end dgpu market won't help them..much).

The report is not well written, no sources, just need to fill the quote of postings?

You wish that were true. Intel still rules the low end. 60% i believe is the figure. You remember Intel, the company that everyone knows about, including every AMD customer. Same goes for Microsoft (vs Linux and Mac)... and nVIDIA (versus AMD in the discrete category). nVIDIA desktop GPU marketshare went up 10%. nVIDIA cant live on graphics cards alone, but they are still the big dog.

btw, AMD's Fusion APU's are eating into its own discrete graphic sales. Intel and nVIDIA both don't have that problem.

Tegra 2 is old shit that should have been phased out 6 months ago! And tablets over $300 - 400 are pure stupidity.

i guess i wasn't clear enough when i said that. i was refering to lower end discreet gfx market. amd has come out with new 6xxx series low end card while nvidia has done nothing. amd's got new 6450, 6570, 6670 and 6770(and some other oem ones) they all are slightly improved version of their 5xxx counterparts . meanwhile nvidia has gt 430, gt 440, gts 450 and gt 550ti (and may be gt520something) against those cards. except 550ti all of them are older models and have higher price but lower performance. and 550ti is priced too high for its performance.
and then there are amd's own desktop apus that have discreet class igps. nvidia has no equivalent there.
in q3, intel does lead Overall gfx market with 60.4% with amd in second 23%, nvidia last 16.1%. nvidia's share in dgfx was up more than 10%.

amd's success with fusion apus is largely in the mobile(laptop, ultraportable etc) market. while the desktop apus are very capable, pc companies keep building pcs with low end dgfx. i've seen plenty of pcs with core i5 or i7 or pentium with radeon hd 64xx, 65xx or 66xx cards e.g. this one or this one. those gpus still have a market there. fusion is slowly creeping into low end dgfx market but amd is handling it in a good way. they offer asymmetric cfx with llano igp and low end dgfx, combine them and deliver better performance(for gfx capability at that price). besides, as long as amd is facing production problem and goes through the changes right now, amd's low end cards will stay a little longer.
nvidia otoh has nothing there.their loss is because they're getting out of igp market, has no newer model entry level cards, too focussed on their socs, android, tegra etc. their low end cards are old, pricey and plain bad.
intel has none of the problem of amd or nvidia. they simply put igp in their cpus and sell them.

Once Trinity and Ivy Bridge desktop CPUs with their improved on-die GPUs come out next spring, I think a lot of the low-end discrete GPUs will see their sales dry up. Trinity's GPU is supposed to be something like 30% more performance than Llano's, and IB's is supposed to be 60% more than SB's.

However I'm personally waiting for those 28nm high-end GPUs for my next DT build . Not too many gamers enthusiastic about iGPUs that I know of..

Which is precisely why you should start your own company, develop a chip better than the Tegra 2 and then release it in a tablet for less than $300 - $400. You'll be rich. Let me know when its available.

I wonder if Tegra 2 lower sales has anything to do with their announcement of tegra 3 before Tegra 2 was even available for purchase. I was going to get a tegra 2 tablet, and then they said tegra 3 was going to be available only a few months later, so I decided to wait.

wth????

you really have no clue. Your posting pure BS. First off the 384bit is only related to the memory systems. Nvidia (by design) uses a wider bus than AMD. Its not a drawback at all, it actually is very useful. Fermi is great as a GPU but Nvidia makes much more money in their tesla brands. AMDs 256bit memory bus is sufficient for their design, Nvidia uses a wider memory bus for theirs. To put it in terms that are easier to understand. AMDs 256 bus is like dual channel ram while Nvidia's 384bit bus is more equivalent to intels triple channel systems. See not at all a draw back to have a wider memory bus, it can actually be very useful.

Hopefully you have a better understanding of it now so in the future you wont sound so foolish!

i was using the memory bus width only for reference, hence the 'for example part'. other factors play into gfx performance as well but memory bus width is an important factor for calculating memory bandwidth - that's why i singled it out. with memory bandwidth i can compare two different cards more easily instead of worrying about shaders, clocks or rops stuff like that.
again, for reference only, here is nvidia's newest 192-bit 550ti vs amd's old 128 bit 5770 comparison:
http://www.tomshardware.com/review [...] 92-16.html
http://www.tomshardware.com/review [...] 92-17.html
550 ti trades blows with the 5770 despite having higher memory bus width and costing more. both nvidia and amd have their strong and weak points.
besides, this wasn't even the point i was making, i was talking about how oems losing interest in nvidia's new gpus can hurt its business.
fermi is...well, it's something nvidia brought out, took them a second try to polish it a bit more but it still wasn't good enough. it has raw power but no finesse. but that's just my opinion.
nvidia may make more money with their tesla brand but they are still lagging in the overall gfx market behind intel and amd. their recent gain in the discreet gfx market is offset by that fact.

Tegra2 has two problems- no NEON, and no reliable playback for most HD (720 or 1080) video. Both are killers. Every other new single or dual core SoC used today have both. Tegra2 was Nvidia's fix, but the chip is way late, and way over engineered (4-cores, when 2-cores are all most OEMs need at the moment, for cost and programming reasons).

Tegra2 was advanced for its time, and still makes for a great tablet, if you don't care about HD video. Tegra3 is going to be advanced for its time too, but a poor fit for the tablets most companies want to build today. By the time companies wish to use Tegra3, better choices will exist from Nvidia's competitors.

Nvidia is out of sync with the market, plain and simple. Also Nvidia, for a company with a killer rep for graphics on the desktop, delivers decidedly average GPU performance on the tablet- a clear embarrassment. Nvidia should have had a Tegra2+, with NEON and better HD decode, selling a year ago. Nvidia bet the bank on TSMC having its new 28nm process ready for 1H2011, and this was a gigantic mistake.

Nvidia's ARM competitors are not bleeding edge, like Intel and AMD are. The ARM SoC marketplace is all about cost, and the right features at the right time. Nvidia is too used to paying shill tech sites to trumpet PR rubbish that hides its shortcomings in the PC space. Who would build a tablet today using Tegra2? No-one, because the video weakness is completely unacceptable in any new tablet, regardless of cost.

I think when it comes to arm gpus noone does it better than imagination's power vr especially not nvidia!!!

After you visit the link you can remove your foot from your mouth. UBS doesn't make random predictions and is a very well known name. Since you didn't seem to understand what was said in the article this is the short translation: Nvidia's revenue prediction was lowered due to the high amount of tablets in the market. A basic understanding of supply and demand is all that is required to understand why this would lower their revenue outlook.

This article didn't say anything negative regarding the Tegra 3 chip nor the Asus Transformer Prime that is to be released in a few days. This is purely from a business standpoint to help investors decide how to invest their money.

http://www.ubs.com/1/e/about/ourprofile.html

Link didn't show up

It can play UT3engine for sure.
An Asus transformer with Tegra3 will replace my current Asus eee 1204N pc. But it would be nice to run Ubuntu on it.

Its such a strange thing to use as a measure of metric. Its a manipulation for your agenda. I could easily say AMD uses many more shaders than Nvidia therefore "that's not really a good sign of efficiency" as you put it or "cards are quite powerful but they're not exactly well rounded" They use 800 steam processors (5770) to match a mere 192 from Nvidia (550ti). I could go on and on to manipulate for an agenda. But its useless other than that. Why compare these total radically different architectures this way You calling 384bit a drawback is absolutely absurd. Its nothing to complain about, its all good! You compare the value, and although your all around opinion of Nvidia isnt high, they are doing much better than their competitor in the dGPU market they are currently selling in.

And to be clear, per individual sale Nvidia makes a huge profit off their tesla products. But its a much lower volume item. They make twice the revenue from their Gforce line. This is their current bread and butter. They are doing very very well here. They make much more money than AMDs GPU division. Why is this? Does AMD sell their cards at a much lower profit? Its clear Nvidia has higher margins, how do they manage this? You may not like their prices but they are extremely competitive. AMD lately is barely making money in their GPU division, and thats just a sad fact.

As far as lagging in the overall market? Nvidia is no longer in the chip-set business and therefor they have no igp. The overall market intel is the number 1 graphics because of their igp. This doesnt mean much of anything. Nvidia no longer sell igps, they only sell discrete (add-on) GPUs. In this market, Nvidia has actually gained share this quarter. They are nimber one here, in the segment they actually compete it.

So to sum it up. No matter what your opinion is, or how much you disslike Nvidia, They are outselling AMD while making more profits from each sale. Your distaste means little at the end of the day. With the truth at hand, it appears to me your force feeding a negative "opinion" is from a personal distaste based on emotion. Its unfounded in reality. You are entitled to your own opinion and thats fine i dont care to change that. I am only here to make clear of what is real and whats hog wash!

Really? Let's try this then:
Apple really s*cks!!! Tegra3 rocks!!1!!one... There's no way apple can beat this tegra3 sh*t!!!

Nvidia's knight in shining armour is windows 8 running on arm processors. Couple that with a NVgaming zone as a distribution platform, they could start trying to make their own version of desktop PC's.
It'll come down to which is cheaper to manufacture.
Its not gonna be long until Mobile/tablet GPU's surpass current gen console hardware and if the same level of optimisation occurs for tegra devices that went on consoles you could get successful ports with awesome performance. Nvidia saw the squeeze from AMD and Intel and has responded. Only problem is mobile tech tends to be integrated totally meaning it has to be replaced in contrast to desktops which can be upgraded as new tech comes along. If that happens they move into low margin PC/laptop devices and possibly away from their core GPU buisness. I just cant see how tegra relates back to geforce/quatro as well as their GPU tech has helped them design tegra. Unless it allows them to integrate ARM into Geforce but that is no different to using an atom to power a GTX580