• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Radeon HD 7800 Series Inbound for March, NVIDIA Kepler in April: Report

H

hclarkjr

Lifer
http://www.digitimes.com/news/a20120215PD219.html
Monica Chen, Taipei; Joseph Tsai, DIGITIMES [Thursday 16 February 2012]
AMD is set to launch performance-level Radeon HD 7870 and 7850 GPUs in the first half of March, while Nvidia is set to launch 28nm Kepler in April, according to sources from graphics card makers.
Although AMD is also set to launch its highest-end Radeon HD 7990 (New Zealand), there is still no firm launch schedule.
Nvidia is expected to have at least eight models of graphics cards in April including GTX690, GTX660 and GTX640 with GK110, GK104 and GK106 cores.
Since Intel will not introduce Ivy Bridge until after April, when PC replacement may start, Nvidia's trailing behind will not affect the company much and the two players should see their 28nm competition at the fiercest in April.
 
So then is the gtx690 going to be a single gpu card or dual? How are tey going to launch 8 models on the same month?
 
And why would they launch the 690 but not the 680? Unless the 690 is based on the 660 or something?
 
GK110 is GK104x2

I believe they are pulling GK110 ahead of schedule, it was supposed to be Q3 originally. This probably means GK104 is not faster than Tahiti...
 
There's debate over the part name but rumors indicate that GK110 or GK112 is a new part, and it is due in Q3. GK110 could very well be a dual gpu card, that would make GK112 the successor due late this year. GK100 was cancelled.

Everything else in that article seems to have a lot of errors, or completely contradict leaks from other sources
 
lavaheadache said:
April releases? sheesh, there goes all those *wait* recommendations
Click to expand...

With TSMC's 28nm "Yield problems" (according to Nvidia), you will be waiting a lot longer to actually get one.

http://semiaccurate.com/2012/02/15/nvidia-comes-somewhat-clean-on-their-quarterly-conference-call/

Nvidia is the only one making mention of yield problems right now.. Hrm.. and then there was that whole bit of dumping early 28nm capacity.

So, we have Nvidia complaing about yield problems, AND they are short on 28nm capacity.. This makes me think that we all are in for a big shocker --- Nvidia will bring out hampered chips that will have very low supply.
 
Jionix said:
With TSMC's 28nm "Yield problems" (according to Nvidia), you will be waiting a lot longer to actually get one.

http://semiaccurate.com/2012/02/15/nvidia-comes-somewhat-clean-on-their-quarterly-conference-call/

Nvidia is the only one making mention of yield problems right now.. Hrm.. and then there was that whole bit of dumping early 28nm capacity.

So, we have Nvidia complaing about yield problems, AND they are short on 28nm capacity.. This makes me think that we all are in for a big shocker --- Nvidia will bring out hampered chips that will have very low supply.
Click to expand...

It makes me wonder if 28nm yields are low because of TSMC, or if because nVidia is having issues of their own and covering it up by pushing it off onto TSMC.

AMD is shipping chips and has not complained of yields at all. Although we don't know how many they are shipping. But once OEM's start grabbing them up, that rate is going to shoot up pretty quick.
 
Stuka87 said:
It makes me wonder if 28nm yields are low because of TSMC, or if because nVidia is having issues of their own and covering it up by pushing it off onto TSMC.

AMD is shipping chips and has not complained of yields at all. Although we don't know how many they are shipping. But once OEM's start grabbing them up, that rate is going to shoot up pretty quick.
Click to expand...

Well NV had issues that AMD didn't have with regards to 40nm, so potentially something similar could happen again. Not to say that TSMC didn't have 40nm issues in general, since we all know it did, just NV suffered more from them than AMD when it came to 5000 vs 400 series cards.
 
Stuka87 said:
It makes me wonder if 28nm yields are low because of TSMC, or if because nVidia is having issues of their own and covering it up by pushing it off onto TSMC.

AMD is shipping chips and has not complained of yields at all. Although we don't know how many they are shipping. But once OEM's start grabbing them up, that rate is going to shoot up pretty quick.
Click to expand...

AMD is shipping, but they have really staggered this release. Seems to be more staggered than usual anyway, although my memory may not be accurate. It just kinda feels that AMD may also have yield issues, which is why they only have a handful of models available.
 
Same old small die vs big die argument. If recent past is any indication, small die will win round 1 and big die will win round 2.
 
sandorski said:
AMD is shipping, but they have really staggered this release. Seems to be more staggered than usual anyway, although my memory may not be accurate. It just kinda feels that AMD may also have yield issues, which is why they only have a handful of models available.
Click to expand...

Your memory definitely isn't accurate.
4 chips in 6 months reaches its end today, with the launch of the final chip in AMD’s Evergreen stack
Click to expand...
5000 series launch.
 
AMD pushes HD 7950 graphics card launch to February


It's a nice fantasy to believe that one company is having more trouble than another, when they are both using a 3rd party fab.
People have short memories.
CHIP DESIGNER AMD has decided to delay the launch of its HD 7950 graphics card to early February.This comes after news that the HD 7970, which AMD 'launched' on 22 December, will not be available until 9 January. This wasn't received too well by the media and customers as it does help to have some stock to sell on launch day.To avoid another bashing for a 'paper launch'


​
Click to expand...
 
I question the validity of the information in this article. Nvidia launching 3 new chips, essentially 8-9 new products all within a month time frame???? That would be a logistical hell to pull off with any sort of decent organization and coordination.

But even if true, I don't the launching of any one particular product gives heed to how any of these said products will perform.
 
tviceman said:
I question the validity of the information in this article. Nvidia launching 3 new chips, essentially 8-9 new products all within a month time frame???? That would be a logistical hell to pull off with any sort of decent organization and coordination.

But even if true, I don't the launching of any one particular product gives heed to how any of these said products will perform.
Click to expand...

Perhaps,

690 = gk104 x2

The rest will be gk104 derivatives.
 
Is it at all possible if they release so many SKUs that maybe only 2 or 3 are actually discrete desktop and the others are mobile parts and/or HPC?

With the APUs eating up the mobile sector, I'd picture a frenzy to stay more active in that sector, also the giant margins in the HPC sector.
 
Perhaps Nvidia is switching to the "9" for top single GPU nomenclature as did AMD going from 5870 to 6970 for single GPU high end.

GTX580 to GTX690.
 
railven said:
Is it at all possible if they release so many SKUs that maybe only 2 or 3 are actually discrete desktop and the others are mobile parts and/or HPC?
Click to expand...

I believe the article was specifically talking about desktop gaming-oriented GPU's. But it is entirely possible the article was covering more than just desktop consumer cards.


Keysplayr said:
Perhaps Nvidia is switching to the "9" for top single GPU nomenclature as did AMD going from 5870 to 6970 for single GPU high end.

GTX580 to GTX690.
Click to expand...

This is entirely possible. But I think they should skip the 6 and 7 series altogether, and jump up to 8. That would really throw a curve ball at everyone.
 
bryanW1995 said:
Same old small die vs big die argument. If recent past is any indication, small die will win round 1 and big die will win round 2.
Click to expand...
The GTX 580 is quite a bit faster than an HD 6970, but it's also proportionally more expensive. The gap between the GTX 590 and HD 6990 is smaller because in the end TDP rules the designs of these cards. Perhaps we'll see 3-slot reference designs in the future...

Anyway the halo tier is fairly irrelevant. The biggest market is OEMs, specifically portables. In this space one of the biggest movers is Apple and not a single Mac product currently uses Nvidia graphics. There are rumors swirling that Apple's Ivy Bridge updates will carry Nvidia graphics, but take it with a grain of salt for now.

A quick Newegg search returns:

110 laptops w/Radeon graphics
98 laptops w/GeForce graphics
92 dekstops w/Radeon graphics
62 desktops w/GeForce graphics

The popularity of Radeon graphics on desktop systems is no doubt because of Fusion APUs. I'm not saying small die vs. big die isn't important. It's very important...it's why the GTX 460 (GF104 332mm^2) so quickly replaced the GTX 465 (GF100 529mm^2). It's also why the GeForce 560 Ti 448 exists...because big dies have lower yields and Nvidia needed to do something with cut GPUs that didn't make the grade as 570s or 580s. This is a bigger problem when you pay TSMC per wafer and not for the yield.
 
HALO tier is never irrelevant, volume (revenue) shipments are low end OEM and mainstream cards, but high profit (big contribution towards net profit) margins have always been in the performance and enthusiast segment.

Market surveys of recent times (taking into account IGPU destroying low-end discrete) have the best selling cards are in the $300 range.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top