When I hear the name of nVidia’s next-gen GPU code named Kepler all I can hear is Cave Johnson from Portal 2 saying, “Now if you're part of control group Kepler-7, we planted a tiny microchip, about the size of a postcard, into your skull.” It is interesting in that Kepler is supposed to be nVidia’s 28nm GPU offering. True a 28nm GPU is a lot smaller than a post card but the other half of Cave Johnson’s statement does seem to have some very nVidia parallels… “Most likely you've forgotten it's even there, but if it starts vibrating and beeping during this next test let us know, because that means it's about to hit about 500 degrees so we're gonna need to go ahead and get that out of you pretty fast.”
Now, I am saying this with humor, but we have seen some rather hot GPUs from nVidia over the years. If Kepler has the same issues that Fermi had then the first run could be quite the space heater. Of course, if you ask the typical gamer they will more than likely tell you that they do not care as much about the heat as long as the FPS is through the roof.
Which brings us to the real crux of the issue, nVidia has already admitted that they will be at least a month or more behind their original timeline for this release (like Fermi) and they will be far behind AMD’s Southern Islands. In the gaming and GPU market it is often the person that gets to market first that wins the acceptance. This is not like the DX11 shift when it was Fermi Vs the Other Islands from AMD where there were almost no compelling games using that technology. Now we are looking at a few good DX11 games so we are going to see people looking to pounce on the next generation products as long as they are faster than what is currently on the market from both companies and the prices are right.
nVidia really needs to get on the ball to get their timelines back in order or they do stand to lose quite a few Holiday sales to AMD which will hurt their own sales numbers when they finally do get Kepler out the door.
Discuss this in our Forum
CES 2012 Las Vegas, Nevada – Just when you thought our CES coverage was over we bring you our coverage of AMD and a few of the products that they wanted to show off during CES. AMD’s booth this year was somewhat out of the way. Although many companies choose to be on the main floor of the show AMD was up in what almost seemed like a back corner in the meeting room area. This did not mean that they did not bring out some cool stuff. It just seemed like they really were not looking to deal with the madness of the show floor.
It seems that Apple and Qualcomm wanted to try and lock out competition when it comes to the use of TSMC (The Taiwanese Semi-Conductor Manufacturing Corporation). The pair attempted to buy themselves a nice exclusive deal with TSMC with a large bundle of cash (over $1 billion). However, TSMC decided that there was more money to be had in keeping things open. The original deal from Qualcomm and Apple was to make sure the pair had access to production facilities. TSMC replied by saying that they could always expand if there was a need, but that they would internally fund it.
If you were looking for any additional confirmation that AMD is reorganizing for a major rally then the announcement that they are canceling CPUs before they are launched might be of interest for you. It appears that AMD will be outright cancelling their Krishna and Wichita 28nm CPUs. The stated reason, Global Foundries will not be ready with 28nm in time.
There were some interesting announcements for AMD at the Globalfoundries event today; although it was not AMD that made these, but Greg Bartlett, Senior VP of Technology and Integration Engineering at GF. Gregg was showing off a few slides of upcoming technology that GF will be involved in and let some things out. We are certain these were previously approved by AMD that would like as much press as possible about their new APUs (which appear to be very good).
The things that were let to slip were the fact that AMD’s next generation Trinity will be 32nm instead of 22nm or even 28nm. The new Trinity will be based on a derivative of the Bulldozer core architecture and, according to AMD, will be roughly 50% faster than the current Llano APUs. The new APUs will feature AMD’s next generation 28nm HD7000 series graphics (which was another confirmation although not a formal announcement). The trinity parts are slated to be pushed towards the mobile market (where AMD seriously needs a better offering) and should begin shipping in early 2012.
It will be interesting to see AMD and Intel compete for this new market space. We have already watched some entertaining Facebook posts from both sides about this very issue.
Discuss in our Forum
According to a few articles on the internet and a few things that are rumbling along the “usual lines” nVidia could already have working 28nm Silicon. Although no one seems to know what this silicon is a good guess would be Tegra. Jen-Hsun Huang, nVidia’s CEO, has been quotes are saying they have an entire team working to make this move much smoother than the one to 40nm.
If you were not around for that one let’s just say it was a mess and nVidia lost a lot of money due to bad yields and other issues that were spawned by TSMC (Taiwanese Silicon Manufacturing Company) and their own rough transition to this die size. However, if you remember during this troubling time for both nVidia and TSMC there was an announcement that TSMC had already begun work moving to 28nm.
This announcement was made by TSMC around September/October which would make the timing for this about right. So, if TSMC has gotten their 28nm process right and nVidia is truly moving Tegra to 28nm first (and has working silicon) then the logical guess for this working Silicon is Tegra. The question now is which Tegra? It is highly unlikely to be Kal-El, but considering the claims that Jen-Hsun is making, a quad core ARM based SoC (System on Chip) that uses less power than the current dual core, we have to wonder. This is speculation of course but we do have some evidence for this supposition. It would also seem to fit the delays of this new mobile chip that we have heard about (now they are saying Holiday Season). It would not be beyond the realm of probability to see these SoCs ready for December and products out in January/February time frame which again could mean the Kal-El will show up as 28nm and not 40nm… But to be honest it is VERY unlikely.
Discuss in our Forum
AMD has begun shipping their 28nm GPUs to OEMs in preparation for the official launch in January at the 2012 CES. This news is good for AMD as they have been promising to be first to market with 28nm for most of the year. It is also good to know that some of the issues with TSMC’s 28nm process that we have heard about are not going to slow things up.
With all the news about the major ARM players like Samsung, nVidia, Texas Instruments etc. It is nice to finally hear the news about one of the largest, yet least talked about. At one time if you bought a cell phone (this was before the Smart Phone age) you had Qualcomm under the hood (and in many cases AMD as well). However, lately all we hear about is Hummingbird and Tegra (with an occasional mention of Snapdragon). That is until yesterday.
Qualcomm has announced that they are prepping to launch a new series of 28nm processors which they claim are 50% faster than “other ARM based offerings in the mobile CPU market”. This is very healthy boast to say the least. Right now all of this is in the pre-production stages so we do not have any phones or designs to tell you about, but we do have a little bit of information about the processor itself.
The new proc will be dubbed the S4 MSM 8960 and will hit the market as a dual core offering with a quad core to follow. Instead of an on/off switch for power the CPU is going to be built with something like a dial. This will allow for dynamic power steps to be sent to the processor. I have not heard of this being done quite this way so it should be interesting to see how it works out.
Inside the 8960 you will have a new GPU, this is the Adreno 225 which is a unified shader design (think AMD/ATi) and will support DX9.x and possibly shader model 3. Now we wonder what kind of radio they are going to build into this as that has been one of Qualcomm’s strengths (the Cellular and Wireless Radio built into the SoC) and something that Samsung and nVidia have been trying to get done as well. We know that nVidia has been buying up wireless technology patents with the hopes integrating their own radios into their Tegra SoC, but have not announced anything on that just yet.
For now it would seem that Qualcomm has taken a step out front in the performance arena. We do think that their GPU offering is probably not up to the same standard as nVidia’s Tegra, but then again many people looking to buy a phone do not need it to play graphics heavy games… I mean really, how much GPU does Angry Birds or words with friends use?
Discuss in our Forums