TweakTown NewsRefine News by Category:
We all know that Pat Gelsinger is leaving Intel for EMC. This was an announcement made the other day and on the surface did not seem to be much of a big deal. But, as with many things what is in a press release is often not even close to the real truth behind the matter.
If sources are to be believed Pat left Intel not for greener pastures, but because he failed to deliver Larrabee. This corresponds with other rumors we heard about techs leaving for failing to product a working Larrabee by the stated deadline.
Of course the issue of Larrabee being delayed would not be so bad if it was not for the CPU Vs. GPU war between Intel and nVidia. This little tiff brought attention to Larrabee for what it was supposed to be, and also for what it was not, working.
Of course as with any rumor or confidential source all of this must be taken with a grain of salt. Still there are two things we know; Larrabee is nowhere near ready and Pat Gelsinger has left Intel for EMC.
Makes you wonder doesn't it?
If you are tired of heating how nVidia has pushing Cuda support into everything then I have some good news for you. The guys over at AMD have penned a deal with Pixelux. They guys are the one that developed the Digital Molecular Matter engine they have (and still are) a key part of the way Lucas Film and Lucas Arts create effects.
Recently Pixelux was allowed to release an end user plug-in version of DMM for Maya that allows pretty much anyone (that can afford and use Maya) to create effects the same way the Pixelux does.
So what does this have to do with AMD? Well you see AMD in their usual fashion has been pushing an open platform; this is OpenCL (Cuda is also OpenCL just nVidia's version). Well Pixelux is now going to work with AMD to develop OpenCL acceleration for their DMM engine.
This is a very cool development for AMD. Normally they end up taking a back seat to the much more vocal companies like Intel and nVidia. This time they are the ones getting the support for an industry standard rendering application and one that could lead to more.
There has been a lot of talk about Apple's Snow Leopard. Some good, some bad but most of the talk has been in the middle. One of the things that is good in Snow Leopard but is not generally talked about is Grand Central Dispatch.
This handy little piece of software enables the OS to efficiently utilize multiple cores. Now OSX has always been able to handle SMP (Symmetric Multi-Processing) due to its BSD roots but it has had an issue dealing with multiple threaded applications. Yes it could handle them but not always the most efficiently.
Bring in Grand Central Dispatch to fix this problem. It allows for a very effective method for dealing with multi-threaded applications and allows OSX to finally take full advantage of Nehalem's Hyper Threading.
So why am I tell you all of this? Well because Apple has done something with GCD that they usually do not do. They have pushed it out to open source. This means that anyone can use the code base to develop applications that can take advantage of this new technology.
That is a pretty cool thing and not a normal Apple move. Maybe they are opening up finally...Nah probably not.
I think it is time for another award around here. This one goes to Steve Jobs. That is right, the master mind behind Apple. Apple is arguably one of the most successful marketing companies on the face of the planet. They could hype a pair of old shoe laces into something revolutionary and spectacular. They know the market they are playing to and know how to cater to their frenzy.
So when Steve made the comment "Originally, we weren't exactly sure how to market the Touch." many people thought he was joking or simply covering something up. I do not think ANYTHING at Apple is ever even considered without knowing how market it and exactly who it is marketed to. The comment came in response to questions about why a camera was not included with the iPod Touch. Apple chose to shove a low end camera into the iPod Nano instead.
Jobs said in the same comment "What happened was, what customers told us was, they started to see it as a game machine." I am not sure how this is true when the iPhone was not see that way until the 3G S. Granted the touch has all of that now.
Still this week's award for most foolish (and obvious) inconsistent statement goes to Steve Jobs.
Ok on the rumor meter we are keeping track of the ones that come true and the ones that do not.
For this update we have;
No new iPod Touch (delayed launch) - could be due to camera issues reported earlier
New iPod touch is out but not with the camera mentioned. So far it is a small update (using similar hardware to the iPhone 3G S) prices will shift down a little 8GB will be $199 32GB $299 and 64GB will be $399.
iPod Nano - Getting bigger screen and Nike + improvements not larger capacity
iPod Nano is getting a camera and microphone (on the 8GB version) instead of the iPod Touch, and FINALLY FM Raido. Of course the Zune HD will have HD radio so the Nano is going to be behind in that regard. The Nano did indeed get an update to Nike +, and also has a voice over control.
HD video to the iPhone - Rumor is help up by Harmon Kardon listing HD output as an option on one of their dock products.
Nothing on this yet.
iPod Classic - Getting a 160GB flavor
True 160GB will come out at $249.99 spot
As the camera was the "One more thing" I have serious doubts that the iPod Touch will get a camera any time soon. We will have more as things develop.
Hmmm, methinks AMD might have a new marketing strategy. So think on this...AMD cannot compete one on one with Intel. If you put them up clock for clock there is very little chance that an AMD CPU will beat the correspondingly clocked Intel CPU.
So if you are AMD what do you do? Well you start by pricing your products lower than the other guy and say that Dollar for Dollar you have the better product. Now that marketing strategy is great, but really will only last so long. So what else can you do?
To put it simply to take CPUs with Cores and Cache that do not meet full QC. This is not to day they are bad, just that they do not meet full QC. You market these as less expensive products so as to mitigate the normal loss you would have with these "failed" parts.
Now this is a good enough solution as you can sell them for less to the consumer but it is not enough to combat Intel and its dominance. To do that you need an extra...edge. This comes in the form of small BIOS tweaks that magically turn on extra cores and cache that is disabled. Suddenly that X3 you bought is a Quad and that Cacheless Athlon II X4 is a Phenom II X4. Not a bad move at all, but a tad sneaky.
Still when you are dealing with a performance powerhouse like Intel you need all the help you can get. You are still not beating them clock for clock but you are able out price them in a very underhanded way.
The future looks bright for ATi, Microsoft and DX11. Windows 7 is receiving some of the best press that MS has received in the last 5-7 years. This is a great comeback after the technical press and bloggers condemned Windows Vista before it even hit the shelves. Of course this has all changed.
DX11 is taking off much faster than DX10 did and we have even seen some game developers postpone games to recode their engines to take advantage of the new DX. We know of roughly 4 games that should be DX11 and available at the time that Windows 7 Launches.
Two of these titles Grid and DiRT are popular games that should help to bolster sales of Windows 7 for gaming. But on the opposite side we have ATi already showing off DX11 hardware and with highly anticipated titles (like Grid and DiRT). Meanwhile nVidia might show off DX11 capable hardware in September. ATi has already promised DX11 hardware on the shelves for the Windows 7 Launch.
We also see faster driver support from hardware manufacturers. At the same time in Vista's life many hardware companies were still waiting and reluctant to port their existing drivers over. Not so with Windows 7. Many products are already shipping with working and stable drivers for the new OS despite it not actually available for purchase yet.
All of this spells a very good holiday season for these three companies as tech enthusiasts scramble to get DX11 hardware, games and OS as quickly as possible. Unless nVidia can get something out in working form very soon they might lose some of that money this year and have to wait for next.
It has been awhile since we have given out our version of the Darwin award. Well if you have been wondering who the lucky recipient is this time you can let out your long held breath (Cyanotic Blue is really not your color) our winner is none other than Apple.
Yes Apple, after the news of their new Malware application in Snow Leopard got out (something that Apple Ads say they do not need) we now have a picture from Gizmodo that belies the "Macs Don't Crash" claim.
This is a picture inside an Apple Store. The difference is that the screens that smugly announce who is next in line to talk to a genius are showing a Kernel Panic. This (as they say in the guide) is completely impossible. But the impossible seems to have happened.
I would love to see this in an ad somewhere
The Foundry race jut heated up; TSMC just announced the production of 64MB SRAM on three 28nm process nodes. This roughly matches the same timeline that Global Foundries is looking to keep.
So far the time line puts the three nodes up and running in risk production by Q3. The Low Power process should be in risk production by the end of Q1 2010 while the High-K metal gate low power and performance versions will be out in Q1 and Q3 respectively.
TSMC says that they have managed to get their High-K Metal Gate Process going by using a "gate last" method. They claim this will make theirs superior in terms of transistor characteristics, performance and manufacturability.
I am not sure what the results of this will be but I am sure the bidding for nVidia's business will be very interesting for the 28nm process next year.
nVidia is going after AMD and their "Dragon" platform. To do this they are leveraging the new Intel P55 Chipset.
You see they really do not have a complete platform solution like AMD did and now they are rapidly building one.
Referred to as "Power of 3" in the PR Slides that have gotten out into the public eye, nVidia claims that P55 with built in SLI support is perfect for their new SLI + PhysX design.
They start off with a mainstream board that has 1 full x16 PCI-e 2.0 slot and a 2nd x4 electrical slot for an additional PhysX enabled GPU. The next level is the performance level. This has at least two x16 physical slots that can adjust to x8 each with SLI. The board should also have one additional x4 electrical slot for a PhysX enabled card.
The top tier is the Extreme, this will have up to 3 x16 PCI-e Slots with full x16 in each provided by nVidia chips like the BR-03 or NF200. This will allow for 3-way SLI. But the Extreme board will usually also have a fourth x4 slot for,...you guessed it; a PhysX enabled GPU.