NVIDIA Tegra
Published by Codrut Nistor, on April 15th, 2008, in the categories: News, Video Cards
When I found out about NVIDIA's upcoming 1 billion tranzistors GPU, I considered it only a hint related to what's going to come, especially since there were no serious details available about it. Anyway, that was happening only a few days ago, and now we have more than just a cryptic codename, but technical details are still secret. So...what do we have, after all? It's a more decent codename, "Tegra," which may also end up as being the name of the future 1 billion tranzistors-powered videocard.
The story goes like this - NVIDIA applied for the "Tegra" trademark with the US patent and trademark office, and everyone is expecting a major product to arrive. The problem is that NVIDIA's description of the upcoming device is only slightly better than nothing - what do you understand by "G & S: Integrated circuits?" If you know more than me, feel free to share it with all of us here, because I have no clue!
For now, NVIDIA didn't comment on the story, so there plenty of room left for rumors and speculations. Are we going to get the first GPU with integrated physics processing unit? Is it possible for the 9800 GX2 to be only the beginning of what's coming up next? I know there are plenty of possibilities, so don't be shy and drop your version below - what should we expect from NVIDIA's Tegra?
The story goes like this - NVIDIA applied for the "Tegra" trademark with the US patent and trademark office, and everyone is expecting a major product to arrive. The problem is that NVIDIA's description of the upcoming device is only slightly better than nothing - what do you understand by "G & S: Integrated circuits?" If you know more than me, feel free to share it with all of us here, because I have no clue!
For now, NVIDIA didn't comment on the story, so there plenty of room left for rumors and speculations. Are we going to get the first GPU with integrated physics processing unit? Is it possible for the 9800 GX2 to be only the beginning of what's coming up next? I know there are plenty of possibilities, so don't be shy and drop your version below - what should we expect from NVIDIA's Tegra?
Intel’s Spring Price Cuts
Published by Bogdan Alex, on April 15th, 2008, in the categories: CPU
AMD holds on just fine thanks to its strategic product pricing. Their desktop quad-cores revolve around the $250 spot and things are about to get cheaper as Deneb approaches the launch. We can fairly say that AMD managed to somehow recover after a disappointing financial year in 2007. Intel, however, prepares a new price cut for its CPUs, aiming to keep the archenemy in check for the next several months.
Intel’s new price cuts will be effective as of Sunday April 20th. They will significantly affect the prices of the older 65nm CPUs, while the 45nm parts will see a minimal price drop. I mean minimal like $2 cheaper.
In this respect, the high-end 45nm model that will get cheaper is the Core 2 Quad Q9550 at 2.83GHz with 12MB L2 cache. It drops from $527 to $525. Keep in mind that these are only the wholesale prices. The cheaper Core 2 Quad Q9450 clocked at 2.66GHz with 12MB cache will drop from $319 to $317, whereas the Core 2 Quad Q9300 clocked at 2.5GHz with 6MB cache will drop from $270 to $268.
Older models such as the Q6700 with a 2.66GHz clock speed and 8MB cache and a 1066MHz FSB speed will drop from the current $527 to $268. Now that’s what I call drastic! Sticking to the ultra popular CPUs, we see the 65nm Kentsfield Q6600 with its 2.4GHz clock and 8MB cache dropping from $250 to $227.
All dual-core 45nm E8x0 series Penryn based models will also see the $2 drop. The older 65nm dual-core E6850 at 3GHz and 4MB cache drops to $266, the E6750 at 2.66GHz and 4MB cache drops to $187 and the E6550 at 2.33GHz and 4MB cache drops to $167.
But that’s not all. Intel really wants to put some more pressure on AMD by dropping prices for the 4-series Core 2 Duo processors. Starting with the E4700 with 2.60GHz clock, 800MHz FSB speed and 2MB cache drops $2 to $137, the E4600 with 2.40GHz clock, 800MHz FSB speed and 2MB cache drops from $139 to $117 while the E4500 with 2.20GHz clock 800MHz FSB speed and 2MB cache drops $2 to $117.
Things are looking good for Intel and we should expect some more price cuts when Nehalem enters the scene, later this year.
Intel’s new price cuts will be effective as of Sunday April 20th. They will significantly affect the prices of the older 65nm CPUs, while the 45nm parts will see a minimal price drop. I mean minimal like $2 cheaper.
In this respect, the high-end 45nm model that will get cheaper is the Core 2 Quad Q9550 at 2.83GHz with 12MB L2 cache. It drops from $527 to $525. Keep in mind that these are only the wholesale prices. The cheaper Core 2 Quad Q9450 clocked at 2.66GHz with 12MB cache will drop from $319 to $317, whereas the Core 2 Quad Q9300 clocked at 2.5GHz with 6MB cache will drop from $270 to $268.
Older models such as the Q6700 with a 2.66GHz clock speed and 8MB cache and a 1066MHz FSB speed will drop from the current $527 to $268. Now that’s what I call drastic! Sticking to the ultra popular CPUs, we see the 65nm Kentsfield Q6600 with its 2.4GHz clock and 8MB cache dropping from $250 to $227.
All dual-core 45nm E8x0 series Penryn based models will also see the $2 drop. The older 65nm dual-core E6850 at 3GHz and 4MB cache drops to $266, the E6750 at 2.66GHz and 4MB cache drops to $187 and the E6550 at 2.33GHz and 4MB cache drops to $167.
But that’s not all. Intel really wants to put some more pressure on AMD by dropping prices for the 4-series Core 2 Duo processors. Starting with the E4700 with 2.60GHz clock, 800MHz FSB speed and 2MB cache drops $2 to $137, the E4600 with 2.40GHz clock, 800MHz FSB speed and 2MB cache drops from $139 to $117 while the E4500 with 2.20GHz clock 800MHz FSB speed and 2MB cache drops $2 to $117.
Things are looking good for Intel and we should expect some more price cuts when Nehalem enters the scene, later this year.
GeForce 9600 GSO – In May
Published by Codrut Nistor, on April 14th, 2008, in the categories: Video Cards
Currently, NVIDIA's 9600 lineup only includes the 9600 GT card, but things are going to change in May. While the 9600 GT cards offer a considerable performance increase over the 8600 GT(according to NVIDIA, it's 90 percent), the 8600 GTS doesn't really have a 9600 GTS equivalent. Anyway, the upcoming 9600 GSO card is going to fill a void, but not quite that one above the 9600 GT...
Looking back, I first have to remind you some obvious things - being more powerful when compared to the 8600 GT, the 8600 GTS cards were also more expensive, but when talking about the 9600 GT and the 9600 GSO, it's the other way around. How's that?
Well, it seems the 9600 GSO will be a reloaded 8800 GS - both cards use the G92 core with 96 stream processors, but with the 9600 GSO, hardware producers are free to design their own PCB, and set their own clocks, so seeing such cards pushed to the limit, trying to grab some market share back from ATI, would be a real delight, and since May is not that far...I hardly wait to see!
Looking back, I first have to remind you some obvious things - being more powerful when compared to the 8600 GT, the 8600 GTS cards were also more expensive, but when talking about the 9600 GT and the 9600 GSO, it's the other way around. How's that?
Well, it seems the 9600 GSO will be a reloaded 8800 GS - both cards use the G92 core with 96 stream processors, but with the 9600 GSO, hardware producers are free to design their own PCB, and set their own clocks, so seeing such cards pushed to the limit, trying to grab some market share back from ATI, would be a real delight, and since May is not that far...I hardly wait to see!
NVIDIA Wages War against Intel
Published by Bogdan Alex, on April 14th, 2008, in the categories: Video Cards
What do we know about NVIDIA and Intel? Well, NVIDIA possesses an important share of the GPU market, but it’s still behind Intel. We know this is because the integration of onboard GPUs in laptops and other mobile stuff. The PC is definitely NVIDIA’s ally. Or not? Since NVIDIA is making motherboard chipsets, looks like things are a bit more complicated than they appear.
Jensen Huang, NVIDIA CEO, is betting it all on the games industry as related to PCs, while Intel is supposed to adopt an opposing strategy – they’ll support anything other than game-related stuff. Jon Peddie’s research company has come to the conclusion that the graphic industry in Q4 2007 sold 31.65 million discrete video card solutions, and the entire discrete GPU market generated $6.57 billion and keeps growing. An important part of this money goes to NVIDIA or AMD/ATI, but most of it is really making its way in Intel’s pocket. However, this is not about the money for Intel. Instead, Intel is looking to generate a solid marketing image, but NVIDIA is no stranger to this kind of strategy.
NVIDIA is currently improving on the signature image with its GeForce series while Intel did the same with the Centrino mobile family, and when it comes to marketing ploys and strategies, all’s going towards a colossal clash, as both companies are doing this part of the job in their own exquisite ways.
But keep in mind that Intel is at least 10 times bigger than NVIDIA when we think about the amounts of money earned per year, not to mention that Intel is preparing NVIDIA’s and AMD/ATI’s slayer – the Larabee GPU that’s supposed to come in 2009. Looks like NVIDIA is going the Spartan way in this war.
Jensen Huang, NVIDIA CEO, is betting it all on the games industry as related to PCs, while Intel is supposed to adopt an opposing strategy – they’ll support anything other than game-related stuff. Jon Peddie’s research company has come to the conclusion that the graphic industry in Q4 2007 sold 31.65 million discrete video card solutions, and the entire discrete GPU market generated $6.57 billion and keeps growing. An important part of this money goes to NVIDIA or AMD/ATI, but most of it is really making its way in Intel’s pocket. However, this is not about the money for Intel. Instead, Intel is looking to generate a solid marketing image, but NVIDIA is no stranger to this kind of strategy.
NVIDIA is currently improving on the signature image with its GeForce series while Intel did the same with the Centrino mobile family, and when it comes to marketing ploys and strategies, all’s going towards a colossal clash, as both companies are doing this part of the job in their own exquisite ways.
But keep in mind that Intel is at least 10 times bigger than NVIDIA when we think about the amounts of money earned per year, not to mention that Intel is preparing NVIDIA’s and AMD/ATI’s slayer – the Larabee GPU that’s supposed to come in 2009. Looks like NVIDIA is going the Spartan way in this war.
NVIDIA APX 2500 HD
Published by Bogdan Alex, on April 13th, 2008, in the categories: Video Cards
The green guys at NVIDIA might take over the world some day. OK, not really, but for now, they are planning to make a big hit out of their mobile-oriented GPU – the APX 2500 HD.
Mabe you’ve heard about the APX 2500 HD video and graphics processing platform back in February this year, when NVIDIA cared to reveal some details and showcased a demo of a iPhone-like handheld device. Here’s the demo, in case you haven’t seen it.
Recently, NVIDIA disclosed some more info on the upcoming APX 2500 HD architecture. The video chip has a power consumption of 50 mW which allows it to run up to 10 hours with a single battery charge. As such, we can fairly say that APX 2500 is clearly aiming at mobile phones, PDAs, and GPS devices. NVIDIA even acknowledged the fact that the first devices to integrate the chip will be a series of GPS devices that could be launched later this year. Phones and PDAs are supposed to wait until 2009.
NVIDIA also informs that the APX 2500 is a single chip solution which has a package slightly smaller than a dime. Believe it or not, this small-form factor is able to accommodate a ARM11 MP core with a maximum clock speed of 750MHz, a GeForce class GPU, a high-definition audio and video processor, memory, and IO system and an image signal processor all in the same package. The chip is HD-ready and can play high-definition video flawlessly through HDMI 1.2 with additional dual display capabilities to view video on both the device and a high-definition display.
Unfortunately, you won’t see 1080p support with this generation. The APX 2500 HD is capable of 720p resolutions in H.264, MPEG-4, and VC-1 formats, while the GeForce GPU (most likely a miniaturized GF 6000 family GPU) is capable of producing 3D graphics with features such as programmable pixel shaders and vertex & lighting, as well as coverage sampled anti-aliasing.
I reckon that in a couple of years we could play DX10/11 games directly on a handheld device. NVIDIA should already be considering a portable game console, if not a portable player or some iPhone destroyer.
Mabe you’ve heard about the APX 2500 HD video and graphics processing platform back in February this year, when NVIDIA cared to reveal some details and showcased a demo of a iPhone-like handheld device. Here’s the demo, in case you haven’t seen it.
Recently, NVIDIA disclosed some more info on the upcoming APX 2500 HD architecture. The video chip has a power consumption of 50 mW which allows it to run up to 10 hours with a single battery charge. As such, we can fairly say that APX 2500 is clearly aiming at mobile phones, PDAs, and GPS devices. NVIDIA even acknowledged the fact that the first devices to integrate the chip will be a series of GPS devices that could be launched later this year. Phones and PDAs are supposed to wait until 2009.
NVIDIA also informs that the APX 2500 is a single chip solution which has a package slightly smaller than a dime. Believe it or not, this small-form factor is able to accommodate a ARM11 MP core with a maximum clock speed of 750MHz, a GeForce class GPU, a high-definition audio and video processor, memory, and IO system and an image signal processor all in the same package. The chip is HD-ready and can play high-definition video flawlessly through HDMI 1.2 with additional dual display capabilities to view video on both the device and a high-definition display.
Unfortunately, you won’t see 1080p support with this generation. The APX 2500 HD is capable of 720p resolutions in H.264, MPEG-4, and VC-1 formats, while the GeForce GPU (most likely a miniaturized GF 6000 family GPU) is capable of producing 3D graphics with features such as programmable pixel shaders and vertex & lighting, as well as coverage sampled anti-aliasing.
I reckon that in a couple of years we could play DX10/11 games directly on a handheld device. NVIDIA should already be considering a portable game console, if not a portable player or some iPhone destroyer.
PAGES