admin2
12-08-2009, 05:28 PM
Intel's decision to indefinitely delay its Larrabee graphics processor is a setback for the Linux community, delaying improved performance for users of open source operating systems. Competitors Nvidia and AMD, however, saw their share prices rise following the news.
To the Windows and Mac world, the end of Larrabee may not matter. They have plenty of advanced graphics options. That's not true for Linux users, who rely upon Intel's integrated graphics for the vast majority of their needs.
Larrabee, which was Intel's program to compete with standalone graphics processor unit (GPU) chips from Nvidia and AMD, was expected to include Intel's usual support for open source operating systems. That would have made Larrabee the choice for desktop Linux users, especially those working against virtualized servers.
Alas, it was not meant to be. Yesterday, Intel said the project was being discontinued, having earlier indicated that Larrabee was behind schedule. My colleague, Tony Bradley, has suggested the chip was really the victim of legal issues that Intel was unable to resolve.
In late September, Intel demonstrated the processor to a mixed reception at its developer's forum.
If not as a processor, Larrabee lives on as a software development platform for use inside Intel and externally, an Intel spokesman said. Back in September, an Intel executive also indicated that Larrabee technology would find its way into an Intel CPU, which still appears likely.
On the stock markets early Monday, Reuters reported that shares of AMD rose 7.9 percent to $8.50 on the New York Stock Exchange while Nvidia rocketed nearly 14 percent to $16.25 on the NASDAQ, and Intel fell 0.5 percent to $20.35.
"This is not the end of Larrabee as a graphics processor -- this is a pause," said Jon Peddie, president of Jon Peddie Research, a market research firm specializing in graphics cards, in a blog entry. The next version of Larrabee could be offered as a co-processor that could work along with CPUs for faster processing of high-performance computing applications, Peddie said.
Analysts also said some attributes from existing and future Larrabee architectures could make their way inside future integrated graphics chips that go into CPUs. The benefits of such integration would be similar to performance improvements provided by graphics processors as compute engines to accelerate gaming and Flash applications, said Dean McCarron, principal analyst at Mercury Research. Processors with elements of Larrabee could appear in the 2011-to-2012 timeframe, he said.
Intel in the past has said that some technology behind Larrabee may be used to produce graphics cores that can be integrated into CPUs. The company is set to launch its first processor with integrated graphics later this month, an Intel spokesman said.
However, graphics cards are very difficult to design, said Nathan Brookwood, principal analyst at Insight 64.
"You want to make sure you have it right before you integrate it into a CPU," Brookwood said. The next generation of Larrabee will first need to come out as a stand-alone graphics card and deliver proven performance after which it can be integrated into CPUs. A flawed product could screw up a whole generation of CPUs, Brookwood said.
Peddie said that integrating the current Larrabee architecture into processors could be impossible, as it is comparable to putting more cores inside the CPU. However, he said it is conceivable that Intel might integrate some of Larrabee's specialized processors uniquely needed for graphics, like texture mapping processors, inside future CPUs.
But it was a smart decision by Intel to cancel the initial version of Larrabee, analysts said. The sales expectations around Larrabee as a discrete graphics card weren't high, and perhaps the processor didn't meet Intel's metrics of success, so it made financial sense to delay the product. "I don't see it as that big of a deal," McCarron said.
Competitive pressure also may have forced Intel back to the drawing board, Brookwood said. Larrabee was designed to compete with graphics cards from Advanced Micro Devices and Nvidia, which already offer products that deliver faster performance. Larrabee is already late, and it would not have been a competitive product by the time it hit the market, he said.
"Intel certainly has the resources to continue to invest in this area. I won't be surprised if they go back to the drawing board and design a next-generation part," Brookwood said.
Intel invested a lot of time and money in the development of Larrabee, but none of that is lost, Peddie said. Larrabee is an interesting architecture and paves the way for "serious potential and opportunity" as a significant co-processor in high-performance computing, Peddie said.
"Rather, that work provides the building blocks for the next phase," Peddie said.
Original article taken from Yahoo News.
To the Windows and Mac world, the end of Larrabee may not matter. They have plenty of advanced graphics options. That's not true for Linux users, who rely upon Intel's integrated graphics for the vast majority of their needs.
Larrabee, which was Intel's program to compete with standalone graphics processor unit (GPU) chips from Nvidia and AMD, was expected to include Intel's usual support for open source operating systems. That would have made Larrabee the choice for desktop Linux users, especially those working against virtualized servers.
Alas, it was not meant to be. Yesterday, Intel said the project was being discontinued, having earlier indicated that Larrabee was behind schedule. My colleague, Tony Bradley, has suggested the chip was really the victim of legal issues that Intel was unable to resolve.
In late September, Intel demonstrated the processor to a mixed reception at its developer's forum.
If not as a processor, Larrabee lives on as a software development platform for use inside Intel and externally, an Intel spokesman said. Back in September, an Intel executive also indicated that Larrabee technology would find its way into an Intel CPU, which still appears likely.
On the stock markets early Monday, Reuters reported that shares of AMD rose 7.9 percent to $8.50 on the New York Stock Exchange while Nvidia rocketed nearly 14 percent to $16.25 on the NASDAQ, and Intel fell 0.5 percent to $20.35.
"This is not the end of Larrabee as a graphics processor -- this is a pause," said Jon Peddie, president of Jon Peddie Research, a market research firm specializing in graphics cards, in a blog entry. The next version of Larrabee could be offered as a co-processor that could work along with CPUs for faster processing of high-performance computing applications, Peddie said.
Analysts also said some attributes from existing and future Larrabee architectures could make their way inside future integrated graphics chips that go into CPUs. The benefits of such integration would be similar to performance improvements provided by graphics processors as compute engines to accelerate gaming and Flash applications, said Dean McCarron, principal analyst at Mercury Research. Processors with elements of Larrabee could appear in the 2011-to-2012 timeframe, he said.
Intel in the past has said that some technology behind Larrabee may be used to produce graphics cores that can be integrated into CPUs. The company is set to launch its first processor with integrated graphics later this month, an Intel spokesman said.
However, graphics cards are very difficult to design, said Nathan Brookwood, principal analyst at Insight 64.
"You want to make sure you have it right before you integrate it into a CPU," Brookwood said. The next generation of Larrabee will first need to come out as a stand-alone graphics card and deliver proven performance after which it can be integrated into CPUs. A flawed product could screw up a whole generation of CPUs, Brookwood said.
Peddie said that integrating the current Larrabee architecture into processors could be impossible, as it is comparable to putting more cores inside the CPU. However, he said it is conceivable that Intel might integrate some of Larrabee's specialized processors uniquely needed for graphics, like texture mapping processors, inside future CPUs.
But it was a smart decision by Intel to cancel the initial version of Larrabee, analysts said. The sales expectations around Larrabee as a discrete graphics card weren't high, and perhaps the processor didn't meet Intel's metrics of success, so it made financial sense to delay the product. "I don't see it as that big of a deal," McCarron said.
Competitive pressure also may have forced Intel back to the drawing board, Brookwood said. Larrabee was designed to compete with graphics cards from Advanced Micro Devices and Nvidia, which already offer products that deliver faster performance. Larrabee is already late, and it would not have been a competitive product by the time it hit the market, he said.
"Intel certainly has the resources to continue to invest in this area. I won't be surprised if they go back to the drawing board and design a next-generation part," Brookwood said.
Intel invested a lot of time and money in the development of Larrabee, but none of that is lost, Peddie said. Larrabee is an interesting architecture and paves the way for "serious potential and opportunity" as a significant co-processor in high-performance computing, Peddie said.
"Rather, that work provides the building blocks for the next phase," Peddie said.
Original article taken from Yahoo News.