In California, more specifically in San Jose, introduced Nvidia CEO and cofounder Jen-Hsun Huang today on the GPU Technology Conference a major breakthrough before - CUDA x86. Anyone who has ever been explained in more detail with modern graphics cards and PC gaming, which is expected CUDA PhysX or better to be a concept is, it is here to have running on the graphics card (physics) calculations.
But as so often is strong competition for parts in the economy before the graphics accelerator is not just modern, why CUDA had been exclusively run on Nvidia graphics cards from the GeForce 8 generation. While there was also an owner of an AMD Radeon card, the user of an Intel GMA rely on the physics effects, but these were then calculated by the CPU - at a pace, which brought many players to give to rather entirely on the listed brightened the effects that this could also the fastest computing monster Core i7 or Phenom II X6 to its knees. This was the processing of the unit due to the x86 processor, which represents an old floating-point unit, which is but ever since the introduction of Intel's SSE unit little more attention. But this brake had a meaning, so you had to find an incentive to make the customer the GeForce graphics cards tasty, if this could not on price or efficiency, not to mention the volume achieved.
This wall has now fallen, Nvidia is now working with Fortran on an x86 compatible compiler, which can speed up the CUDA commands well on normal processors, without excessive resource consumption. While PhysX for Nvidia is almost irrelevant, since the calculations are likely to CUDA Tesla cards or ordinary servers soon have a larger market share than Nvidia's graphics products. This promises to be slow now, what Nvidia has for the identities ultimately end the year: Man wanted away from hardware and toward software that offers the customer the advantages and disadvantages. Sun. Nvidia PhysX Intel Havok Engine is indeed superior to worlds, for the loss of a DirectX graphics chip producers would have a more than bad on price and innovation of video cards - even if projected that the share of external graphics cards would soon be vanishingly small. Germany is still a stronghold of separate graphics cards, while almost the entire U.S. space consists of integrated GPUs.
But as so often is strong competition for parts in the economy before the graphics accelerator is not just modern, why CUDA had been exclusively run on Nvidia graphics cards from the GeForce 8 generation. While there was also an owner of an AMD Radeon card, the user of an Intel GMA rely on the physics effects, but these were then calculated by the CPU - at a pace, which brought many players to give to rather entirely on the listed brightened the effects that this could also the fastest computing monster Core i7 or Phenom II X6 to its knees. This was the processing of the unit due to the x86 processor, which represents an old floating-point unit, which is but ever since the introduction of Intel's SSE unit little more attention. But this brake had a meaning, so you had to find an incentive to make the customer the GeForce graphics cards tasty, if this could not on price or efficiency, not to mention the volume achieved.
This wall has now fallen, Nvidia is now working with Fortran on an x86 compatible compiler, which can speed up the CUDA commands well on normal processors, without excessive resource consumption. While PhysX for Nvidia is almost irrelevant, since the calculations are likely to CUDA Tesla cards or ordinary servers soon have a larger market share than Nvidia's graphics products. This promises to be slow now, what Nvidia has for the identities ultimately end the year: Man wanted away from hardware and toward software that offers the customer the advantages and disadvantages. Sun. Nvidia PhysX Intel Havok Engine is indeed superior to worlds, for the loss of a DirectX graphics chip producers would have a more than bad on price and innovation of video cards - even if projected that the share of external graphics cards would soon be vanishingly small. Germany is still a stronghold of separate graphics cards, while almost the entire U.S. space consists of integrated GPUs.