Nvidia to Hit the x86 CPUs With CUDA Capability
Your CPU will be able to do CUDA.
We've heard rumors that Nvidia been dipping its toe into the x86 CPU market, and today the graphics company made an announcement related to that – but it's not what you think.
Nvidia CEO Jen-Hsun Huang revealed that the company will bring its CUDA programming language to "any computer, or any server in the world," with the help of the Portland Group (PGI).
Specifically, this means that systems without Nvidia GPUs will be able to process CUDA code, giving the company its answer to Microsoft's DirectCompute and the more open OpenCL.
Nvidia says that its CUDA without a GPU will run best on multicore CPUs and will be ideal for servers.
(Source: Electronista.)
8
Comments
Read more
- cuda ,
- gpgpu ,
- x86 ,
- opencl ,
- directcompute
Nvidia's Future GPUs Will Be Kepler, Then Maxwell
- Kinect Director: FPS Not Played on PCs Anymore
- PlayStation Team Wants Android Engineers
- HP Launches Budget and Touch All-in-One PCs
- PS3 Now Has Ability to Play 3D Blu-ray Movies
- HP: Our WebOS Devices Will be Similar to the iPad
- Seagate's Industry First 1.5 TB Portable Drive
- Michael Jackson-Themed MMOG Announced
- Intel Threatens to Sue HDCP Crack Users
- Windows HPC Server 2008 R2 is Done and Ready
Nvidia Tegra 3 Almost Done, Tegra 4 On the Way
- Rumor: Nintendo 3DS to Pack Two ARM CPUs
- No Super Street Fighter IV PC Thanks to Piracy
- VIDEO: Plants vs. Zombies in WoW: Cataclysm
- Updated: Tom's Hardware 2010 Charts
- More Than Two Million Have Downloaded IE9 Beta
- LaCie Shows Super Accurate IPS LCD Monitor
- Good Old Games Re-Launches, Today
- Intel Launches Scavenger Hunt for iPhone Users
- AMD Tells Girls How to Attract a Geeky Boyfriend
Sponsored
See more
Latest news
Graphics Cards Previous news
Partners
Nice news, but maybe you want to explain more. There are already many solutions to run code on multicore configurations, such openmp for shared memory and mpi for separate memory per core. What?How? More detaisl please.
I hope nvidia is not jut trying to reinvent the wheel.
What, the same nVidia who crippled CPU-based PhysX just so they can make their products look good?
What, the same nVidia who crippled CPU-based PhysX just so they can make their products look good?
Okay maybe that's unfair, however it cannot be hard to update the old code from x87 to SSE... can it?
I thought the whole point was GPUs were better at the sort of computations the whole GPGPU revolution was founded on. Isn't it therefore a but redundant to run CUDA on CPU? Where are the gains?
Besides, there are many solutions which work for the CPU. I have been doing some basic parallel programs and I think that it will be difficult to make auto-parallelization more efficient than the one done even by an average scientist.
To be honest I'm a bit confused by this news..running GPU-optimised code on a CPU? There's got to be some performance penalty, even with multiple cores at hand.
Also, I didn't see anything mentioned about running CUDA on non-nVidia GPU's. Perhaps it's down to patents and other legal issues. Still, wouldn't that option be more efficient, especially from an energy consumption perspective? I reckon even an onboard GPU such as the AMD/ATI 785g running translated CUDA would give a lot of quad-core processors a run for their money - the GPU is much more optimised for this type of task. Maybe we'll see this theory put to the test in an upcoming review