NVIDIA cuda2.0 CPU physx

Http://www.hkepc.com /? Id = 1214 & FS = c3nl

NVIDIA will launch the Cuda June 17 platform on the same day with the Alibaba geforce GTX 200 family in March 2.0, adding support for precision (Double Precision) computing, in order to provide a more accurate computing result, take the weather image as an example, dr. david Kirk noted that a study using geforce 8800 GTX for WRF-mode image computing compared with Pentium D 2.8 GHz found that geforce 8800 GTX has a 17-fold performance ratio, cost per flop is about 1.2X. It is worth noting that the study also pointed out that the growth ratio has not been any optimized for the GPU, and I believe that the performance growth will be improved in the future. Study article: http://www.mmm.ucar.edu/wrf/WG2/michalakes_lspp.pdf

When we see this information, we should have more friends who can use Cuda for development ~ With Dual-precision, it provides better support for biological and fluid simulation.

Http://www.hkepc.com /? Id = 1245 & FS = Rell

NVIDIA announced the launch of the tegar series of cross-platform CPU architecture with arm 11:
NVIDIA announced at the Taipei press conference on the 2 th that it has launched a single-chip electrical product that is comparable to the high quality and online user experience of desktop computers, and bring the PC perception computing efficiency into various types of manual operations, and formally cross-process the operator, the middleware uses the arm 11 architecture and adds it to the NVIDIA geforce Processing Kernel core, and is named as the tegra series processor. It is deployed on the smartphone, mid, and netbook marketplaces, its Single-chip design incorporates the processor, Zookeeper core, and north-south features. It is expected that the first NVIDIA tegra processor product will be officially available by the end of this year.

So cool ~ In this way, it is unnecessary to exchange host and device data, which is at least much less than the current time. You don't have to pay for intel if you don't use the X86 architecture. Haha ~ This is suspected ~ Maybe we saw that Intel had failed in the past. NVIDIA first chose embedded to try ...... Hope ......

Http://www.hkepc.com /? Id = 1210 & FS = Rell

NVIDIA integrated physx Project
AMD wants physx skills !? NVIDIA promotional materials

Asked whether physx can be applied to ATI products, Manju Hegde pointed out that this is a very easy task, for example, the original ageia release completes the development of the new physx notebook in just one month, as long as AMD is willing to open a cuda-like API, they intend to add physx computing functions to ATI products to benefit their users.

If AMD and NVIDIA are combined, Intel should be uncomfortable ...... Competitive