NEWS CATEGORIES:



NEWS ARCHIVE >>
SOFTPEDIA REVIEWS >>
Home / News / Technology / Video

Video


NVIDIA Rolls out the Second-gen Tesla

Offers more memory, bandwidth and processing power

By Traian Teglet, Technology News Editor

16th of June 2008, 15:17 GMT

Adjust text size:


The NVIDIA Tesla
Enlarge picture
After ATI proudly released its new over 1 TFlop capable StreamFire 9250 graphics card, NVIDIA has followed AMD's actions and has announced today its second generation Tesla card. As with AMD's offer, NVIDIA's new product is also going to provide a considerable performance boost as compared to its previous version, which was capable of providing only 518 GFlops.

The new T10P graphics processing unit presents a massive die
that comes with 1.4 billion transistors and no less than 240 processing cores, which is higher than the previous 128 cores-enabled 8-series GPUs.

The new entry-level Tesla product remains an add-in card, designed for business users who demand high graphics computing power. The card will integrate 102 GB/s memory bandwidth and a power consumption rating of 160 watts, better than the 170 watts of the previous generation. According to NVIDIA, due to thermal restrictions, the company was forced to clock the C1060 GPU at only 1.33GHz, instead of 1.5GHz, in the blade. Because of this restriction, the C1060 will not be capable of achieving 1 TFlop and is estimated that will only bring 900 Gflops.

Compared to the original C870, which had a MSRP of $1,500, the new C1060 will be offered for $1,699.

Besides its obvious performance improvements, the T10P will also deliver 64-bit double-precision capability, which is required in most fluid dynamics and financial stream processing applications. According to NVIDIA, if the card hadn't been featured with 64-bit double-precision capability, it wouldn't have been capable of achieving the 1 TFlop level, but rather only that of 100 Gflops. Apparently, double-precision calculations dramatically decreased the card's performance.

According to the Santa-Clara based company, the graphics manufacturer intends to release a beta version of CUDA that developers can apply to multi-core CPUs. NVIDIA claims that the CUDA has been downloaded 60,000 times so far, but they don't believe that this exact number of developers are working on HPC applications who, they believe, are rather "playing" with CUDA, trying to create "consumer applications."

TAGS:

Tesla | NVIDIA | CUDA | graphics card


Rating:
Good (3.5/5) 2 vote(s) so far    

Read by 552 user(s) | Add comment | Link to this article
Subscribe to news | Print article | Send to friend

© Copyright 2001-2008 Softpedia
Contact:

 

 

SEARCH THE NEWS ARCHIVE :




Today's News
| Yesterday's News | News Archive


MORE RELATED ARTICLES:


Nvidia and AMD Go Against Intel over USB 3.0 Specification

More Specs for NVIDIA's GTX 280 and GTX 260

Nvidia Boss Thinks the PC and Gaming Consoles Can Get Along

NVIDIA's Next-Generation Cards Scores Revealed

The GTX 280 Gets Full Range of Tests

NVIDIA Finally Rolls out GTX 280 and GTX 260

Get Ready for the Next Intel Ready NVIDIA Chipset

User opinions:

No user comments yet.
Be the first to express your opinion using the form below!

Share your opinion:

Your Name:
Your Email Address:
(will not be used for commercial purposes)
Solve this to prove you're not a bot: =
Your review/opinion:

 






SUBMIT PROGRAM   |   ADVERTISE   |   GET HELP   |   SEND US FEEDBACK   |   RSS FEEDS   |   ENTER NEWS SITE   |   ENGLISH BOARD   |   ROMANIAN FORUM