Advertisement

News Calendar

<< May 2008 >>
S M T W T F S
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31

Advertisement

Site Search

Latest Forum Posts

Tech News From Around The Web

Qimonda Bringing GDDR5 to ATI's Radeon 4000-series

Posted on May 21, 2008 8:59 AM by Rob Williams

AMD has decided to lead the pack where GPU memory is concerned, and Qimonda is the company to deliver. The Radeon 4000 series, which will launch next month, will be the first to feature GDDR5, an evolutionary upgrade to GDDR4, which will feature faster speeds, improved bandwidth and more efficient power features.

Launch GDDR5 parts will be offered in speeds from 900MHz (3.6Gb/s) up to 1GHz (4.0Gb/s) and will run on 1.5v, compared to the 1.8v for GDDR4. Launch cards will not use memory speeds beyond 4.0Gb/s on the 16M x 32 arrangement, however, as the faster and higher-density chips will not be available to Q3 and Q4 (which go up to 5Gb/s).

Both the speed increases and feature-additions are what makes any new DRAM type worthwhile, and GDDR5 does have a few tricks up its sleeves. New is an error detection mechanism, which is similar to ECC server memory. It calculates an eight-bit CRC algorithm on each data burst and instantly allows a calculation repeat on a detected error, improving efficiency.

Then we have clamshell mode, which enables a 32 I/O to be reduced to 16 I/O, allowing each DQ to connect to a separate IC in x16 mode. This feature is a little confusing, as it makes no difference in bandwidth, so the question is, what's the purpose? I'll be checking with Qimonda to see if they have an answer up their sleeves.


Munich, May 21, 2008 - Qimonda AG, a leading manufacturer of memory products, today announced that the company has won AMD as launch partner for the new graphics standard GDDR5. Qimonda already started mass production and the volume shipping of GDDR5 512Mbit components with a speed of 4.0Gbps to AMD, a leading global provider of innovative processing solutions in the computing, graphics and consumer electronics markets.

Source: Qimonda


NVIDIA's GTX 200-series Specs Revealed

Posted on May 21, 2008 8:11 AM by Rob Williams

While NVIDIA's 9-series launch seemed a bit lackluster, given that they were primarily an over-glorified 8-series (though I can't say enough good about the 9600 GT), the companies' upcoming GTX 200 cards are looking to wipe the disgruntled looks off all our faces.

According to DailyTech, two cards will launch initially, the GTX 260 and GTX 280. As you'd expect, the GTX 280 is the highest offering, which will include 240 stream processors (compared to 128 on the 9800 GTX), a huge 512-bit memory bus width, support for up to 1 GB (although it should be possible for companies to add more). On top of those, the unified shaders are to perform 50% faster than previous generation cards.

The lower-spec'd GTX 260 removes 48 stream processors, for 192 total, decreases the memory bus to 448-bit and also lowers the memory to 896MB. Despite the decreases, it's hard to assume that this card will be "gimped" by any standard. Going by specs alone, even the GTX 260 should be absolutely stellar.

The downside, if there is one, is that DX10.1 will not be supported. This is a bit striking, but goes to show NVIDIA doesn't have much faith in it, or feels the need for inclusion. It would have been nice to see it added either way though. Embargo is set to lift on June 18th, which is right around the same time that ATI will lift theirs for the Radeon 4000-series.


The GTX 280 enables all features of the D10U processor; the GTX 260 version will consist of a significantly cut-down version of the same GPU. The D10U-30 will enable all 240 unified stream processors designed into the processor. NVIDIA documentation claims these second-generation unified shaders perform 50 percent better than the shaders found on the D9 cards released earlier this year.

Source: DailyTech


Intel to Launch Q9650 in Q3 to Replace QX9650

Posted on May 21, 2008 7:47 AM by Rob Williams

According to Chinese hardware site HKEPC, Intel will be discontinuing the current high-end QX9650 Extreme processor in Q3, with an identically-specced mid-range offering, appropriately named, Q9650. The new CPU will be identical all-around to the QX9650, including the 3.0GHz clock speed, 12MB cache and 1333MHz FSB, but will retail for $530USD... half of the current Extreme processor.

Also noted is that the E8600 will show face around the same time, but we knew that already. It's a Dual-Core offering at 3.33GHz and will retail for $224. In addition to these launches, the Q9550 is to decrease to $316, while the Q9400 will see a new $266 price tag. Probably not-so-surprisingly, the Q9300 and Q6700 will fade out at this time.

The interesting thing is that while the QX9650 will essentially become the Q9650, there will be no "high-end" offering available, except for the ultra-high-end QX9770, which currently retails for around $1,500. I don't think many in their right mind would pay a $1,000 premium at that time for a 200MHz boost, so we have to wonder if Intel has something else up their sleeve. But given that Nehalem will be right around the corner at that time, the next Extreme edition we see may very well be one based on the new architecture.


Core 2 Quad when Q9650 core the arteries are 3GHz, supports 1333MHz FSB, to construct 12MB L2 Cache (6MB x 2), highest TDP is 95W. Will lie in the frequency multiplication with the Core 2 Extreme QX9650 distinction to lock, has limited its ultra frequency ability. The selling price aspect, Core 2 Quad Q9650 every thousand unit prices are $530 US dollars, compares Core 2 Quad QX9650 every thousand unit prices to reach as high as $999 US dollars amiable.

Source: HKEPC (Translated to English)


Gateway's XHD3000 Now Works Under Linux (Somehow)

Posted on May 21, 2008 7:16 AM by Rob Williams

When I reviewed Gateway's XHD3000 a few months ago, I complained that Linux would not work with it at the native resolution. To me, that was a big problem, since I occasionally benchmark under Linux on our test rigs. Other than that though, the monitor was fantastic.

Well, sometime between then and now, the problem fixed. For fun, I hooked up the XHD3000 to my main machine, just to play around with dual-display possibilities, but in doing so, I expected the NVIDIA driver to default to 1280x800... the same resolution I've always been stuck to. I was quite surprised when I saw the native resolution of 2560x1600 become available, though, and applying it delivered the promised result.

I also tested the monitor by itself afterwards to make sure that it wasn't only enabled because I had it as a secondary display, and it again worked just fine. So we might very-well be at a point where all 30-inch displays should work under Linux with little issue, and from what I can tell, we might have NVIDIA to thank. For those interested in the monitor in question, you can click to read the review below. I liked it enough to buy one, which may very well speak for itself.

Gateway, not content to let Dell and others hog the limelight in the display market, have recently launched their XHD3000 monitor. It offers a wide-range of connectivity options along with bright picture and an incredible upscaler. When said and done, this is one model we can heartily recommend.

Source: Gateway XHD3000 Review


May 21st Tech Roundup

Posted on May 21, 2008 1:00 AM by Rob Williams







    Competitions, Complete Systems & Et cetera
  • 10+ Things I Learned With my First DSLR - TheTechLounge
  • Toshiba Satellite A305-S6845 - Computer Shopper


News Archives