Asus ROG Matrix GeForce GTX 580 Graphics Card Details Leaked

Subject: Graphics Cards | May 25, 2011 - 08:04 PM |
Tagged: ROG, nvidia, GTX 580, asus

X Fastest, a Chinese language technology website today posted images of the Asus ROG Matrix version of the NVIDIA GTX 580 graphics card. The three-slot (no, that is not a typo) graphics card is claimed to have a 16 phase VRM design, GPU clock of 816 MHz, shader clock of 1632 MHz, and a memory clock of 4008 MHz. Further, the card contains 1.5GB of GDDR5 memory on a 384 bit bus. All this power is delivered via two 4x2 PCI-E connectors (8 pin). The following table compares the claimed Asus card's speeds to NVIDIA's reference design.

  Asus ROG Matrix GTX 580 NVIDIA Reference Design
GPU Clock  816 MHz  772 MHz
Shader Clock  1632 MHz  1544 MHz
Memory Clock  4008 MHz (effective)  4008 MHz (effective)
Memory Amount and Bus  1.5 GB GDDR5, 384 bit bus  1.5 GB GDDR5, 384 bit bus
PCI-E Connections  Two 4x2 PCI-E (8 pin) connector  One 6 pin, One 8 pin connector

 

The card features several overclocker friendly features, including nodes to directly measure voltage, hardware buttons to increase/decrease voltage to the card, and a “safe mode” button that promises to restore the card to factory settings located next to the uppermost DVI output.

Asus580MatrixROG.jpg

Asus seems to have gone big with the Matrix GTX 580, and and the three slot cooler should provide ample air cooling for overclocking. On the other hand, one has to question why Asus did not choose to implement a water block for this overclocker-oriented card, as water cooling could allow overclockers to run the card at the same or higher speeds (especially when voltage increases are necessary) and with lower temps. Further, the three slot design may prohibit enthusiasts from using them in SLI configurations, depending on their motherboard layouts. You can see more images of the card here.
Source: X Fastest

Revisiting quad-gpus and the Law of Diminishing Returns

Subject: Graphics Cards | May 24, 2011 - 12:44 PM |
Tagged: quad sli, quad crossfire, sli, crossfire, nvidia, amd

With SLI and CrossFire we all hoped to see direct scaling so that a quad GPU setup would be somewhere in the neighbourhood of 4x better than a single GPU.  That has proven to be incorrect, not only is the scaling nowhere near that it has been discovered that in some cases going beyond 2 GPUs can actually reduce performance.  

As the hardware and drivers evolve, it is worth revisiting the scaling performance of both AMD and NVIDIA which is why [H]ard|OCP grabbed two GeForce GTX 590s and two AMD Radeon HD 6990s, both dual GPU cards.  In three of the five games tested they ran into at least one issue, a strike right off the bat.  Read on to see how they rate the value of the two manufacturers based on the performance they saw once they'd resolved the problems.

H_quad.gif

"How does NVIDIA's GeForce GTX 590 SLI Quad-GPU compare to AMD's Radeon HD 6990 CrossFireX Quad-GPU? We will find out if these "if-money-didn't-matter dream video card setups" will deliver the gameplay experience we all expect."

Here are some more Graphics Card articles from around the web:

Click Here to go to Video Cards   Graphics Cards

Source: [H]ard|OCP

KFA2 Launches Two NVIDIA GeForce GTX 560 Graphics Cards

Subject: Graphics Cards | May 21, 2011 - 12:04 AM |
Tagged: nvidia, kfa2, GTX 560, graphics

large560EXOC_image_3.jpg

Not to be left out of the slew of NVIDIA GeForce GTX 560 releases, KFA2 announced two new NVIDIA graphics cards to their current graphics card lineup.  Both are based on the Geforce GTX 560 GPU; however, one card is overclocked and fitted with an aftermarket heatsink and fan combo (the other is a standard single, centered, and shrouded fan design).  Labeled the KFA2 GeForce GTX 560 1GB 256bit and the KFA2 GeForce GTX 560 EX OC 1GB 256bit, the DirectX 11 cards offer the following specifications:

  GeForce GTX 560 1GB 256bit GeForce GTX 560 EX OC 1GB 256bit
CUDA Cores 336 336
GPU Clock 810 MHz 905 MHz
Shader Clock 1620 MHz 1810 MHz
Memory Clock 2004 MHz 2004 MHz
Memory 1 GB GDDR5 on 256-bit bus 1 GB GDDR5 on 256-bit bus
Memory Bandwidth 128.3 GB/s 128.3 GB/s
Texture Fill Rate 45.3 Billion/s 50.6 Billion/s

 

The two new cards seem to be positioned (specifications wise) between purely reference cards and the highest clocked GTX 560 cards of their competitors.  The street price will ultimately determine if they are worth picking up versus other brands with higher clocks or reference clocks but aftermarket cooling.  KFA2 states that the cards will be available online and in retail stores throughout Europe, and are backed by a two year warranty.

Source: KFA2

NVIDIA drops $163 million in generated revenue

Subject: General Tech | May 19, 2011 - 08:35 AM |
Tagged: nvidia, quarter, income

We have been discussing the changes to the graphics market on the front page and on the Podcast, and as expected NVIDIA's income has shrunk.  Last year NVIDIA was generating $800 million but saw revenue drop bu over $100 million, in perspective SemiAccurate pegs their professional graphics division at about $200 million.  If NVIDIA is going to be able to keep their R&D team working on chips several generations ahead of the current products on the market, which they need to in order to be competitive, they had better hope that their foray into the mobile chip market is lucrative enough to pay the bills.

scroogemcduck.jpg

"Nvidia (NASDAQ:NVDA) published their results last Thursday topping analyst estimates and six days later the stock was down 10%. What happened?

The numbers were pretty good. Revenue was up and Tegra™ finally started to get traction, more than 3 times up but there are some red lights. First their revenues are down YoY. Second, their GPU business is down YoY and last, but not least, their professional business revenue is more or less flat for the last quarter."

Here is some more Tech News from around the web:

Tech Talk

Source: SemiAccurate

Gigabyte Unveils NVIDIA GeForce GTX 560 Overclock Edition Graphics Card

Subject: Graphics Cards | May 19, 2011 - 01:33 AM |
Tagged: nvidia, GTX 560, graphics

Coinciding with the NDA lift on the NVIDIA GeForce GTX 560, Gigabyte announced its enthusiast class Overclock Edition graphics card based on new the GTX 560 GPU.

GTX560GigabyteOC.jpg

The new Overclock Edition replaces the reference design's cooler with Gigabyte's own WindForce 2X variant, which they claim reduces the noise of the card under full load to 31db. Further, the heatsink used direct heat pipe technology, which means that the heat pipes that carry heat away from the GPU and into the fins physically contact the GPU itself. Both fans produce 30.5 CFM of airflow to quickly dissipate the heat of the overclocked GTX 560 GPU, Gigabyte was able to clock the card at a 830 MHz GPU clock and a 4008 Mhz memory clock from the factory. Gigabyte claims to improve overclocking capability by 10% to 30% thanks to it's "Ultra Durable" copper PCB technology and power switching enhancements.

The full specification of the GeForce GTX 560 Overclock Edition are as follows:

 

 

Model Number GV-N56GOC-1GI
Core Clock 830 MHz
Shader Clock 1660 MHz
Memory Amount 1 GB
Memory Type GDDR5
Memory Bus 256 bit
Card Bus PCI-E 2.0
Process Technology 40 nm
Card Dimensions 43mm (h) x 238mm (l) x 130mm (w)
Power Requirements Minimum 500 Watt PSU required
DirectX Support 11
Outputs

1x HDMI and Display Port via adapter(s)

1x mini HDMI

2x DVI

1x VGA (via adapter)

 

Gigabyte is a popular motherboard manufacturer for enthusiasts and it seems that they are striving to gain that same level of consumer brand loyalty with their graphics cards.  Do you have a Gigabyte graphics card in your rig?

Source: Gigabyte

Rumor: More details on Amazon tablets, yes: plural

Subject: General Tech, Mobile | May 18, 2011 - 02:05 PM |
Tagged: tegra, nvidia, kal-el, amazon

At the beginning of the month we reported that Amazon seems to be moving into the tablet space with an order for hundreds of thousands of touchscreens per month. There is now more evidence that the Kindle manufacturer is looking specifically to do an Android tablet due to the processors rumored to be included. We think you will be smiling very soon.

18-Mmm-amazon.png

Roadrunner Stew: Water, Roadrunner, Diced Apple 

According to a source from Boy Genius Report: Amazon will be releasing two tablets potentially as earlier as second-half 2011. Their entry-level model, “Coyote”, will contain an NVIDIA dual-core Tegra 2 which for those following the tablet space knows that is not an entry-level part. Their higher-up model, "Hollywood", is said to contain NVIDIA’s upcoming Kal-El, the quad-core successor to the Tegra 2 with an integrated GPU reportedly 5-fold faster than those found on the Tegra 2.
 
It should be noted that an entry-level tablet for Amazon does not necessarily suggest what price Amazon would be aiming for. Amazon could still have their Kindle line as their true entry-level product which would allow them to skip straight to the higher performing Tegra 2 and Kal-El parts for their actual tablet line as they could point to the Kindle for those who want less than a Tegra 2. So if you happen to be looking for an Android tablet and can wait a few months you might wish to hold off and see what wily Amazon is cooking up.
Source: PCMag

Fermi, Fermi, Fermi! Nobody pays attention to Tesla and the M2090 GPU Coprocessor

Subject: General Tech | May 18, 2011 - 08:39 AM |
Tagged: nvidia, gpu coprocessor, tesla

It is always the flashy brother that everyone notices, even if you've never met them ... say the GTX590.  However the other brother shouldn't be ignored because it turns out Telsa is pretty cool among the server crowd.  Where once the humble math coprocessor went the M2090 GPU coprocessor races past, with a specially made, not bin sorted 40nm Fermi GPU running at 1.3GHz and GDDR5 at 1.85GHz which can pull some interesting ECC tricks and of course a ful 512 CUDA Cores.  If you think that is a lot of power, NVIDIA told The Register they are recommending one M2090 per CPU core, not per physical CPU.  

ElReg_nvidia_tesla_m2090_gpu.jpg

"GPU chipmaker Nvidia knows that it has to do more to grow its Tesla biz than slap some passive heat sinks on a fanless GPU card and talk up its CUDA parallel-programming tools. It has to keep delivering price/performance improvements, as well.

And that's exactly what it's doing with the new Tesla M2090 GPU coprocessor."

Here is some more Tech News from around the web:

Tech Talk

Source: The Register

The vanilla GTX560 arrives, does it fit well between the GTX460 and the GTX 560 Ti?

Subject: Graphics Cards | May 17, 2011 - 11:14 AM |
Tagged: nvidia, GTX560, gtx560 ti

The GTX560 Ti did not receive glowing rewards, not because it didn't perform but because the price was too high compared to the Radeon cards it competes against.  Now with the vanilla card available at $200, and with higher peak pixel fill rates, higher rasterization rates, and more memory bandwidth than the Ti version the card is not simply squeezed into an already tight market segment but actually has some interesting abilities.  The similarly priced Asus Radeon HD 6870 TOP can't keep up with the new GTX560, but the gap is not huge.  The Tech Report recommends waiting a bit before considering this card, they feel it is likely to drop below $200 which would make it a very good deal indeed.

TR_msi560.jpg

"Say hello to Nvidia's latest $199 graphics card. Is this a worthwhile step up from the GeForce GTX 460 1GB, and is it a better deal than AMD's Radeon HD 6870?"

Here are some more Graphics Card articles from around the web:

Graphics Cards

NVIDIA To Release Updated GTX 590 Cards In June

Subject: Graphics Cards | May 17, 2011 - 10:43 AM |
Tagged: nvidia, hardware, graphics

gtx590.jpg

The current GTX 590

VR-Zone reports that NVIDIA is gearing up to deliver a revised edition GTX 590 in June to combat the overheating problems that some overclockers fell victim too using certain drivers.  PC Perspective did not run into the issue when overclocking their card; however, VR-Zone stated in an earlier article that:

"NVIDIA has sent out a cautionary to their partners regarding possible component damage due to high temperature when running Furmark 1.9 as it bypasses the capping detection. . . .  This is something not able to fix through drivers nor it is just applicable to GeForce GTX 590."

Fortunately for overclockers, NVIDIA is planning to re-engineer aspects of the design, including new inductors, which should help with the over-current protection issues.  This new design will also effect the size and dimensions of the current GTX 590 PCB, which means that current third party heat sinks and water blocks made for the (current) GTX 590 will not fit.

It is nice to see that NVIDIA is sticking by it's technology and updating its hardware to fix issues.  Overclockers especially, will benefit from this updated model.

 

Source: VR-Zone

Tidbits from NVIDIA's Q1 conference call

Subject: Editorial, General Tech, Graphics Cards, Processors, Mobile | May 13, 2011 - 03:49 PM |
Tagged: nvidia, conference call

NVIDIA made their quarterly conference call on May 12th which consisted of financial results up to May 1st and questions from financial analysts and investors. NVIDIA chief executive officer Jen-Hsun Huang projected that future revenue from the GPU market would be “flattish”, revenue from the professional market would be “flattish”, and revenue from the consumer market would be “uppish”. Huang did mention that he believes that the GPU market will grow in the future as GPUs become ever more prevalent.

13-1.jpg

How's the green giant doing this quarter? Read on for details.

For the professional market, NVIDIA discussed their intention to continue providing proof-of-concept applications to show the benefit of GPU acceleration which they hope will spur development of GPU accelerated code. Huang repetitively mentioned that the professional market desires abilities like simultaneous simulation and visualization and that a 10% code-rewrite would increase performance 500-1000%, but current uptake is not as fast as they would like. NVIDIA also hinted that GPUs will be pushed in the server space in the upcoming future but did not clarify on what that could be. NVIDIA could simply be stating that Tesla will continue to be a focus for them; they also could be hinting towards applications similar to what we have seen in recent open sourced projects.

For consumers, Huang made note of their presence in the Android market with their support of Honeycomb 3.1 and the upcoming Icecream Sandwich. Questions were posed about the lackluster sales of Tegra tablets but Huang responded stating that the first generation of tablets were deceptively undesirable due to cost of 3G service. He went on to say that the second wave of tablets will be cheaper and more available in retail stores with Wi-Fi only models more accessible to consumers.

13-2.jpg

nVihhhhhhhhhdia. (Image by Google)

The bulk of the conference call was centered on nVidia’s purchase of Icera though not a lot of details were released being that the purchase is yet to be finalized. The main points of note is that as of yet, while NVIDIA could integrate Icera’s modems onto their Tegra mobile processors, they have no intention of doing so. They also stated they currently have no intention of jumping into the other mobile chip markets such as GPS and near-field communications due to the lesser significance and greater number of competitors.

13-3.jpg

I think the new owners like the color on the logo.

The last point of note from the conference call was that they expect that Project Denver, NVIDIA’s ARM-based processor, to be about 2 generations away from accessible. They noted that they cannot comment for Microsoft but they do reiterate their support for Windows 8 and its introduction of the ARM architecture. The general theme throughout the call was that NVIDIA was confident in their position as a player in the industry. If each of their projects works out as they plan, it could be a very well justified attitude.

Source: NVIDIA