According to 3D Center Nvidia are planning a radically different architecture on Kepler. The design will retain the same basic hierarchy present on the Fermi cards (400/500 series) but will employ a different implementation of the CUDA layout. The first cards expected are going to be based off the GK104 architecture which has its specifications listed below, this will include the anticipated GTX 680. Interestingly both the HD 7970 and GTX 680 will possess 32 ROPs and 128 TMUs while in terms of Graphics cores Nvidia’s card has 1536 CUDA cores and the HD 7970 has 2048 stream processors. The GTX 680 is also only expected to pack 2GBs of GDDR5 instead of the 3 on the HD 7970, whilst the HD 7970 also contains a larger memory bus at 384bits instead of 256. Whether any of this will make a difference we are yet to see since Nvidia are claiming a performance win but the specifications seem to show it will be close.

SIMD Hierarchy

• 4 Graphics Processing Clusters (GPC)
• 4 Streaming Multiprocessors (SM) per GPC = 16 SM
• 96 Stream Processors (SP) per SM = 1536 CUDA cores

TMU / Geometry Domain

• 8 Texture Units (TMU) per SM = 128 TMUs
• 32 Raster OPeration Units (ROPs)

Memory

• 256-bit wide GDDR5 memory interface
• 2048 MB (2 GB) memory amount standard

Clocks/Other

• 950 MHz core/CUDA core (no hot-clocks)
• 1250 MHz actual (5.00 GHz effective) memory, 160 GB/s memory bandwidth
• 2.9 TFLOP/s single-precision floating point compute power
• 486 GFLOP/s double-precision floating point compute power
• Estimated die-area 340mm²

Source: 3DCenter.org

Rambus, announced today it has signed a patent license agreement with NVIDIA.

The agreement covers the use of Rambus patented innovations in a broad range of integrated circuit (IC) products offered by NVIDIA. In addition, the two companies have settled all outstanding claims, including resolution of past use of Rambus’ patented innovations. The term of this agreement is five years; other details are confidential.

“This is an important license agreement as it settles our differences and allows us to move forward with NVIDIA, the leader in visual and parallel computing,” said Harold Hughes, president and chief executive officer at Rambus. “Looking forward, we have the opportunity to focus on developing innovative solutions in concert with our licensees to help bring compelling, innovative products to market.”

We are not sure what this means for Nvidia and Nvidia’s products but some are speculating this could mean the use of Rambus’ XDR2 memory in Nvidia’s latest graphics cards instead of GDDR5.

Source: Press Release

The latest rumours on Nvidia’s Kepler 28nm architecture have been compiled into a neat table by the team over at CPU World. The table suggests that the 600 series cards will be based off three main chips the GK 104, GK 106 and GK 110. There will also be 7 single GPU cards and then 1 dual GPU card, the GTX 690. Although it is very important to stress right now that these are simply rumours about Kepler that have no specific evidence for proof. They have originated from somewhere in the Far East and were released on the same day AMD released its 2012-2013 roadmap.

CPU World summarises the performance expectations of Kepler based on these rumours:

The GTX680 will offer approx 45% better performance than AMD Radeon HD7970, which is it’s direct competitor. GTX670 will be a 20% improvement over HD7950, while GTX660Ti will perform 10% better than HD7950. GTX660 is said to perform similar to GTX580. GTX650Ti will equal the performance of GTX570, and GTX650 will replace GTX560. Finally, GTX640 will have GTX550Ti performance. This ties in with previous rumors, which have all indicated similar performance expectations.

Source: CPU-World

eTeknix Tech Review Of 2011 – The industry’s most action packed year yet?

2011 was a jam packed year for the tech industry, with exciting developments across the board from Intel’s release of the revolutionary Sandy Bridge architecture at the very start of the year to AMD’s release of the HD 7970 at the very end of the year. Not to mention everything in between including the release of the Windows 8 developer preview, the Thailand floods and AMD’s Bulldozer release. Now a month has passed from 2012 and we have had enough time to reflect on all the major developments of 2011 and would like to bring you our topical review of the year.

We will be taking a look at the year on a month by month basis giving you what we think were the most important developments in each month. As you might expect there is a particular emphasis on Intel, AMD and Nvidia as they tend to be the source of most developments within the hardware market. But we also musn’t forget the contribution given to the excitement of 2011 by the likes of HP with the whole TouchPad saga, the rather messy PR battle undertaken between the giants MSI and Gigabyte as well as the sneak peak at what is to come from Microsoft with the Windows 8 developer preview.

The best all-round Intel CPU yet?

We hope you enjoy our tech review of 2011 and here is what’s on offer:

Quarter 1

January : The Sandy Bridge launch: A revolution in the CPU market?

February: The B2 Revision Sandy Bridge Motherboard saga and the introduction of Thunderbolt to Apple computers.

March: AMD and Nvidia fight it out with the dual GPU giants: The HD 6990 and GTX 590 get launched.

Quarter 2

April: AMD board partners get existing boards ready for AM3+while AMD launches the HD 67X0 cards.

May: AMD partners debut 990FX motherboards whilst Intel hits back with the introduction of Z68.

June: AMD launches desktop APUs under the codename Llano while Nvidia dominate the mobile GPU market with the launch of the GTX 580M.

Quarter 3

July: The Thai floods send hard drive prices on the way to the roof while AMD hit back at Nvidia with the HD 6990M

August: HP announce the TouchPad cancellation whilst ASUS debut the monster MARS II graphics card.

September: Microsoft teased us with the Windows 8 developer preview whilst MSI and Gigabyte get their dirty linen out in public over PCI-E Gen3 motherboards.

Quarter 4

October: AMD launch Bulldozer, Or is that Faildozer?

November: Intel raises the bar once again as Sandy Bridge-E arrives.

December: AMD redeem themselves with launch of the world’s fastest single GPU card: the HD 7970.

AMD's saving grace?

Nvidia have released the 295.51 beta drivers for their GPUs. The aim of this new driver update is to bring significantly better performance in SLI configurations by adding new or updating existing SLI profiles. Nvidia has also continued to update its 3D Vision profiles for a very large selection of games. As usual the driver package comes in several different forms depending on your operating system so here are links to the applicable downloads:

Windows 7 or Windows Vista 32 Bit

Windows 7 or Windows Vista 64 Bit

Windows XP 32 Bit

Windows XP 64 Bit

Full release notes are as follows:

New in 295.51

NVIDIA SLI

• Adds or updates SLI profiles for :
• 7554
• Kingdoms of Amalur: Reckoning
• Mass Effect 3
• Oil Rush
• Red Faction: Armageddon

3D Vision

• Adds support for 3D Vision windowed mode on DLP HDTVs and on Optimized for GeForce passive 3D monitors
• Adds or updates 3D Vision profiles for:
• The Adventures of Tintin – rated Good
• Afterfall: Insanity – rated Fair
• Black Prophecy – rated Not Recommended
• Chaos Online – rated Fair
• Da Vinci Online – rated Fair
• DotA 2 – rating upgraded to Fair
• Dungeon and Dragons: Daggerdale – rated Fair
• Ferrari Project – rated Not Recommended
• The First Templar – rated Fair
• Fractal – rated Excellent
• FreeStyle 2 – rated Good
• The Haunted Halls Reach – rated Excellent
• King Arthur II – rated Not Recommended
• Kingdoms of Amalur: Reckoning – rated Fair
• Kong Fu Ying Xiong – rated Fair
• Love Beat – rated Good
• Marvel Super Hero Squad Online – rated Good
• Microsoft Flight – rated Fair
• Perpetuum – rated Fair
• Post Apocalyptic Mayhem – rated Fair
• Prototype 2 – rated Fair
• Q.U.B.E. – rated Excellent
• Quan Qiu Shi Ming – rated Excellent
• rFactor 2 (mod mode) – rated Good
• Saints Row: The Third – Rated Not Recommended
• Shadow Company: Left for Dead – rated Not Recommended
• Special Force – Rated Good
• Tera – rated Good
• Top Gun: Hard Lock – rated Not Recommended
• Tropico 4 – rated Good
• Worms Blast – rated Good

NVIDIA PhysX

• Updates PhysX System Software to 9.11.1111

HD Audio

• Updates HD Audio to version 1.3.12.0

Key Bug Fixes and Enhancements

• Fixes instances where the GeForce GTX 590 fan unnecessarily increased to 100%.
• Fixes instances of driver timeouts while simultaneously watching multiple videos.
• Adds edit/delete option for custom resolutions.

Other Details

• Supports multiple languages and APIs for GPU computing: CUDA C, CUDA C++, CUDA Fortran, OpenCL, DirectCompute, and Microsoft C++ AMP.
• Supports single GPU and NVIDIA SLI technology on DirectX 9, DirectX 10, DirectX 11, and OpenGL, including 3-way SLI, Quad SLI, and SLI support on SLI-certified Intel and AMD motherboards.

AMD’s 28nm technology has been out for a while and has brought substantial performance gains over 40nm technology from both AMD and Nvidia. Some are saying Nvidia will have a 28nm range of cards superior to AMD and others are saying they won’t be able to match AMD’s performance. Either way there has been no proof, no benchmarks, no footage, or anything of that nature to prove that Nvidia has one over on AMD when it comes to 28nm GPUs. According to ChipHell, one of the largest and most influential tech forums in Asia Nvidia has been up to dirty tricks across the internet in several big forums. These tricks include spreading unfounded (or unfounded to public knowledge) rumours about price, performance of the Nvidia 28nm GPUs in comparison to AMD ones in an attempt to hurt HD 7900 series graphics card sales.

Chinese forums have recognized patterns in this spurt of anti-AMD and pro-HD 7970 imaginary-competitor product posts and banned scores of user accounts engaged in this activity. There’s a very potent reason why ChipHell’s observations shouldn’t be met with cynicism, or made light of. In China, competing factions in the ever-ruling Communist party have been known to use what is known as “the 50 cent army”, where for a pittance, scores of internet forum posters can be engaged to perform coordinated PsyOps, to shape public opinion, and maintain public support. It is only natural, then, that competing companies in the extremely competitive Chinese domestic market resort to similar tactics, including those distributing NVIDIA GeForce graphics cards, or, as Chip Hell alleges, NVIDIA itself.

Source: Hard|OCP

NVIDIA today released a new version of its CUDA parallel computing platform, which will make it easier for computational biologists, chemists, physicists, geophysicists, other researchers, and engineers to advance their simulations and computational work by using GPUs.

The new NVIDIA CUDA parallel computing platform features three key enhancements that make parallel programing with GPUs easier, more accessible and faster. These include:

- Re-designed Visual Profiler with automated performance analysis, providing an easier path to application acceleration
- New compiler, based on the widely-used LLVM open-source compiler infrastructure, delivering up to 10 percent speed up in application performance
- Hundreds of new imaging and signal processing functions, doubling the size of the NVIDIA Performance Primitives (NPP) library

“The new visual profiler is amazing,” said Joshua Anderson, lead developer of the HOOMD-blue open source molecular dynamics project. “With just a few clicks, it performs an automated performance analysis of your application, highlights likely problem areas, and then provides links to best-practice suggestions on improving them. It makes it quick and easy for virtually all developers to accelerate a broad range of applications.”

“The LLVM complier gave me an almost immediate 10 percent performance speed up, just by recompiling my existing real-time financial risk analysis code,” said Gilles Civario, senior software architect at the Irish Centre for High-End Computing. “I can only imagine the additional performance gains I can achieve with additional tuning using the new CUDA release.”

Among the new features of the latest CUDA parallel computing platform release (available free of charge on the NVIDIA developer web site here) are:

New Visual Profiler – Easiest path to performance optimization

The new Visual Profiler makes it easy for developers at all experience levels to optimize their code for maximum performance. Featuring automated performance analysis and an expert guidance system that delivers step-by-step optimization suggestions, the Visual Profiler identifies application performance bottlenecks and recommends actions, with links to the optimization guides. Using the new Visual Profiler, performance bottlenecks are easily identified and actionable.

LLVM Compiler – Instant 10 percent increase in application performance

LLVM is a widely-used open-source compiler infrastructure featuring a modular design that makes it easy to add support for new programming languages and processor architectures. Using the new LLVM-based CUDA compiler, developers can achieve up to 10 percent additional performance gains on existing GPU-accelerated applications with a simple recompile.

In addition, LLVM’s modular design allows third-party software tool developers to provide a custom LLVM solution for non-NVIDIA processor architectures, enabling CUDA applications to run across NVIDIA GPUs, as well as those from other vendors. (CUDA to be available on AMD cards? Intepret this how you like)

New Image, Signal Processing Library Functions – “Drop-in” Acceleration with NPP Library

NVIDIA has doubled the size of its NPP library, with the addition of hundreds of new image and signal processing functions. This enables virtually any developer using image or signal processing algorithms to easily gain the benefit of GPU acceleration, with the simple addition of library calls into their application. The updated NPP library can be used for a wide variety of image and signal processing algorithms, ranging from basic filtering to advanced workflows.

Source: Press Release

Nvidia has been rather absent from the next-gen GPU market so far and CES 2012 saw them more centred around the mobile platform rather than its Kepler section.

Yet this didn’t stop news reporters from asking questions, for which NordicHardware got answers. While talking to a senior Nvidia official about the Kepler GPU and it’s competition, AMD’s Southern Islands, a statement caught the attention very effectively:

“Honestly, we expected more from our competitor’s new architecture.”

When stating this, the official indicated two things related:

• The performance levels showed by AMD’s Southern Islands GPU family are within the expectations Nvidia had set
• Nvidia’s new architecture will be a lot more powerful than that of Southern Islands due to the fact that Nvidia prepared Kepler to go up against a faster architecture from AMD

Nvidia seems to be determined to bring a faster and better next-gen GPU, now whether this is just publicity or fact remains to be seen.

Source: NordicHardware

Inside a NZXT H2 chassis (black or white) is installed an Intel Core i7-2600K processor on an ASRock Z68 Pro3 motherboard and cooled by a Scuhe APSALUS2 120 closed-loop water cooler.

Also present are 8GB of DDR3-1333MHz memory (2x4GB), ESLA Gladiac GeForce GTX 560 Ti 1GB graphics, an Antec Earthwatts Green 650W PSU, a Pioneer DVD writer, an Intel SSD 320 Series 80GB on which is installed the OS (Windows 7 Home Premium 64-bit) and a 2TB storage HDD holding Users and Documents folders.

Now available, the Still Tower Gamer Edition desktop PC can be bought for ¥129,780, or $1,685.

Source: TPU

While it is obvious from all forms of testing that Intel’s graphics royally fail at gaming, it is also just as apparent that their graphics cores are quite capable in a workstation environment being particularly strong at encoding and rendering. Currently Nvidia’s Quadro series and AMD’s FirePro series dominant the workstation environment. Using Nvidia’s FX 580 as a baseline (which costs rougly £160/$180) Intel claims the HD P3000 graphics it has integrated onto its newest Xeon CPUs offer the same if not better performance resulting in no need to have an FX 580 and therefore a $150-$200 saving.

This should give Intel CPUs even more of a competitive edge in the workstation environment if they didn’t have it already. Since now cheaper AMD alternatives are not that cheap as you would still need to purchase a separate GPU for them. The new Professional grade Intel graphics are going to get certification for PhotoShop, AutoCAD, Solid Edge and 3DS Max which will really give Intel a foothold in the workstation graphics market.

Source: VR-Zone

With Version 0.5.8 comes two new features along the regular fixes provided.

The first one is a render test that applies sufficient load ( but not stress) on the GPU to pull it out of PCI-Express link-state power-management, ensuring the Bus information is accurate. If information is displayed incorrectly (i.e. PCI-Express bus link speed or PCIe version displayed incorrectly) just click on the “?” button next to the field to launch the load test.

The second feature is ASIC quality, designed for NVIDIA Fermi (GF10x and GF11x GPUs) and AMD Southern Islands (HD 7800 series and above), aimed at advanced users, hardware manufacturers, and the likes. TechPowerUp found the ways in which AMD and NVIDIA segregate their freshly-made GPU ASICs based on the electrical leakages the chips produce (to increase yield by allotting them in different SKUs and performance bins), and they’ve found ways in which ASIC quality can be quantified and displayed. You’ll find this feature in the context menu of GPU-Z. They’re also currently working on implementing this feature on older AMD Radeon GPUs.

Here’s a full log of the changes made:

• Added explanation about PCI-Express power savings and 3D render test to accurately measure bus config under load
• Added function to display ASIC quality for Fermi and Southern Islands. (Located in the GPU-Z system menu)
• Fixed crash on older ATI cards
• Added voltage monitoring for HD 7970
• Improved real-time clock monitoring for HD 7970
• Fixed OpenCL detection for AMD Antilles, Whistler, Seymour, Blackcomb
• Improved default clock reading for AMD HD 7970 and Fusion
• Added support for AMD FirePro V7900, HD 6930, HD 7690M, HD 6410D
• Fixed Intel Sandy Bridge IGP to be DirectX 10.1, 32 nm
• Added support for NVIDIA Tesla C2075, GeForce GT 630M

Available now, you may download the regular GPU-Z 0.5.8 here, or choose to get the ROG theme GPU-Z 0.5.8 here.

Source: TPU

Alienware are attempting to lead the way when it comes to the compact gaming system. Their new X51 Ultra gaming systems will feature an ultra small form factor measuring in at just 343 x 95 x 318 mm (HxWxD). The X51 Ultra systems will feature Sandy Bridge processors a choice of either the i3-2120 or i7 2600. Slightly disappointing that there is no i5 2400 available or any other i5 as it means you can only go to either end of the spectrum of expensive and entry whereas the i5′s would provide a nice spot in the middle. In terms of the graphics Alienware haven’t gone mad: recognising a small form factor impedes airflow they have gone with mid range graphics cards offering either the GTX 555 or GT 545 from Nvidia (both OEM units). The front of the case features an optical drive, USB and audio ports.

The X51 Ultra comes with 4GB of 1333MHz as standard, again slightly disappointing as Sandy Bridge supports higher frequency RAM, also disappointing to see up to 8GB is the maximum available when it is so easy to get hold of a 16GB RAM kit these days. In terms of storage you can equip a 1TB mechanical drive or some SSD drives. Connectivity includes 7.1-channel HD audio with Waves MAXX Audio Software, gigabit Ethernet, Wireless b/g/n, optical and coaxial SPDIF audio outputs. The graphics card gives out HDMI and DVI outputs.

Source: Dell

The GTX 580 successor is almost here.

A rumour originating in China is picking up pace and credibility. It suggests that while Nvidia will release the entire GeForce 600 series in late March or early April, the GTX 680 which will be the flagship single GPU model is going to get an early release at the end of February. Maybe an attempt by Nvidia to claw back some of the market it is losing to the HD 7970 by offering enthusiasts the chance to get an Nvidia alternative within just a month or so.

The rumour claims GTX 680 will offer similar performance to AMD’s Radeon HD 7970. Clock speed is 780 MHz – no higher than GTX 580 – while it sports 2GB memory. While this suggests a 512-bit or 256-bit memory bus, NVIDIA has in the past mated 1GB to 192-bit with GTX 550 Ti, so it is possible GTX 680 has a 384-bit memory interface. Little else is revealed about the mysterious GTX 680.

The source of the rumour is notable forum member Napoleon from Chiphell, who has a history of leaking GPU related information. That said, this is nothing but a rumour - amidst other conflicting rumours – and must be considered as such.

Source: Chiphell via VR-Zone

According to various sources Nvidia has stated that it does not want to do a ‘Paper Launch’ of its 28nm Kepler graphics card. It wants to avoid the ‘AMD tactic’ of launching a product before there is enough stock to get to manufacturers. Therefore, Nvidia will only launch Kepler when there is enough stock to provide consumers with, and this date is predicted (by Nvidia) to be between the end of March and start of April. Obviously, Nvidia will be aiming for the late march launch so it can undermine and take a bit of hype from the Ivy Bridge launch which is scheduled for April the 8th.

The company originally scheduled GeForce Kepler for the launch in the first quarter (i.e. fiscal fourth quarter 2012), Quadro Kepler was scheduled for the second quarter (i.e. fiscal first quarter 2013) and Tesla Kepler was scheduled for the third quarter (second fiscal quarter 2013). But the new schedule could see GeForce Kepler arriving in the second quarter, putting Nvidia slightly behind schedule.

Source: VR-Zone

After mentioning more than once that it was time for Nvidia to step up, CES buzzed with talks these past days and the company’s next generation GPU architecture, both for the desktops and notebooks, is coming out in early Q2 this year.

No precise date is marked, but it has been heard several times that April was the month, coincidentally, it’s the same month Intel is planning to release Ivy Bridge. Nvidia sure took its time, AMD took the first place with the HD 7970, the first 28nm to be out ‘in the wild’. Yet Nvidia isn’t far back, with a good boost, the GTX 580 can still keep up with its rival for now.

Several sources and website around the web mention that Nvidia’s Kepler is superior to AMD’s Southern Islands, but as of today AMD is the undisputed king with the only next-gen GPU in stores. However, if Kepler does turn out to be better than Tahiti, or any other 28nm GPU from AMD, you can count on a beefy price tag.

Looking back on the previous so called “leaked roadmap”, we doubt that Nvidia will keep a scheme of “entry level first, high-end last” and will most probably counter AMD with some sort of equivalent graphics card.

April is certainly going to be a month of surprise and suspense, as was reported previously, the dual-GPU 28nm from AMD is set to launch around April as well (Radeon HD 7990 equipped with 2xHD 7970 GPUs). We certainly look forward to this new ‘clash of the titans’, mark your calendars, the showdown is coming.

Source: Fudzilla