Posts tagged with graphics

NVIDIA 196.75 graphic drivers crash and burn your precious cards

Posted on March 5, 2010 3:11 AM

Word has it that NVIDIA's 196.75 graphic drivers is the main reason behind computers and graphic cards dying while someone is playing StarCraft II Beta. When contacted, a Blizzard Entertainment tech support representative did confirm that these drivers were the main cause for stopping the graphic card's fans from working, which resulted in the card itself as well as computer to overheat and shut down. It goes to show that you should always wait for a while after a new driver is released before installing one yourself to make sure that all known issues are worked out beforehand.

NVIDIA Ion, The Next-Generation

Posted on March 1, 2010 9:03 PM

NVIDIA has just made its next-generation Ion graphics processors official. You might have heard that Ion-Next was slower than the first generation Ion... well this is wrong. In fact, there are two versions of Ion, a 16-core and an 8-core versions. Obviously, rumored performance had been benchmarked on the 8-core version. Anyhow, the next-generation Ion still has the same quality than the first one: it bring high performance graphics to Netbooks, without battery life compromise when compared to a plain vanilla Netbook with Intel integrated graphics. NVIDIA is talking about 15x graphics performance against Intel's graphics chips (GMA 945GSE, Intel's GMA 3150) or 7x faster in video conversion (GPU-Badaboom versus CPU-iTunes).

NVIDIA considers external graphics accelerators

Posted on February 8, 2010 9:42 AM

For years, NVIDIA has kept eye on the potential of external graphics accelerators, although nothing concrete is mentioned just yet. Many questions still need to be answered including whether the size of its target market is large enough to make such a venture feasible. While there are many scenarios to be considered, we think that the market is effectively very small. The future of portable computing is definitely not adding stuff via an external PCI-E bus. If NVIDIA sees an market opportunity, it can deliver a product very quickly, after all external GPUs are just like PCI-E based ones. Would you get one and at what price?

MSI R5870-PM2D1G achieves highest 3DMark Vantage P-Score

Posted on December 4, 2009 11:01 AM

The MSI R5870-PM2D1G is one GPU that you really should take a look at if you're a serious gamer - after all, it is the current record holder when it comes to highest 3DMark Vantage P-Score of 24416 ever recorded based on a single GPU card, thanks to the efforts of famous overclocker Deanzo from New Zealand. Relying on the MSI Afterburner overclocking software to overclock the clock and GPU’s voltage alongside LN2 to keep proceedings running crazily cool at -180 degrees Celsius, this GPU achieved a phenomenal 1380 MHz GPU clock speed in order to break the previous 3DMark Vantage P-Score.

NVIDIA iRay, RealityServer

Posted on October 20, 2009 2:33 PM

NVIDIA has just announced a 3D cloud infrastructure initiative this morning. The idea is to render incredibly realistic graphics for a number of applications (home design, architecture, product design….) in the cloud and send the images back as a movie stream or most likely a single frame (rendering preview happens in seconds, instead of hours) to a client that has little or no 3D capabilities. The dataset, typically too large to reside on the client, would live remotely on the RealityServer, a solution that NVIDIA has been pushing for some time now.

Overall, I liked the idea and the applications looked like they could work in the real world. From Home Design to Computer Assisted Design (CAD), iRay could bring value to those who use Mental Ray (a bunch of people) for offline rendering. Keep in mind that there is probably an initial data upload time that could be considerable, depending on the project. This could be a turn-off for anyone who updates a few big files during each iteration, as the upload might outlast the rendering time.

The ATI Radeon HD 5800 is the fastest GPU in town

Posted on September 23, 2009 3:18 AM

The Next generation AMD GPU has arrived: The ATI Radeon HD 5870 should have been called Radeon 2X: when compared to the Radeon HD 4890, It has twice the number of stream processors, twice the transistors, twice the texture and pixel blending power and is worth twice as much... The theoretical numbers are impressive, and actual numbers are too. It is as fast as the GeForce GTX 295 that has two GPUs. If you have the choice, a single GPU solution is always better than a dual-GPU solution that has the same level of performance. Combine the Radeon 5870 speed with a price of $379 and you get a very compelling offering from AMD. That's not it: did we mention that this is the only DX11 GPU on the market?

By any metrics, this is the best high-end graphics card today.

MSI N260GTX Batman: Arkham Asylum

Posted on September 19, 2009 8:18 AM

Batman fans that are looking forward to the Batman: Arkham Asylum game will certainly welcome MSI’s N260GTX-T2D896 Batman: Arkham Asylum Edition. The card features a Harley Quinn image on the card, and more importantly, the game itself. Sporting the NVIDIA GTX260 GPU and 896MB of 448-bit GDDR3 high-performance memory, it’ll certainly do the game justice (providing the rest of your gaming rig is up to speed too). This card is a limited edition card with only 1600 pieces release, and will be available only in Europe, Japan, Taiwan, Australia and some countries in South East Asia. If you’ve been holding back on that graphics card upgrade, this will probably tempt you just enough to splash your hard-earned cash.

Crysis on iPhone

Posted on September 14, 2009 8:21 PM

Just how powerful are AMD's next generation GPUs? Strong enough to run Crysis-level graphics on the iPhone - as demonstrated by AMD using their next-gen platform’s 2.5 teraFLOPS of floating-point power and DirectX 11 abilities. Does this signal the death knell for dedicated portable consoles like the Sony PSP or Nintendo DSi, or could it just prove to be the catalyst that helps such dedicated machines boast even superior graphics in addition to solid gameplay without getting interrupted by calls halfway through an intense gaming session?

AMD chip drives 6 displays, might be useful for 10 people on the planet

Posted on September 11, 2009 10:42 AM

This morning, AMD demonstrated in grand fanfare its new chip's ability to drive 6 displays. And because you can pack four graphics cards in a (big) PC, that's a total of 24 screens at once, driven by a single computer. As a graphics enthusiast, I always enjoy watching technical demos like this. As a consumer, however, I think that it's pretty useless for most people. Note that the resolution for the tiled setup is "only" 6500x2500, which is a little more than four 30" displays in a quad configuration.

The demonstration was done in the contact that AMD seems ready to launch their DX11 chip ahead of Nvidia and that's where the real excitement is. Check Dean Takahashi's video in the full post.

Tim Sweeney: The End of The GPU Roadmap

Posted on August 11, 2009 11:03 AM

Epic Games founder and 3D engine architect Tim Sweeney has presented what he calls "The end of the GPU roadmap", where he essentially says that GPU as we know them are too limited, and predicts that by 2020 developers will switch to a more flexible massively parallel programming model where all fixed functionality (texture filtering, anti-aliasing, rasterization) have been replaced with a software implementation, backed by massive computing power. There's no denying that such a perspective is exciting for software engineers, and T. Sweeney commands a lot of respect: the Unreal engine was one of the best software 3D engine before we all jumped onto the GPU boat and the Unreal Engine3 is used in about 150 games today.

OpenGL jumps to 3.2 at SIGGRAPH

Posted on August 3, 2009 6:00 AM

The Khronos Group has announced OpenGL 3.2! The new version of the graphics API has new core features that were previously accessible via OpenGL Extensions. The support for anti-aliased render targets or better cubemap filtering will improve image quality, while things like depth_clamp will make it (a little) easier for developers to implement shadow volumes. There is now "core" support for Geometry Shaders, a new feature introduced in DX10 that is only marginally used.

With OpenGL 3.2, the port from DX9 and DX10 should be easier as the API exposes practically all the hardware features. Does it mean that we can get decent gaming on Mac now? OpenGL will also interact nicely with OpenCL (a general purpose compute API), which would make it close to the upcoming DX11. Note that OpenGL 3.2 runs on hardware that has shipped since 2006 (Nvidia G80 and ATI R6xx class).

iPhone SDK 3.1 Beta 3 available with augmented-reality friendly API

Posted on July 27, 2009 3:25 PM

Apple has made its SDK 3.1b3 available to developers. Today's news is that Apple is making a video API available so that developers can retrieve the video feed from the internal camera to build augmented reality apps. The last time we checked, AcrossAir is already doing this, but it might get easier, who knows. If you wonder what Augmented Reality is, check this demo of AcrossAir.

Interestingly enough, Apple has not approve existing augmented reality apps, may be because they want everyone using the same API. Here’s another example:

 

Caustic hardware ray-tracing to be shown at SIGGRAPH

Posted on July 22, 2009 3:01 PM

You might have seen the first CausticOne card on this site some time ago: it was a prototype of a hardware ray-tracing accelerator that will be demonstrated at SIGGRAPH this summer. Caustic Graphics is not the first company to try accelerating ray-racing, and one would think that it would require a massive chip with hundreds of cores to build something that works. A quick look at the card tells us that this is not what's happening: no massive chip, no super-fast graphics memory, cooling or power connector in sight. So, how do they do it?

Caustic says that it has found another way: they have realized that the data traversal for rendering was very expensive to do on a CPU and inconvenient to do on a GPU. Caustic is using a proprietary algorithm and hardware (mainly a special memory controller) to make ray-tracing faster. The hardware is there not to accelerate the actual computing (it can too), but mainly to accelerate how their algorithm works. In the end, it is true that most things are memory-bandwidth limited. It often come down to “how fast can you mova data?”