TweakTown NewsRefine News by Category:
It looks like TSMC is now facing an issue which would affect the supply of 20nm technology for video cards with AMD or NVIDIA chips. The Taiwan-based foundry is having difficulties to meet both the chipmaker's demand for the new process, therefore delaying the productions until December 2014.
NVIDIA wanted to release high-end 20nm based Maxwell GPUs, but because of this the graphics chipmaker is forced to re-engineer its GM204 and GM206 chipsets to work with the existing 28nm process.
Because of this delay, we shouldn't be expected any newer video cards until December 2014.
If you've got a spare $1500, and were in the market for a new GPU, then you might want to consider AMD's Radeon R9 295X2. AMD has started selling its dual-GPU video card, which is now on e-tailers and retailers' shelves.
The Radeon R9 295X2 is a dual-GPU powerhouse of a video card, powered by two Hawaii-based GPUs. The Radeon R9 295X2 has a total of 5632 GCN2 stream processors, 352 TMUs, 128 ROPs, a total of 8GB of RAM which is spread across two 512-bit memory interfaces. As for cooling, we have a factory-fitted water-cooling solution build by Asetek.
Now that little devil on your shoulder? It's whispering to you: "buy two of them, you know you want to".
We've only just seen the announcement of the NVIDIA GeForce GTX Titan Z, and the proper release of the AMD Radeon R9 295X2 dual-GPU card, but now we're hearing about AMD's next-generation single GPU, the Radeon R9 390X.
If the rumors are true, the Radeon R9 390X has a codename of Bermuda, with a family name of Pirate Islands. The Pirate Islands-based GPUs will be baked onto AMD's 20nm process, succeeding the Volcanic Island GPUs we all know and love today. The Radeon R9 390X would be quite the beast, featuring 4224 Stream Processors, 264 Texture Units (TUs), 96 ROPS, a Core Clock of around 1GHz, Memory Clocks hitting 7GHz or so, and a super-wide 512-bit memory bus.
Comparing this against the Radeon R9 290X which is already powerful, the R9 390X will trump it, easily. The R9 290X for comparison sake had 2816 SPs, 176 TUs, 64 ROPS, and similar core and memory clocks depending on the card (third-party cards can overclock very high).
Details on NVIDIA's GeForce GTX 880 have begun to leak, with a report from Tyden.cz claiming that the GeForce GTX 880 will be a Maxwell-based GPU that will dominate the single GPU market. The GeForce GTX 880 will move from the family name of GK110 (the 'K' is for Kepler) to the GM204 part. GM204 is the GK110 equivalent of the Kepler family, but based on the Maxwell architecture.
The leaked specs will see the DirectX 12-capable GPU feature a streaming multiprocessor Maxwell, or SMM, SIMD design that is found in the also Maxwell-based GeForce GTX 750 Ti. The GeForce GTX 880 will just feature more of those SMMs, spread across multiple graphics processing clusters (GPCs) and see some extra performance when compared against the GTX 780 Ti. We should expect something along the lines of these specs:
- 20 nm GM204 silicon
- 7.9 billion transistors
- 3,200 CUDA cores
- 200 TMUs
- 32 ROPs
- 5.7 TFLOP/s single-precision floating-point throughput
- 256-bit wide GDDR5 memory interface
- 4 GB standard memory amount
- 238 GB/s memory bandwidth
- Clock speeds of 900 MHz core, 950 MHz GPU Boost, 7.40 GHz memory
- 230W board power
This morning Sapphire Technology launched a new version of its R9 290 Radeon GPU that features a new cooler and some factory boosted performance. The new R9 290 Vapor-X OC video card features several of Sapphire's leading technology including its high-performance Vapor-X cooling solution, which keeps the factory-overclocked card running cool.
The new R9 290 Vapor-X OC GPU is based on AMD's Graphics Core Next technology and features 2560 Stream Processors, 4GB of GDDR5 memory clocked at 1400MHz, and a high-speed engine clock of 1030MHz. Sapphire says that these features combine to create industry-leading performance, stability, and longevity. The Vapor-X Cooler utilizes a "Tri-X" structure that is made from five heat-pipes, three fans, and efficient cooling fins.
We know that the Radeon R9 295X2 is coming, very very soon, but VideoCardz.com has an exclusive piece detailing AMD's upcoming dual Hawaii-based GPU. It's quite the beast, so prepare your hearts, wallets and electric bills.
The Radeon R9 295X2 is the first dual-GPU card from AMD based on its impressive Hawaii architecture, where AMD has baked in two fully loaded Hawaii cores, instead of cutting them down for the usual thermal and power consumption issues. We have a total of 8GB of RAM spread across its impressive 512-bit memory bus.
The full specs you can see above, comparing it against the Radeon R9 290X, the just-announced GeForce GTX TITAN Z, and the slightly older but still very relevant GeForce GTX 780 Ti. The new R9 295X2 features two 8-pin PCIe power connectors, with a total TDP of 500W. This is quite high considering the R9 290X has a TDP of 250W, as does the GTX 780 Ti. But, we have two full Hawaii cores here, versus the usual cut down of cores that we see in most dual-GPU cards.
With last weeks introduction of a new GPU core from NVIDIA as well as the announcement of the new $3000 GeForce GTX TITAN Z, AMD has some catching up to do in the world of high-end PC graphics. Today the company has reveled yet another teaser that is pointing towards a premium dual-GPU video card that would most likely be based on dual Radeon R9 290X GPUs.
The company kicked off its new #2betterthan1 campaign by posting a silhouetted image of what appears to be an industrial-themed Radeon Video card on Twitter. If the speculation of dual R9 290X Hawaii GPU's on a single card is true, then that also means that AMD has designed one hell of a new cooler for the card, as it is quite difficult to cool just one of those GPUs. Could this mean that we are about to see the launch of one of the first factory liquid-cooled GPUs ever? This could be the case, but more than likely, AMD has added more heat-pipes to the existing cooler, and beefed up the fans to keep everything cool.
GTC 2014 - At the GPU Technology Conference, NVIDIA unveiled a few new technologies that could see us in a world of 8-way GPUs in consumer PCs within the next couple of years.
Right now, the limitations in place in the form of chipsets, PCIe bandwidth and power consumption simply don't allow this to happen. Starting with NVLink, which promises some impressive tech - such as 5-12x the bandwidth of PCIe, it could happen.
Then we have 3D Memory, another technology which could enable 8-way GPUs. We have incredibly fast memory bandwidth, much faster than what is available today, and it will only get better, and especially as it has 400% better energy efficiency.
Then we have Pascal, NVIDIA's next-generation GPU. It combines NVLink and 3D Memory technology, while sprinkling some magical dust on top as it is only one-third the size of a PCIe card.
GTC 2014 - NVIDIA just took the wraps off of its incredible new GPU, the GeForce GTX TITAN Z, but what better way to show it off by displaying Unreal Engine 4 running in real-time.
We have a video above, and while it's not super impressive in terms of physical destruction within the environment, the lighting and model detail is nothing short of impressive.
GTC 2014 - Jen Hsun Huang, the co-founder and CEO of NVIDIA, has just unveiled the new GeForce GTX TITAN Z GPU. What does the Z variant of an already incredible GPU do?
Well, how does 12GB of VRAM sound? GeForce GTX TITAN Z features two GK110 cores, 12GB of VRAM, 5,760 CUDA cores (2,880 cores per GPU), 8 TeraFLOPS of performance and a price tag of $2999. The TITAN Z's two GPUs run at the same clock speed, thanks to NVIDIA deploying dynamic power balancing.
This means that both GPUs will never see a performance bottleneck. We should see a card that is both cool and quiet, instead of hot and loud. NVIDIA is using low-profile components with a ducted baseplate that channels turbulence and improves acoustic quality.
We will have more on this soon, but I'm hoping I can score four of them before I go (please?).