Video Cards & GPUs News - Page 295

MSI has made its GeForce GTX 1080 Ti Gaming X graphics card official, with a proper tease at card with its spruced up cooler both on, and off - as well as the technical specifications, so we have some GPU and VRAM clock speeds to share.

MSI will be offering its usual three overclocking modes: Silent, Gaming, and OC modes. Silent Mode sees the GTX 1080 Ti Gaming X clocked at 1480/1582MHz on the GPU, and 11000MHz on the 11GB of GDDR5X. Gaming Mode ramps up the GPU clocks to 1544/1657MHz, with the GDDR5X clocks at 11016MHz (16MHz overclock), while the OC Mode clocks the card at 1569/1683MHz, and the GDDR5X at 11124MHz.

The card doesn't look as long as usual, with MSI doing some great work on the custom PCB on the GTX 1080 Ti Gaming X.

AMD hosted its Tech Summit in China recently, with CEO Lisa Su announcing their mid-range Ryzen 5 availability, with demos and more on show. There were some new details released on Radeon RX Vega, and because I foolishly had a night off from work - I missed this story, but here it is!

Scott Herkelman, the former General GeForce Manager at NVIDIA and now VP of AMD took the stage, with his segment concentrated solely on Radeon RX Vega. We now know that Radeon RX Vega will definitely come in both 4GB and 8GB variants, both running the new HBM2 memory that should hopefully pave the way for various improvements. HBM2 uses far less space on the PCB, so we should expect shorter cards that are magnitudes faster than the Radeon RX 480.

The big stress for new GPU architectures and the latest and greatest graphics cards, like NVIDIA's fresh GeForce GTX 1080 Ti and AMD's upcoming Radeon RX Vega - are the new APIs in Vulkan and DirectX 12. Futuremark has just announced that there is a new API Overhead test that replaces Mantle with Vulkan - with the API Overhead feature test running DX11, DX12, and Vulkan.

Futuremark explains in its press release: "The 3DMark API Overhead feature test measures API performance by making a steadily increasing number of draw calls. The result of the test is the number of draw calls per second achieved by each API before the frame rate drops below 30 FPS. The purpose of the test is to compare the relative performance of different APIs on a single system. The API Overhead feature test is not a general-purpose GPU benchmark, and it should not be used to compare graphics cards from different vendors. For more details, please read the 3DMark Technical Guide".

The new API Overheard test is available for 3DMark Advanced and Professional Edition, with the standalone versions of 3DMark requiring a manual update. The Steam versions will update automatically, just like my test beds all have - but the new API Overhead test is not available for the 3DMark Basic Edition, or the Steam demo.

Manli is showing off their upcoming GeForce GTX 1080 Ti graphics card, with a blower-style fan. It has the same tech specs as the GTX 1080 Ti Founders Edition from NVIDIA, with the same 11Gbps of GDDR5X clocked at 11Gbps.

Manli GeForce GTX 1080 Ti 11GB Blower specifications:

MSI will be gracing the world with no less than 5 custom GeForce GTX 1080 Ti graphics cards over the next few months, and before we jump into those models - I have a big question I need answered: where the hell are the Lightning cards, MSI? I want the GTX 1080 Lightning, and GTX 1080 Ti Lightning cards.

MSI have released previous generation cards with Lightning Edition cards, and they were usually one of the best SKUs of that generation on the market. Many people, including myself, have waited for MSI to unveil the GeForce GTX 1080 Lightning - and now that we're seeing the GeForce GTX 1080 Ti enter the market, my question on where the Lightning Edition cards are is stronger than ever. MSI... where are they? Can we have them soon? The last one we saw was MSI's GeForce GTX 980 Ti Lightning, which was one of the best GTX 980 Ti cards ever.

Now, onto MSI's custom GeForce GTX 1080 Ti graphics cards: starting with the higher-end Gaming X, which we have the SeaHawk EK, SeaHawk, Armor, and Aero brands coming soon.

EVGA is on a winning streak today with in-house overclocker 'Kingpin' hitting 3GHz+ on the EVGA GeForce GTX 1080 Ti Founders Edition, but now the company has just unleashed more details on three of its upcoming custom GTX 1080 Ti graphics cards.

EVGA will have 3 different models of custom GeForce GTX 1080 Ti cards, with the GTX 1080 Ti SC, GTX 1080 Ti SC2, and GTX 1080 Ti FTW graphics cards.

The GTX 1080 Ti FTW3 Gaming will have its clocks at 1569/1683MHz, with 8+8-pin PCIe power connectors, and a 10+2-phase design - it'll also be the most customized GTX 1080 Ti card from EVGA.

Well... I don't even know what to say. The previous world record on the GeForce GTX 1080 Ti saw the GPU hitting speeds of over 2.5GHz - but now, legendary overclocker Kingpin - who seems to not sleep, or at least overclocks GPUs in his sleep - has hit a mind blowing 3GHz on the NVIDIA GeForce GTX 1080 Ti Founders Edition.

Kingpin used LN2 to keep the GPU cool enough for it to clock past 3GHz, hitting 3024MHz on the dot - the pixel fillrate throughput reaches 229.3 GPixel/s and texture fillrate throughput of 583.7 GTexel/s.

The new 3024MHz GPU clock on the overclocked GTX 1080 Ti Founders Edition also marks a new record for the Pascal architecture, which had previously been pushed to 3012MHz on the GTX 1060.

A few days ago we reported that multi-GPU support in Vulkan was exclusive to Windows 10, but Khonos Group has reiterated that this isn't true - and that multi-GPU support under Vulkan will work with Windows 7, Windows 10, and Linux.

Khronos Group explains: "Some of the Khronos GDC slideware mentioned that for Vulkan multi-GPU functionality, Windows Display Driver Model (WDDM) must be in Linked Display Adapter (LDA) mode. That was not a very clear statement that has caused some confusion. And so it is worth clarifying that:"

They continued: "Khronos always strives to make its' specifications as cross platform as possible. Of course, what products ship on which OS is up to the implementers of each specification, but Khronos is already aware of vendor plans to ship multi-GPU functionality on platforms other than Windows 10, including Linux".

The latest rumors on AMD's upcoming Radeon RX Vega is that there will be 7 different SKUs on offer, something that was spotted by Phoronix, inside of the new Linux driver released by AMD.

In the first 100 patches released, the first Vega support was based on the GFX9 architecture, which is very different to GFX8 - or the Polaris architecture. The new architecture required around 40,000 new lines of code, something that included: "new video BIOS interface, new hardware intellectual property, support for video decode using UVD (UVD 7.0), support for video encode using VCE (VCE 4.0), support for 3D via RadeonSI, power management, full display support using DC, and support for SR-IOV virtualization".

The new patches include 7 new Vega IDs, which will include Radeon Instinct and Radeon PRO products based on the Vega architecture. As for the IDs, this is what we have:

We've been hearing rumbles of the Radeon RX 500 series for a while, but now we're getting more specific news on what GPU the refresh will use. We're still looking at the Polaris GPU architecture, but they're using a new node process, known as 14LPE and 14LPP.

AMD shifting its new Radeon RX 500 series to the 14LPE/14LPP nodes won't be the same as the larger shift from 28nm to 14nm FinFET, but it will be a nice little bump. Inside of the Radeon RX 580, we're now expecting the higher-clocked Polaris 20 XTX GPU, with 2304 stream processors, 144 TMUs, and 32 ROPs with 8GB of GDDR5 RAM at 8Gbps.