There was a new Direct Rendering Manager (DRM) update for Linux pushed out yesterday, with AMD sliding in some major Vega feature support into the open source OS.
We now know that AMD have some interesting things planned for Vega, including GPU sensors, partial resident textures, network visualization, non-contiguous VRAM mapping, and more. But it's the internal specs of Vega 10 that we're all here for, so without getting your excitement meter up too much - please, take some of this salt and throw it over your shoulder.
Vega 10 will supposedly rock 64 next-gen compute units, each with 64 GCN stream processors - with a total of 4096 next-gen GCN stream processors in 4 divisions, each with a single shader engine. Every 1024 stream processor shader engine has two Asynchronous Compute Units, one render back end and 4 texture blocks.
Inside of each texture block are 16 texture mapping units, providing a total of 256 TMUs - while Vega 10 has the ability of supporting 8 independent work threads simultaneously. With Vega 10 clocked at 1.5GHz, we could expect a monstrous 12.5 TFLOPs of FP32 compute performance, and the high-speed 8GB of HBM2 with what I think will be the start of the show in High Bandwidth Cache (HBC), AMD could have one of the fastest graphics cards on the market with its Radeon RX Vega.
We've been hearing more and more about AMD's first truly next-gen GPU architecture in years, Vega - but now the company is getting official in PR statements for the Q1 2017 results.
In AMD's recent Q1 2017 highlights release, the company said: "AMD's "Vega" GPU architecture is on track to launch in Q2, and has been designed from scratch to address the most data- and visually-intensive next-generation workloads with key architecture advancements including: a differentiated memory subsystem, next-generation geometry pipeline, new compute engine, and a new pixel engine".
Exciting stuff, that's for sure. I'm sure we're going to see something unveiled just before, or during Computex - with a physical launch in the weeks after, sometime in late June.
In order to playback 4K content on Netflix, you need a few things: Windows 10, an Intel Kaby Lake-based CPU, and you have to use Microsoft's oh-so-meh Edge browser. If you don't, Netflix will load up to 1080p and that's it.
But, things are changing - slowly, and in not such a good way with NVIDIA preparing Netflix 4K support in the near future. In order for the Netflix 4K content support, you'll need:
- NVIDIA Driver version exclusively provided via Microsoft Windows Insider Program (currently 381.74).
- NVIDIA Pascal based GPU, GeForce GTX 1050 or greater with minimum 3GB memory
- HDCP 2.2 capable monitor(s). Please see the additional section below if you are using multiple monitors and/or multiple GPUs.
- Microsoft Edge browser or Netflix app from the Windows Store
- 25Mbps (or faster) internet connection.
NVIDIA seems to be providing very access to Netflix 4K content, as there are some large caveats right now: you need a GTX 10 series card with over 3GB of RAM, and now you can use the Netflix app from the Windows Store (instead of requiring Edge only), and you won't be able to use SLI at all.
It's not just SLI users that get the elbow, but multi-monitor users on a single GPU or multi-GPU system where the GPUs will need to not be linked together (SLI), for 4K content to flow. Secondly, all of the active monitors need to be HDCP 2.2 compliant, and if they're not (or even one of them isn't) then the Netflix quality will be locked to 1080p.
This shift is showing NVIDIA is committed to 4K content, and has officially shown AMD up with their new Radeon RX 500 series which can't run Netflix content in 4K... for now. Let's hope Vega gets some built-in Netflix 4K love.
Things aren't looking good for AMD with all of these lukewarm rumors on Radeon RX Vega, with the latest purported results of the HBM2-based graphics card barely edging out NVIDIA's GeForce GTX 1070 in 3DMark TimeSpy.
The leaked result on Radeon RX Vega was powered by AMD's own Ryzen 7 1800X processor, with the system scoring 5950. If we compare this to our results, the GTX 1070 scores 5703, while the GTX 1080 pushes 6709, and the GTX 1080 Ti leaps ahead with 8312.
AMD's upcoming Radeon RX Vega seems to have its GPU clocked at 1GHz base, and 1.2GHz boost - at least in reference form. There's 8GB of HBM2 with 512GB/sec of memory bandwidth, and a TDP of 225W. If these results are true, Vega isn't exactly a GTX 1080 killer, let alone a GTX 1080 Ti killer. But then remember that AMD could be under delivering with performance right now, and things will change with tuned drivers and final hardware.
We're beginning to hear more on AMD's upcoming Radeon RX Vega graphics card, with the TechPowerUp GPU database being updated, and the upcoming Radeon RX Vega is now listed with 'relative performance' results at 1080p against various graphics cards.
TPU's relative performance chart shows the RX Vega being 5% slower than NVIDIA's stock GeForce GTX 1080, and 27% slower than the GTX 1080 Ti, at 1080p. Radeon RX Vega beats the Radeon R9 Fury X by 31%, and is just 11% faster than the GTX 1070 at 1080p. These results should be taken with massive handfuls of salt, and real-world testing will need to happen to ensure we're not being fooled.
The listing shows that AMD has 8GB of HBM2 on a 2048-bit memory bus, with the Vega GPU clocked at 1GHz stock, while it can be boosted up to 1.2GHz in reference form. The 8GB of HBM2 is clocked at 1GHz, and shouldn't be confused with the 11GHz+ on the GDDR5X on the new GTX 1080 11Gbps and GTX 1080 Ti cards because they use GDDR5X compared to the HBM2 tech on AMD's upcoming Radeon RX Vega.
AMD is close to the release of its next-gen Radeon RX Vega graphics card, with the latest on the new Vega GPU is that there have been some Radeon RX Vega packaging spotted in a special edition, with Quake Champions.
Quake Champions is the new first-person shooter from id Software built for the PC, so it would make sense to have the most advanced GPU ever made - with its new Vega NCU and HBM2 tech - to be featured with Radeon RX Vega.
The packaging also lists DirectX 12, Vulkan, and AMD's Radeon Chill technology - which for now, shows that there are no Vega exclusive features that set it apart from the Polaris-based RX 500 series.
NVIDIA is reportedly preparing a new graphics card to compete with AMD's mid-range Radeon RX 550, in what will be the purported GeForce GT 1030, and it'll be aimed at the sub-$100 market.
NVIDIA is already in the sub-$150 market with its GeForce GTX 1050 at $109, but AMD attacked the sub-$100 market with its RX 550 at $79. AMD's Radeon RX 550 has a Polaris 12 GPU, 512 stream processors, and either 2GB or 4GB of GDDR5 RAM. Not bad for $79, but the RX 560 is only $20 more at $99. Still, the RX 550 will appeal to those who don't have the additional $20, or don't want the PCIe power connector - think internet cafes, and budget PCs.
The purported GT 1030 will feature a GP108 GPU, with 512 stream processors, 32 TMUs, 16 ROPs, and 2GB/4GB of RAM on a 128-bit memory bus. NVIDIA won't require an additional PCIe power connector, as the TDP will be just 35W... while the RX 550 chews 50W.
It looks like GALAX will be joining Colorful in the LCD display on a graphics card business, with GALAX's upcoming GeForce GTX 1080 Ti HOF set to include a LCD display.
GALAX will use the LCD display on its upcoming GTX 1080 Ti HOF to display their logo, the BIOS version, VRAM capacity, GPU model and driver installed, temperatures, fan speed, GPU voltage, GPU clock, VRAM clocks, and even custom text. It sounds dreamy.
GALAX will have the highest clocks on its GTX 1080 Ti HOF, with a huge triple-fan cooler and triple-fan design, all on a custom PCB with 8+8+8-pin PCIe power connectors. It's not meant for gamers running Overwatch, but for professional overclockers and enthusiasts.
AMD is still preparing its next-gen Radeon RX Vega graphics card, but before its big launch next month - fingers crossed - the company teased its Vega GPU was used inside of the ridiculously awesome Radeon Pro SSG graphics card, video editing 4K and 8K content at the NAB show in Las Vegas.
AMD used its Radeon Pro SSG with 8K video processing in Adobe Premiere Pro CC 2017, while also using the next-gen card in 4K post-processing with Radeon ProRender, something that's used for rendering high-end graphics. PC World reports that the NAB show is "targeted at the TV and film industry, in which 8K is a growing trend. AMD has been wooing the industry to adopt its GPUs".
8K might seem like it's years away, but with RTG boss and ice cream expert Raja Koduri saying that VR needs 16K per eye @ 240Hz for "true immersion", 8K doesn't seem that foreign anymore.
AMD has just launched its latest Radeon Pro Duo, a new dual-GPU graphics card based on the Polaris architecture. The new Radeon Pro Duo and its dual Polaris GPUs offer up to 11.45 TFLOPs of single-precision compute performance, 4608 stream processors, 32GB of GDDR5 RAM (16GB per GPU) and a TDP of 250W.
If you remember back to Capsaicin 2016, AMD launched the original Radeon Pro Duo, but the 2017 refresh is a much better design, and $500 cheaper. The original Radeon Pro Duo launched for $1499, while the new Radeon Pro Duo is priced at $999. The original Radeon Pro Duo was also much more power hungry with its 350W TDP, while the refreshed one slaps 100W off, dropping to 250W.
This isn't a gaming card by any means, just like its predecessor, with the new Radeon Pro Duo joining the ranks of the other Radeon Pro graphics cards, like the insane Radeon Pro WX 7100 graphics card. AMD's newly launched Radeon Pro Duo and its dual-GPUs are ready for intense professional/creation systems, where they can handle 8K content over a single DP1.4 cable (albeit, at just 30Hz). You can connect two DP1.4 cables for glorious, hair flying back in the wind 60FPS at 8K.