Subject: Graphics Cards | July 27, 2016 - 01:56 AM | Tim Verry
Tagged: solid state, radeon pro, Polaris, gpgpu, amd
UPDATE (July 27th, 1am ET): More information on the Radeon Pro SSG has surfaced since the original article. According to AnandTech, the prototype graphics card actually uses an AMD Fiji GPU. The Fiji GPU is paired onboard PCI-E based storage using the same PEX8747 bridge chip used in the Radeon Pro Duo. Storage is handled by two PCI-E 3.0 x4 M.2 slots that can accommodate up to 1TB of NAND flash storage. As I mentioned below, having the storage on board the graphics card vastly reduces latency by reducing the number of hops and not having to send requests out to the rest of the system. AMD had more numbers to share following their demo, however.
From the 8K video editing demo, the dual Samsung 950 Pro PCI-E SSDs (in RAID 0) on board the Radeon Pro SSG hit 4GB/s while scrubbing through the video. That same video source stored on a Samsung 950 Pro attached to the motherboard had throughput of only 900MB/s. In theory, reaching out to system RAM still has raw throughput advantages (with DDR4 @ 3200 MHz on a Haswell-E platform theroretically capable of 62 GB/s reads and 47 GB/s writes though that would be bottlenecked by the graphics card having to go over the PCI-E 3.0 x16 link and it's maximum of 15.754 GB/s.). Of course if you can hold it in (much smaller) GDDR5 (300+GB/s depending on clocks and memory bus width) or HBM (1TB/s) and not have to go out to any other storage tier that's ideal but not always feasible especially in the HPC world.
However, having onboard storage on the same board as the GPU only a single "hop" away vastly reduces latency and offers much more total storage space than most systems have in DDR3 or DDR4. In essence, the solid state storage on the graphics card (which developers will need to specifically code for) acts as a massive cache for streaming in assets for data sets and workloads that are highly impacted by latency. This storage is not the fastest, but is the next best thing for holding active data outside of GDDR5/x or HBM. For throughput intensive workloads reaching out to system RAM will be better Finally, reaching out to system attached storage should be the last resort as it will be the slowest and most latent. Several commentors mentioned using a PCI-E based SSD in a second slot on the motherboard accessed much like GPUs in CrossFire communicate now (DMA over the PCI-E bus) which is an interesting idea that I had not considered.
Per my understanding of the situation, I think that the on board SSG storage would still be slightly more beneficial than this setup but it would get you close (I am assuming the GPU would be able to directly interact and request data from the SSD controller and not have to rely on the system CPU to do this work but I may well be mistaken. I will have to look into this further and ask the experts heh). On the prototype Radeon Pro SSG the M.2 slots are actually able to be seen as drives by the system and OS so it is essentially acting as if there was a PCI-E adapter card in a slot on the motherboard holding those drives but that may not be the case should this product actually hit the market. I do question their choice to go with Fiji rather than Polaris, but it sounds like they built the prototype off of the Radeon Pro Duo platform so I suppose it would make sense there.
Hopefully the final versions in 2017 or beyond use at least Vega though :).
Alongside the launch of new Radeon Pro WX (workstation) series graphics cards, AMD teased an interesting new Radeon Pro product: the Radeon Pro SSG. This new professional graphics card pairs a Polaris GPU with up ot a terabyte of on board solid state storage and seeks to solve one of the biggest hurdles in GP GPU performance when dealing with extremely large datasets which is latency.
One of the core focuses of AMD's HSA (heterogeneous system architecture) is unified memory and the ability of various processors (CPU, GPU, specialized co-processors, et al) to work together efficiently by being able to access and manipulate data from the same memory pool without having to copy data bck and forth between CPU-accessible memory and GPU-accessible memory. With the Radeon Pro SSG, this idea is not fully realized (it is more of a sidestep), but it will move performance further. It does not eliminate the need to copy data to the GPU before it can work on it, but once copied the GPU will be able to work on data stored in what AMD describes as a one terabyte frame buffer. This memory will be solid state and very fast, but more importantly it will be able to get at the data with much lower latency than previous methods. AMD claims the solid state storage (likely NAND but they have not said) will link with the GPU over a dedicated PCI-E bus. I suppose that if you can't bring the GPU to the data, you bring the data to the GPU!
Considering AMD's previous memory champ – the Radeon W9100 – maxed out at 32GB of GDDR5, the teased Radeon Pro SSG with its 1TB of purportedly low latency onboard flash storage opens up a slew of new possibilities for researchers and professionals in media, medical, and scientific roles working with massive datasets for imaging, creation, and simulations! I expect that there are many professionals out there eager to get their hands on one of these cards! They will be able to as well thanks to a beta program launching shortly, so long as they have $10,000 for the hardware!
AMD gave a couple of examples in their PR on the potential benefits of its "solid state graphics" including the ability to image a patient's beating heart in real time to allow medical professionals to examine and spot issues as early as possible and using the Radeon Pro SSG to edit and scrub through 8K video in real time at 90 FPS versus 17 with current offerings. On the scientific side of things being able to load up entire models into the new graphics memory (not as low latency as GDDR5 or HBM certainly) will be a boon as will being able to get data sets as close to the GPU as possible into servers using GPU accelerated databases powering websites accessed by millions of users.
It is not exactly the HSA future I have been waiting for ever so impatiently, but it is a nice advancement and an intriguing idea that I am very curious to see how well it pans out and if developers and researchers will truly take advantage of and use to further their projects. I suspect something like this could be great for deep learning tasks as well (such as powering the "clouds" behind self driving cars perhaps).
Stay tuned to PC Perspective for more information as it develops.
This is definitely a product that I will be watching and I hope that it does well. I am curious what Nvidia's and Intel's plans are here as well! What are your thoughts on AMD's "Solid State Graphics" card? All hype or something promising?
Subject: General Tech, Graphics Cards | July 25, 2016 - 09:48 PM | Sebastian Peak
Tagged: siggraph 2016, Siggraph, capsaicin, amd, 3D rendering
At their Capsaicin Siggraph event tonight AMD has announced that what was previously announced as the FireRender rendering engine is being officially launched as AMD Radeon ProRender, and this is becoming open-source as part of AMD's GPUOpen initiative.
From AMD's press release:
AMD today announced its powerful physically-based rendering engine is becoming open source, giving developers access to the source code.
As part of GPUOpen, Radeon ProRender (formerly previewed as AMD FireRender) enables creators to bring ideas to life through high-performance applications and workflows enhanced by photorealistic rendering.
GPUOpen is an AMD initiative designed to assist developers in creating ground-breaking games, professional graphics applications and GPU computing applications with much greater performance and lifelike experiences, at no cost and using open development tools and software.
Unlike other renderers, Radeon ProRender can simultaneously use and balance the compute capabilities of multiple GPUs and CPUs – on the same system, at the same time – and deliver state-of-the-art GPU acceleration to produce rapid, accurate results.
Radeon ProRender plugins are available today for many popular 3D content creation applications, including Autodesk® 3ds Max®, SOLIDWORKS by Dassault Systèmes and Rhino®, with Autodesk® Maya® coming soon. Radeon ProRender works across Windows®, OS X and Linux®, and supports AMD GPUs, CPUs and APUs as well as those of other vendors.
Subject: Graphics Cards | July 25, 2016 - 09:30 PM | Sebastian Peak
Tagged: siggraph 2016, Siggraph, Radeon Pro WX Series, Radeon Pro WX 7100, Radeon Pro WX 5100, Radeon Pro WX 4100, radeon, capsaicin, amd
AMD has announced new Polaris-based professional graphics cards at Siggraph 2016 this evening, with the Radeon Pro WX 4100, WX 5100, and WX 7100 GPUs.
The AMD Radeon Pro WX 7100 GPU (Image credit: AMD)
From AMD's official press release:
AMD today unveils powerful new solutions to address modern content creation and engineering: the new Radeon Pro WX Series of professional graphics cards, which harness the award-winning Polaris architecture and is designed to deliver exceptional capabilities for the immersive computing era.
Radeon Pro solutions and the new Radeon Pro WX Series of professional graphics cards represent a fundamentally different approach for professionals rooted in a commitment to open, non-proprietary software and performant, feature-rich hardware that empowers people to create the “art of the impossible”.
The new Radeon Pro WX series graphics cards deliver on the promise of this new era of creation, are optimized for open source software, and are designed for creative professionals and those pushing the boundaries of science, technology and engineering.
The AMD Radeon Pro WX 5100 GPU (Image credit: AMD)
Radeon Pro WX Series professional graphics cards are designed to address specific demands of the modern content creation era:
- Radeon Pro WX 7100 GPU is capable of handling demanding design engineering and media and entertainment workflows and is AMD’s most affordable workstation solution for professional VR content creation.
- Radeon Pro WX 5100 GPU is the ideal solution for product development, powered by the impending game-engine revolution in design visualization.
- Radeon Pro WX 4100 GPU provides great performance in a half-height design, finally bringing mid-range application performance demanded by CAD professionals to small form factor (SFF) workstations
The AMD Radeon Pro WX 4100 GPU (Image credit: AMD)
A breakdown of the known specifications for these new GPUs was provided by AnandTech in their report on the WX Series:
Subject: General Tech | July 22, 2016 - 12:39 PM | Jeremy Hellstrom
Tagged: amd, profits
It is reasonable to expect more in depth analysis from Josh about AMD's earnings this quarter but the news is too good not to briefly mention immediately. AMD brought in $1.027 billion in revenue this quarter, a cool $68.7 million higher than expected, mostly thanks to console sales as these numbers do not include the new Polaris cards which are just being released. This is very good news for everyone, having $69 million in profit will give AMD a bit of breathing room until Polaris can start selling and Zen arrives next year. It also gives investors a boost of confidence in this beleaguered company, something that has not happened for quite a while. Drop by The Register for more numbers and a link to the slides from the AMD financial meeting from yesterday.
"AMD's share price is up more than seven per cent in after-hours trading to $5.60 at time of writing. That's agonizingly close to the magic six-buck mark for the troubled semiconductor giant that this time last year was struggling to look viable."
Here is some more Tech News from around the web:
- The cloud ain't making it rain for Intel right now: Tech giants pause server chip sales @ The Register
- MSI becomes the largest gaming notebook vendor worldwide, says paper @ DigiTimes
- All you need for quantum computing at room temperature is some mothballs @ The Register
- Stagefright-like flaw opens up iPhones and Macs to iMessage hack @ The Inquirer
- Nvidia is mildly excited about its 11 teraflop Titan X GPU and is very calm @ The Inquirer
- Sony Is the Only Remaining Obstacle To PS4-Xbox Cross-Play @ Slashdot
- Spotify Is Now Selling Your Information To Advertisers @ Slashdot
Subject: General Tech | July 21, 2016 - 01:02 PM | Jeremy Hellstrom
Tagged: rx 460, polaris 11, oculus rift, amd
TechARP spotting something unexpected at the Radeon RX 480 launch in Malaysia, a Radeon RX 460. One suspects that the picture below does not represent its final form but it does give you an idea of the dimensions and the outputs which seem to include DVI, DP and HDMI. TechARP were given some of the specs of this AMD Polaris 11 GPU based card, 14 Compute Units, 2 GB of GDDR5 memory on a 128-bit memory bus.
The biggest takeaway is what AMD was doing with it, this was powering an Oculus Rift VR demo so it is safe to say this card meets at least the minimum specs for the headset. Drop by for more pictures and a video.
"We just stumbled upon an actual Radeon RX 460 graphics card. AMD was using it to power a virtual reality demo on an Oculus VR headset. That was our first encounter with the Radeon RX 460, so we had to take off the perspex cover to take a closer look!"
Here is some more Tech News from around the web:
- Microsoft ordered to fix 'excessively intrusive, insecure' Windows 10 @ The Register
- Microsoft tweaks TCP stack in Windows Server and Windows 10 @ The Register
- Making Graphene More Practical @ Hack a Day
- Verizon Begins Charging a Fee Just to Use an Older Router @ Slashdot
- Gorilla Glass 5 promises to survive selfie-height drops - most of the time @ The Inquirer
- Digitimes Research: SoftBank chairman overoptimistic about benefits from acquiring ARM
- Really Scary Telecoms Stuff? Nah – telephony's just an app @ The Register
Subject: General Tech | July 21, 2016 - 12:21 PM | Ryan Shrout
Tagged: Wraith, Volta, video, time spy, softbank, riotoro, retroarch, podcast, nvidia, new, kaby lake, Intel, gtx 1060, geforce, asynchronous compute, async compute, arm, apollo lake, amd, 3dmark, 10nm, 1070m, 1060m
PC Perspective Podcast #409 - 07/21/2016
Join us this week as we discuss the GTX 1060 review, controversy surrounding the async compute of 3DMark Time Spy and more!!
The URL for the podcast is: http://pcper.com/podcast - Share with your friends!
- iTunes - Subscribe to the podcast directly through the Store (audio only)
- Google Play - Subscribe to our audio podcast directly through Google Play!
- RSS - Subscribe through your regular RSS reader (audio only)
- MP3 - Direct download link to the MP3 file
This episode of the PC Perspective Podcast is sponsored by Casper!
Hosts: Ryan Shrout, Allyn Malventano, Jeremy Hellstrom, and Josh Walrath
Introduction: Rethinking the Stock Cooler
AMD's Wraith cooler was introduced at CES this January, and has been available with select processors from AMD for a few months. We've now had a chance to put one of these impressive-looking CPU coolers through its paces on the test bench to see how much it improves on the previous model, and see if aftermarket cooling is necessary with AMD's flagship parts anymore.
While a switch in the bundled stock cooler might not seem very compelling, the fact that AMD has put effort into improving this aspect of their retail CPU offering is notable. AMD processors already present a great value relative to Intel's offerings for gaming and desktop productivity, but the stock coolers have to this point warranted a replacement.
Intel went the other direction with the current generation of enthusiast processors, as CPUs such as my Core i5-6600k no longer ship with a cooler of any kind. If AMD has upgraded the stock CPU cooler to the point that it now cools efficiently without significant noise, this will save buyers a little more cash when planning an upgrade, which is always a good thing.
The previous AMD stock cooler (left) and the AMD Wraith cooler (right)
A quick search for "Wraith" on Amazon yields retail-box products like the A10-7890K APU, and the FX-8370 CPU; options which have generally required an aftermarket cooler for the highest performance. In this review we’ll take a close look at the results with the previous cooler and the Wraith, and throw in results from the most popular aftermarket cooler of them all; the Cooler Master Hyper 212 EVO.
Yes, We're Writing About a Forum Post
Update - July 19th @ 7:15pm EDT: Well that was fast. Futuremark published their statement today. I haven't read it through yet, but there's no reason to wait to link it until I do.
Update 2 - July 20th @ 6:50pm EDT: We interviewed Jani Joki, Futuremark's Director of Engineering, on our YouTube page. The interview is embed just below this update.
Original post below
The comments of a previous post notified us of an Overclock.net thread, whose author claims that 3DMark's implementation of asynchronous compute is designed to show NVIDIA in the best possible light. At the end of the linked post, they note that asynchronous compute is a general blanket, and that we should better understand what is actually going on.
So, before we address the controversy, let's actually explain what asynchronous compute is. The main problem is that it actually is a broad term. Asynchronous compute could describe any optimization that allows tasks to execute when it is most convenient, rather than just blindly doing them in a row.
This is asynchronous computing.
Subject: General Tech | July 16, 2016 - 06:58 PM | Scott Michaud
Tagged: n64, dolphin, libretro, retroarch, vulkan, async shaders, asynchronous compute, amd
While the Dolphin emulator has a lot of mind share, and recently announced DirectX 12 support, they have only just recently discussed working on the open alternative, Vulkan. It looks like the LibRetro developer community will beat them with an update to RetroArch and the LibRetro API. The page for RetroArch 1.3.5 exists as of (according to Google) yesterday, but 404s, so it should be coming soon. It is still in experimental mode, but it's better than nothing.
Interestingly, they also claim that their Vulkan port of Angrylion makes use of asynchronous compute. It's unclear what it uses that for, but I'm sure it will make for interesting benchmarks.
Subject: Graphics Cards | July 16, 2016 - 01:10 AM | Tim Verry
Tagged: rx 470, rx 460, polaris 11, polaris 10, gcn4, esports, amd
At a launch event in Australia earlier this week AMD talked about its Polaris architecture, launched the RX 480 and revealed the specifications for the Polaris 10-based RX 470 and Polaris 11-derived RX 470 GPUs. The new budget GPUs are aimed at 1080p or lower gaming and will allegedly be available for purchase sometime in August.
First up is the AMD Radeon RX 470. This GPU is based on Polaris 10 (like the RX 480) but has some hardware disabled (mainly the number of stream processors). Based on the same 14nm process the GPU has 2,048 cores running at not yet known clocks. Thankfully, AMD has left the memory interface intact, and the RX 470 uses the same 256-bit memory bus pairing the GPU with 4GB of GDDR5 memory on the reference design and up to 8GB GDDR5 on partner cards.
Speaking of the reference design, the reference RX 470 will utilize a blower style cooler that AIBs can use but AMD expects that partners will opt to use their own custom dual and triple fan coolers (as would I). The card is powered by a single 6-pin power connector though, again, AIBs are allowed to design a card with more.
This card is reportedly aimed at 1080p gaming at "ultra and max settings". Video outputs will include DisplayPort 1.3/1.4 HDR support.
Breaking away from Polaris 10 is the RX 460 which is the first GPU AMD has talked about using Polaris 11. This GCNv4 architecture is similar it its larger Polaris sibling but is further cut down and engineered for low power and mobile environments. While the "full" Polaris 11 appears to have 16 CUs (Compute Units), RX 460 will feature 14 of them (this should open up opportunities for lots of salvaged dies and once yields are good enough we might see a RX 465 or something with all of its stream processors enabled). With 14 CUs, that means RX 460 has 896 stream processors (again clock speeds were not discussed) and a 128-bit memory bus. AMD's reference design will pair this card with 2GB of GDDR5 but I would not be surprised to see 4GB versions possibly in a gaming laptop SKU if only just because it looks better (heh). There is no external PCI-E power connector on this card so it will be drawing all of its power from the PCI-E slot on the motherboard.
The reference graphics card is a tiny affair with a single fan HSF and support for DP 1.3/1.4 HDR. AMD further mentions 4K H.264 / HEVC encoding/decoding support. AMD is positioning this card at HTPCs and "eSports" budget gamers.
One other tidbit of information from the announcement was that AMD reiterated their new "RX" naming scheme saying that RX would be reserved for gaming and we would no longer see R9, R7, and R5 branding though AMD did not rule out future products that would not use RX aimed at other non-gaming workloads. I would expect that this will apply to APU GPUs eventually as well.
Naturally, AMD is not talking exact shipping dates or pricing but expect them to be well under the $239 of the RX 480! I would guess that RX 470 would be around the $150 mark while RX 460 will be a sub $100 part (if only barely).
What do you think about the RX 470 and RX 460? If you are interested in watching the whole event, there is a two part video of it available on YouTube. Part 1 and Part 2 are embedded below the break.