Is the GPU in Intel Kaby Lake-G More Polaris than Vega?

Subject: Graphics Cards, Processors | April 9, 2018 - 04:25 PM |
Tagged: Vega, Polaris, kaby lake-g, Intel, amd

Over the weekend, some interesting information has surfaced surrounding the new Kaby Lake-G hardware from Intel. A product that is officially called the “8th Generation Intel Core Processors with Radeon RX Vega M Graphics” is now looking like it might be more of a Polaris-based GPU than a Vega-based one. This creates an interesting marketing and technology capability discussion for the community, and both Intel and AMD, that is worth diving into.

PCWorld first posted the question this weekend, using some interesting data points as backup that Kaby Lake-G may in fact be based on Polaris. In Gordon’s story he notes that in AIDA64 the GPU is identified as “Polaris 22” while the Raven Ridge-based APUs from AMD show up as “Raven Ridge.” Obviously the device identification of a third party piece of software is a suspect credential in any situation, but the second point provided is more salient: based on the DXDiag information, the GPU on the Hades Canyon NUC powered by Kaby Lake-G does not support DirectX 12.1.

dx_diag_comparo-100754201-orig.jpg

Image source: PCWorld

AMD clearly stated in its launch of the Vega architecture last year that the new GPUs supported DX 12.1, among other features. The fact that the KBL-G part does NOT include support for it is compelling evidence that the GPU might be more similar to Polaris than Vega.

Tom’s Hardware did some more digging that was posted this morning, using a SiSoft Sandra test that can measure performance of FP16 math and FP32. For both the Radeon RX Vega 64 and 56 discrete graphics cards, running the test with FP16 math results in a score that is 65% faster than the FP32 results. With a Polaris-based graphics card, an RX 470, the scores between FP32 and FP16 were identical as the architecture can support FP16 math functions but doesn’t accelerate it with AMD’s “rapid packed math” feature (that was a part of the Vega launch).

tomsmath.jpg

Image source: Tom's Hardware

And you guessed it, the Kaby Lake-G part only runs essentially even in the FP16 mode. (Also note that AMD’s Raven Ridge APU that integrated Vega graphics does get accelerated by 61% using FP16.)

What Kaby Lake-G does have that leans toward Vega is support for HBM2 memory (which none of the Polaris cards have) and “high bandwidth memory cache controller and enhanced compute units with additional ROPs” according to the statement from Intel given to Tom’s Hardware.

It should be noted that just because the benchmarks and games that can support rapid packed math don’t take advantage of that capability today, does not mean they won’t have the capability to do so after a driver or firmware update. That being said, if that’s the plan, and even if it’s not, Intel should come out and tell the consumers and media.

The debate and accusations of conspiracy are running rampant again today with this news. Is Intel trying to pull one over on us by telling the community that this is a Vega-based product when it is in fact based on Polaris? Why would AMD allow and promote the Vega branding with a part that it knows didn’t meet the standards it created to be called a Vega architecture solution?

Another interesting thought comes when analyzing this debate with the Ryzen 7 2400G and Ryzen 5 2200G products, both of which claim to use Vega GPUs as a portion of the APU. However, without support for HBM2 or the high-bandwidth cache controller, does that somehow shortchange the branding for it? Or are the memory features of the GPU considered secondary to its design?

This is the very reason why companies hate labels, hate specifications, and hate having all of this tracked by a competent and technical media. Basically every company in the tech industry is guilty of this practice: Intel has 2-3 architectures running as “8th Generation” in the market, AMD is selling RX 500 cards that were once RX 400 cards, and NVIDIA has changed performance capabilities of the MX 150 at least once or twice.

The nature of semi-custom chips designs is that they are custom. Are the GPUs used in the PS4 and Xbox One or Xbox One X called Polaris, Vega, or something else? It would be safer for AMD and its partners to give each new product its own name, its own brand—but then the enthusiasts would want to know what it was most like, and how did it compare to Polaris, or Vega, etc.? It’s also possible that AMD was only willing to sell this product to Intel if it included some of these feature restrictions. In complicated negotiations like this one surely was, anything is feasible.

These are tough choices for companies to make. AMD loves having the Vega branding in more products as it gives weight to the development cost and time it spent on the design. Having Vega associated with more high-end consumer products, including those sold by Intel, give them leverage for other products down the road. From Intel’s vantage point using the Vega brand makes it looks like it has the very latest technology in its new processor and it can benefit from any cross-promotion that occurs around the Vega brand from AMD or its partners.

Unfortunately, it means that the devil is in the details, and the details are something that no one appears to be willing to share. Does it change the performance we saw in our recent Hades Canyon NUC review or our perspective on it as a product? It does not. But as features like Rapid Packed Math or the new geometry shader accelerate in adoption, the capability for Kaby Lake-G to utilize them is going to be scrutinized more heavily.

Source: Various

Considering picking up a vintage GPU?

Subject: Graphics Cards | April 2, 2018 - 02:36 PM |
Tagged: cryptocurrency, graphics cards

It has been a while since the Hardware Leaderboard has been updated as it is incredibly depressing to try to price out a new GPU, for obvious reasons.  TechSpot have taken an interesting approach to dealing with the crypto-blues, they have just benchmarked 44 older GPUs on current games to see how well they fare.  The cards range from the GTX 560 and HD7770 through to current model cards which are available to purchase used from sites such as eBay.  Buying a used card brings the price down to somewhat reasonable levels, though you do run the risk of getting a dead or dying card.  With interesting metrics such as price per frame, this is a great resource if you find yourself in desperate need of a GPU in the current market.  Check it out here.

2018-03-21-image.jpg

"Along with our recent editorials on why it's a bad time to build a gaming PC, we've been revisiting some older GPUs to see how they hold up in today's games. But how do you know how much you should be paying for a secondhand graphics card?"

Here are some more Graphics Card articles from around the web:

Graphics Cards

Source: TechSpot

Eight-GPU SLI in Unreal Engine 4 (Yes There Is a Catch)

Subject: Graphics Cards | March 29, 2018 - 09:52 PM |
Tagged: nvidia, GTC, gp102, quadro p6000

At GTC 2018, Walt Disney Imagineering unveiled a work-in-progress clip of their upcoming Star Wars: Galaxy’s Edge attraction, which is expected to launch next year at Disneyland and Walt Disney World Resort. The cool part about this ride is that it will be using Unreal Engine 4 with eight, GP102-based Quadro P6000 graphics cards. NVIDIA also reports that Disney has donated the code back to Epic Games to help them with their multi-GPU scaling in general – a win for us consumers… in a more limited fashion.

nvidia-2018-GTC-starwars-8-way-sli.jpg

See? SLI doesn’t need to be limited to two cards if you have a market cap of $100 billion USD.

Another interesting angle to this story is how typical PC components are contributing to these large experiences. Sure, Quadro hardware isn’t exactly cheap, but it can be purchased through typical retail channels and it allows the company to focus their engineering time elsewhere.

Ironically, this also comes about two decades after location-based entertainment started to decline… but, you know, it’s Disneyland and Disney World. They’re fine.

Source: NVIDIA

ASRock Enters Graphics Card Market With Phantom Gaming Series of AMD GPUs

Subject: Graphics Cards | March 29, 2018 - 05:45 PM |
Tagged: RX 580, RX 570, RX 560, RX 550, Polaris, mining, asrock, amd

ASRock, a company known mostly for its motherboards that was formerly an Asus sub-brand but is now an independent company owned by Pegatron since 2010 is now getting into the graphics card market with a new Phantom Gaming series. At launch, the Phantom Gaming series is comprised of four AMD Polaris-based graphics cards including the Phantom Gaming RX 550 2G and RX 560 2G on the low end and the Phantom Gaming X RX 570 8G OC and RX 580 8G OC on the mid/high end range.

Phantom Gaming X Radeon RX580 8G OC(L4).png

ASRock is using black shrouds with white accents and silver and red logos. The lower end Phantom Gaming cards utilize a single dual ball bearing fan while the Phantom Gaming X cards use a dual fan configuration. ASRock is using copper baseplates paired with aluminum heatsinks and composite heatpipes. The Phantom Gaming RX 550 and RX 560 cards use only PCI-E slot power while the Phantom Gaming X RX 570 and RX 580 cards get power from both the slot and a single 8-pin PCI-E power connector.

Video outputs include one HDMI 2.0, one DisplayPort 1.4, and one DL-DVI-D on the Phantom Gaming parts and one HDMI 2.0, three DisplayPort 1.4, and one DL-DVI-D on the higher-end Phantom Gaming X graphics cards. All of the graphics card models feature both silent and overclocked modes in addition to their out-of-the-box default clocks depending on whether you value performance or noise. Users can select which mode they want or perform a custom overclock or fan curve using ASRock's Phantom Gaming Tweak utility.

On the performance front, out of the box ASRock is slightly overclocking the Phantom Gaming X OC cards (the RX 570 and RX 580 based ones) and slightly underclocking the lower end Phantom Gaming cards (including the memory which is downclocked to 6 GHz) compared to their AMD reference specifications.

  ASRock RX 580 OC RX 580 ASRock RX 570 OC RX 570 ASRock RX 560 RX 560 ASRock RX 550 RX 550
Cores 2304 2304 2048 2048 896 896 512 512
GPU Clock (MHz) 1380 1340 1280 1244 1149 1275 1100 1183
GPU Clock OC Mode (MHz) 1435 - 1331 - 1194 - 1144 -
Memory (GDDR5) 8GB 8GB 8GB 8GB 2GB 2GB/4GB 2GB 2GB/4GB
Memory Clock (GHz) 8GHz 8GHz 7GHz 7GHz 6GHz 7GHz 6GHz 7GHz
Memory Clock OC Mode (MHz) 8320 - 7280 - 6240 - 6240 -
Texture Units 144 144 128 128 64 64 32 32
ROPs 32 32 32 32 16 16 16 16

The table above shows the comparisons between the ASRock graphics cards and their AMD reference card counterparts. Note that the Phantom Gaming RX 560 2G is based on the cut-down 14 CU (compute unit) model rather than the launch 16 CU GPU. Also, even in OC Mode, ASRock does not bring the memory up to the 7 GT/s reference spec. On the positive side, turning on OC mode does give a decent factory overclock of the GPU over reference. Also nice to see is that on the higher end "OC Certified" Phantom Gaming X cards, ASRock overclocks both the GPU and memory speeds which is often not the case with factory overclocks.

Phantom Gaming Radeon RX550 2G(L1).png

ASRock did not detail pricing with any of the launch announcement cards, but they should be coming soon with 4GB models of the RX 560 an RX 550 to follow later this year.

It is always nice to have more competition in this space and hopefully a new AIB partner for AMD helps alleviate shortages and demand for gaming cards if only by a bit. I am curious how well the cards will perform as while they look good on paper the company is new to graphics cards and the build quality really needs to be there. I am just hoping that the Phantom Gaming moniker is not an allusion to how hard these cards are going to be to find for gaming! (heh) If the rumored Ethereum ASICs do not kill the demand for AMD GPUs I do expect that ASRock will also be releasing mining specific cards as well at some point.

What are your thoughts on the news of ASRock moving into graphics cards?

Also read:

Source: Tech Report

GDC 2018: NVIDIA Adds new Ansel and Highlights features to GeForce Experience

Subject: Graphics Cards | March 21, 2018 - 09:37 PM |
Tagged: GDC, GDC 2018, nvidia, geforce experience, ansel, nvidia highlights, call of duty wwii, fortnite, pubg, tekken 7

Building upon the momentum of being included in the two most popular PC games in the world, PlayerUnknown's Battlegrounds and Fortnite, NVIDIA Highlights (previously known as ShadowPlay Highlights) is expanding to even more titles. Support for Call of Duty: WWII and Tekken 7 are available now, with Dying Light: Bad Blood and Escape from Tarkov coming soon.

For those unfamiliar with NVIDIA Highlights, it’s a feature that when integrated into a game, allows for the triggering of automatic screen recording when specific events happen. For example, think of the kill cam in Call of Duty. When enabled, Highlights will save a recording whenever the kill cam is triggered, allowing you to share exciting gameplay moments without having to think about it.

Animated GIF support has also been added to NVIDIA Highlights, allowing users to share shorter clips to platforms including Facebook, Google Photos, or Weibo.

In addition to supporting more games and formats, NVIDIA has also released the NVIDIA Highlights SDK, as well as plugins for Unreal Engine and Unity platforms. Previously, NVIDIA was working with developers to integrate Highlights into their games, but now developers will have the ability to add the support themselves.

Hopefully, these changes mean a quicker influx more titles with Highlights support, compared to the 16 currently supported titles.

In addition to enhancements in Highlights, NVIDIA has also launched a new sharing site for screen captures performed with the Ansel in-game photography tool.

The new ShotWithGeforce.com lets users upload and share their captures from any Ansel supported game.

shotwithgeforce.PNG

Screenshots uploaded to Shot With GeForce are tagged with the specific game the capture is from, making it easy for users to scroll through all of the uploaded captures from a given title.

Source: NVIDIA

NVIDIA RTX Technology Accelerates Ray Tracing for Microsoft DirectX Raytracing API

Subject: Graphics Cards | March 19, 2018 - 01:00 PM |
Tagged: rtx, nvidia, dxr

The big news from the Game Developers Conference this week was Microsoft’s reveal of its work on a new ray tracing API for DirectX called DirectX Raytracing. As the name would imply, this is a new initiative to bring the image quality improvements of ray tracing to consumer hardware with the push of Microsoft’s DX team. Scott already has a great write up on that news and current and future implications of what it will mean for PC gamers, so I highly encourage you all to read that over before diving more into this NVIDIA-specific news.

Ray tracing has been the holy grail of real-time rendering. It is the gap between movies and games – though ray tracing continues to improve in performance it takes the power of offline server farms to render the images for your favorite flicks. Modern game engines continue to use rasterization, an efficient method for rendering graphics but one that depends on tricks and illusions to recreate the intended image. Ray tracing inherently solves the problems that rasterization works around including shadows, transparency, refraction, and reflection. But it does so at a prohibitive performance cost. But that will be changing with Microsoft’s enablement of ray tracing through a common API and technology like what NVIDIA has built to accelerate it.

04.jpg

Alongside support and verbal commitment to DXR, NVIDIA is announcing RTX Technology. This is a combination of hardware and software advances to improve the performance of ray tracing algorithms on its hardware and it works hand in hand with DXR. NVIDIA believes this is the culmination of 10 years of development on ray tracing, much of which we have talked about on this side from the world of professional graphics systems. Think Iray, OptiX, and more.

RTX will run on Volta GPUs only today, which does limit usefulness to gamers. With the only graphics card on the market even close to considered a gaming product the $3000 TITAN V, RTX is more of a forward-looking technology announcement for the company. We can obviously assume then that RTX technology will be integrated on any future consumer gaming graphics cards, be that a revision of Volta of something completely different. (NVIDIA refused to acknowledge plans for any pending Volta consumer GPUs during our meeting.)

The idea I get from NVIDIA is that today’s RTX is meant as a developer enablement platform, getting them used to the idea of adding ray tracing effects into their games and engines and to realize that NVIDIA provides the best hardware to get that done.

I’ll be honest with you – NVIDIA was light on the details of what RTX exactly IS and how it accelerates ray tracing. One very interesting example I was given was seen first with the AI-powered ray tracing optimizations for Optix from last year’s GDC. There, NVIDIA demonstrated that using the Volta Tensor cores it could run an AI-powered de-noiser on the ray traced image, effectively improving the quality of the resulting image and emulating much higher ray counts than are actually processed.

By using the Tensor cores with RTX for DXR implementation on the TITAN V, NVIDIA will be able to offer image quality and performance for ray tracing well ahead of even the TITAN Xp or GTX 1080 Ti as those GPUs do not have Tensor cores on-board. Does this mean that all (or flagship) consumer graphics cards from NVIDIA will includ Tensor cores to enable RTX performance? Obviously, NVIDIA wouldn’t confirm that but to me it makes sense that we will see that in future generations. The scale of Tensor core integration might change based on price points, but if NVIDIA and Microsoft truly believe in the future of ray tracing to augment and significantly replace rasterization methods, then it will be necessary.

Though that is one example of hardware specific features being used for RTX on NVIDIA hardware, it’s not the only one that is on Volta. But NVIDIA wouldn’t share more.

The relationship between Microsoft DirectX Raytracing and NVIDIA RTX is a bit confusing, but it’s easier to think of RTX as the underlying brand for the ability to ray trace on NVIDIA GPUs. The DXR API is still the interface between the game developer and the hardware, but RTX is what gives NVIDIA the advantage over AMD and its Radeon graphics cards, at least according to NVIDIA.

DXR will still run on other GPUS from NVIDIA that aren’t utilizing the Volta architecture. Microsoft says that any board that can support DX12 Compute will be able to run the new API. But NVIDIA did point out that in its mind, even with a high-end SKU like the GTX 1080 Ti, the ray tracing performance will limit the ability to integrate ray tracing features and enhancements in real-time game engines in the immediate timeframe. It’s not to say it is impossible, or that some engine devs might spend the time to build something unique, but it is interesting to hear NVIDIA infer that only future products will benefit from ray tracing in games.

It’s also likely that we are months if not a year or more from seeing good integration of DXR in games at retail. And it is also possible that NVIDIA is downplaying the importance of DXR performance today if it happens to be slower than the Vega 64 in the upcoming Futuremark benchmark release.

05.jpg

Alongside the RTX announcement comes GameWorks Ray Tracing, a colleciton of turnkey modules based on DXR. GameWorks has its own reputation, and we aren't going to get into that here, but NVIDIA wants to think of this addition to it as a way to "turbo charge enablement" of ray tracing effects in games.

NVIDIA believes that developers are incredibly excited for the implementation of ray tracing into game engines, and that the demos being shown at GDC this week will blow us away. I am looking forward to seeing them and for getting the reactions of major game devs on the release of Microsoft’s new DXR API. The performance impact of ray tracing will still be a hindrance to larger scale implementations, but with DXR driving the direction with a unified standard, I still expect to see some games with revolutionary image quality by the end of the year. 

Source: NVIDIA

HTC announces VIVE Pro Pricing, Available now for Preorder

Subject: General Tech, Graphics Cards | March 19, 2018 - 12:09 PM |
Tagged: vive pro, steamvr, rift, Oculus, Lighthouse, htc

Today, HTC has provided what VR enthusiasts have been eagerly waiting for since the announcement of the upgraded VIVE Pro headset earlier in the year at CESthe pricing and availability of the new device.

vivepro.png

Available for preorder today, the VIVE Pro will cost $799 for the headset-only upgrade. As we mentioned during the VIVE Pro announcement, this first upgrade kit is meant for existing VIVE users who will be reusing their original controllers and lighthouse trackers to get everything up and running.

The HMD-only kit, with it's upgraded resolution and optics, is set to start shipping very soon on April 3 and can be preordered now on the HTC website.

Additionally, your VIVE Pro purchase (through June 3rd, 2018) will come with a free six-month subscription to HTC's VIVEPORT subscription game service, which will gain you access to up to 5 titles per month for free (chosen from the VIVEPORT catalog of 400+ games.)

There is still no word on the pricing and availability of the full VIVE Pro kit including the updated Lighthouse 2.0 trackers, but it seems likely that it will come later in the Summer after the upgrade kit has saturated the market of current VIVE owners.

As far as system requirements go, the HTC site doesn't list any difference between the standard VIVE and the VIVE Pro. One change, however, is the lack of an HDMI port on the new VIVE Pro link box, so you'll need a graphics card with an open DisplayPort 1.2 connector. 

Source: HTC

Asus Introduces Gemini Lake-Powered J4005I-C Mini ITX Motheboard

Subject: Graphics Cards, Motherboards | March 8, 2018 - 02:55 AM |
Tagged: passive cooling, mini ITX, j4005i-c, Intel, gemini lake, fanless, asus

Asus is launching a new Mini ITX motherboard packing a passively-cooled Intel Celeron J4005 "Gemini Lake" SoC. The aptly-named Asus Prime J4005I-C is aimed at embedded systems such as point of sale machines, low end networked storage, kiosks, and industrial control and monitoring systems and features "5x Protection II" technology which includes extended validation and compatibility/QVL testing, overcurrent and overvoltage protection, network port surge protection, and ESD resistance. The board also features a EUFI BIOS with AI Suite.

Asus Prime J4005I-C Mini ITX Gemini Lake Motherboard.jpg

The motherboard features an Intel Celeron J4005 processor with two cores (2.0 GHz base and 2.7 GHz boost), 4MB cache, Intel UHD 600 graphics, and a 10W TDP. The SoC is passively cooled by a copper colored aluminum heatsink. The processor supports up to 8GB of 2400 MHz RAM and the motherboard has two DDR4 DIMM slots. Storage is handled by two SATA 6 Gbps ports and one M.2 slot (PCI-E x2) for SSDs. Further, the Prime J4005I-C has an E-key M.2 slot for WLAN and Bluetooth modules (PCI-E x2 or USB mode) along with headers for USB 2.0, USB 3.1 Gen 1, LVDS, and legacy LPT and COM ports.

Rear I/O includes two PS/2, two USB 2.0, one Gigabit Ethernet (Realtek RTL8111H), two USB 3.1 Gen 1, one HDMI, one D-SUB, one RS232, and three audio ports (Realtek ALC887-UD2).

The motherboard does not appear to be for sale yet in the US, but Fanless Tech notes that is is listed for around 80 euros overseas (~$100 USD). More Gemini Lake options are always good, and Asus now has one with PCI-E M.2 support though I see this board being more popular with commercial/industrial sectors than enthusiasts unless it goes on sale.

Source: Asus

AMD Project ReSX brings performance and latency improvements to select games

Subject: Graphics Cards | March 5, 2018 - 06:07 PM |
Tagged: amd, radeon, Adrenalin, resx

We all know that driver specific and per-game optimization happens for all major GPU vendors, including AMD and NVIDIA, but also Intel, and even mobile SoC vendors. Working with the game developers and tweaking your own driver is common practice to helping deliver the best possible gaming experience to your customers.

During the launch of the Radeon Vega graphics cards, AMD discussed with the media an initiative to lower the input latency for some key, highly sensitive titles. Those mostly focused around the likes of Counter-Strike: GO, DOTA 2, League of Legends, etc. They targeted very specific use cases, low-hanging fruit, which the engineering team had recognized could improve gameplay. This included better management of buffers and timing windows to decrease the time from input to display, but had a very specific selection of games and situations it could address.

And while AMD continues to tout its dedication to day-zero driver releases and having an optimized gaming experience for Radeon users on the day of release of a new major title, AMD apparently saw fit to focus a portion of its team on another specific project, this time addressing what it called “the best possible eSports experience.”

So Project ReSX was born (Radeon eSports Experience). Its goal was to optimize performance for some of the “most popular” PC games for Radeon GPUs. The efforts included both driver-level fixes, tweaks, and optimizations, as well as direct interaction with the game developer themselves. Depending on the level of involvement that the dev would accept, AMD would either help optimize the engine and game code itself locally or would send out AMD engineering talent to work with the developer on-site for some undisclosed period of time to help address performance concerns.

Driver release 18.3.1 which is posted on AMD’s website right now, integrates these fixes that the company says are available immediately with some titles and will be “rolling into games in the coming weeks.”

Results that AMD has shared look moderately impressive.

amd_resx_table.jpg

In PUBG, for example, AMD is seeing an 11% improvement in average frame rate and a 9% improvement in the 99th percentile frame time, an indicator of smoothness. Overwatch and DOTA2 are included as well though the numbers are bit lower at 3% and 6%, respectively, in terms of average frame rate. AMD claims that the “click to response” measurement (using high speed cameras for testing) was as much as 8% faster in DOTA 2.

This is great news for Radeon owners, and not just RX 580 customers. AMD’s Scott Wasson told me that if anything, the gaps may widen with the Radeon Vega lineup but that AMD wanted to focus on where the graphics card lineup struggled more with this level of game. PLAYERUNKNOWN’S BATTLEGROUNDS is known to be a highly unoptimized game, and seeing work from AMD on the driver and at the developer relations level is fantastic.

However, there are a couple of other things to keep in mind. These increases in performance are in comparison to the 17.12.1 release, which was the first Adrenalin launch driver in December of last year. There have been several drivers released between now and today, so we have likely seen SOME of this increase along the way.

Also, while this initiative and project are the right track for AMD to be on, the company isn’t committing to any future releases along these veins. To me, giving this release and direction some kind of marketing name and calling it a “project” indicates that there is or will be continued work on this front: key optimizations and developer work for very popular titles even after the initial launch window. All I was told today was that “there may be” more coming down the pipeline but they had nothing to announce at this time. Hmph.

Also note that NVIDIA hasn’t been sitting idle during this time. In fact, the last email I received from NVIDIA’s driver team indicates that it offers “performance improvements in PlayerUnknown’s Battlegrounds (PUBG), which exhibits performance improvements up to 7% percent” with driver 391.01. In fact, the website lists a specific table with performance uplifts:

nvtable.png

While I am very happy to see AMD keeping its continued software promise for further development and optimization for current customers going strong, it simply HAS TO if it wants to keep pace with the efforts of the competition.

All that being said – if you have a Radeon graphics card and plan on joining us to parachute in for some PUBG matches tonight, go grab the new driver immediately!

Source: Radeon

NVIDIA GeForce 391.05 Hotfix Driver Available

Subject: Graphics Cards | March 4, 2018 - 02:02 PM |
Tagged: nvidia, hotfix, graphics drivers

NVIDIA has published a hotfix driver, 391.05, for a few issues that didn’t make it into the recently released 391.01 WHQL version. Specifically, if you are experiencing any of the following issues, then you can go to the NVIDIA forums and follow the link to their associated CustHelp page:

  • NVIDIA Freestyle stopped working
  • Display corruption on Titan V
  • Support for Microsoft Surface Book notebooks

nvidia-2015-bandaid.png

While improved support for the Titan V and the Microsoft Surface Book is very important for anyone who owns those devices, NVIDIA Freestyle is an interesting one for the masses. NVIDIA allows users to hook the post processing stage of various supported games and inject their own effects. The feature launched in January and it is still in beta, but lead users still want it to work of course. If you were playing around with this feature and it stopped working on 390-based drivers, then check out this hotfix.

For the rest of us? Probably a good idea to stay on the official drivers. Hotfixes have reduced QA, so it’s possible that other bugs were introduced in the process.

Source: NVIDIA

Bitmain could create headaches for NVIDIA, AMD, and Qualcomm

Subject: Graphics Cards | February 28, 2018 - 09:04 PM |
Tagged: bitmain, bitcoin, qualcomm, nvidia, amd

This article originally appeared in MarketWatch.

Research firm Bernstein recently published a report on the profitability of Bitmain Technologies, a secretive Chinese company with a huge impact on the bitcoin and cryptocurrency markets.

With estimated 2017 profits ranging from $3 billion to $4 billion, the size and scope of Beijing-based Bitmain is undeniable, with annual net income higher than some major tech players, including Nvidia and AMD. The privately held company, founded five years ago, has expanded its reach into many bitcoin-based markets, but most of its income stems from the development and sale of dedicated cryptocurrency mining hardware.

There is a concern that the sudden introduction of additional companies in the chip-production landscape could alter how other players operate. This includes the ability for Nvidia, AMD, Qualcomm and others to order chip production from popular semiconductor vendors at the necessary prices to remain competitive in their respective markets.

Bitmain makes most of its income through the development of dedicated chips used to mine bitcoin. These ASICs (application-specific integrated circuits) offer better performance and power efficiency than other products such as graphics chips from Nvidia and AMD. The Bitmain chips are then combined into systems called “miners” that can include as many as 250 chips in a single unit. Those are sold to large mining companies or individuals hoping to turn a profit from the speculative cryptocurrency markets for prices ranging from a few hundred to a few thousand dollars apiece.

Bitcoin mining giant

Bernstein estimates that as much as 70%-80% of the dedicated market for bitcoin mining is being addressed by Bitmain and its ASIC sales.

Bitmain has secondary income sources, including running mining pools (where groups of bitcoin miners share the workload of computing in order to turn a profit sooner) and cloud-based mining services where customers can simply rent mining hardware that exists in a dedicated server location. This enables people to attempt to profit from mining without the expense of buying hardware directly.

bitmain.JPG

A Bitmain Antminer

The chip developer and mining hardware giant has key advantages for revenue growth and stability, despite the volatility of the cryptocurrency market. When Bitmain designs a new ASIC that can address a new currency or algorithm, or run a current coin algorithm faster than was previously possible, it can choose to build its Antminers (the brand for these units) and operate them at its own server farms, squeezing the profitability and advantage the faster chips offer on the bitcoin market before anyone else in the ecosystem has access to them.

As the difficulty of mining increases (which occurs as higher-performance mining options are released, lowering the profitability of older hardware), Bitmain can then start selling the new chips and associated Antminers to customers, moving revenue from mining directly to sales of mining hardware.

This pattern can be repeated for as long as chip development continues, giving Bitmain a tremendous amount of flexibility to balance revenue from different streams.

Imagine a situation where one of the major graphics chip vendors exclusively used its latest graphics chips for its own services like cloud-compute, crypto-mining and server-based rendering and how much more valuable those resources would be — that is the power that Bitmain holds over the bitcoin market.

Competing for foundry business

Clearly Bitmain is big business, and its impact goes well beyond just the bitcoin space. Because its dominance for miners depends on new hardware designs and chip production, where performance and power efficiency are critical to building profitable hardware, it competes for the same foundry business as other fabless semiconductor giants. That includes Apple, Nvidia, Qualcomm, AMD and others.

Companies that build ASICs as part of their business model, including Samsung, TSMC, GlobalFoundries and even Intel to a small degree, look for customers willing to bid the most for the limited availability of production inventory. Bitmain is not restricted to a customer base that is cost-sensitive — instead, its customers are profit-sensitive. As long as the crypto market remains profitable, Bitmain can absorb the added cost of chip production.

Advantages over Nvidia, AMD and Qualcomm

Nvidia, AMD and Qualcomm are not as flexible. Despite the fact that Nvidia can charge thousands for some of its most powerful graphics chips when targeting the enterprise and machine-learning market, the wider gaming market is more sensitive to price changes. You can see that in the unrest that has existed in the gaming space as the price of graphics cards rises due to inventory going to miners rather than gamers. Neither AMD nor Nvidia will get away with selling graphic cards to partners for higher prices and, as a result, there is a potential for negative market growth in PC gaming.

If Bitmain uses the same foundry as others, and is willing to pay more for it to build their chips at a higher priority than other fabless semiconductor companies, then it could directly affect the availability and pricing for graphics chips, mobile phone processors and anything else built at those facilities. As a result, not only does the cryptocurrency market have an effect on the current graphics chip market for gamers by causing shortages, but it could also impact future chip availability if Bitmain (and its competitors) are willing to spend more for the advanced process technologies coming in 2018 and beyond.

Still, nothing is certain in the world of bitcoin and cryptocurrency. The fickle and volatile market means the profitability of Bitmain’s Antminers could be reduced, lessening the drive to pay more for chips and production. There is clearly an impact from sudden bitcoin value drops (from $20,000 to $6,000 as we see saw this month) on mining hardware sales, both graphics chip-based and ASIC-based, but measuring that and predicting it is a difficult venture.

Source: MarketWatch

NVIDIA Job Posting for Metal and OpenGL Engineer

Subject: Graphics Cards | February 18, 2018 - 02:54 PM |
Tagged: opengl, nvidia, metal, macos, apple

Just two days ago, NVIDIA has published a job posting for a software engineer to “implement and extend 3D graphics and Metal”. Given that they specify the Metal API, and they want applicants who are “Experienced with OSX and/or Linux operating systems”, it seems clear that this job would involve macOS and/or iOS.

First, if this appeals to any of our readers, the job posting is here.

Apple-logo.png

Second, and this is where it gets potentially news-worthy, is that NVIDIA hasn’t really done a whole lot on Apple platforms for a while. The most recent NVIDIA GPU to see macOS is the GeForce GTX 680. It’s entirely possible that NVIDIA needs someone to fill in and maintain those old components. If that’s the case? Business as usual. Nothing to see here.

The other possibility is that NVIDIA might be expecting a design win with Apple. What? Who knows. It could be something as simple as Apple’s external GPU architecture allowing the user to select their own add-in board. Alternatively, Apple could have selected an NVIDIA GPU for one or more product lines, which they have not done since 2013 (as far as I can tell).

Apple typically makes big announcements at WWDC, which is expected in early June, or around the back-to-school season in September. I’m guessing we’ll know by then at the latest if something is in the works.

Source: NVIDIA

Radeon Software Adrenaline Edition 18.2.2 Released

Subject: Graphics Cards | February 14, 2018 - 07:00 PM |
Tagged: amd, graphics drivers

AMD has published a new version of their Radeon Software Adrenaline Edition graphics drivers. This one focuses on Kingdom Come: Deliverance, Fortnite, and PlayerUnknown’s Battleground. The first one, Kingdom Come: Deliverance, is an action RPG from Deep Silver and Warhorse Studios. It is the studio’s first game, although its founders came from 2K and Bohemia Interactive.

amd-2016-rx480-candid.jpg

AMD is quoting frame rate increases in the range of ~3-4% with this driver, although PubG can see up to 7% if you compare it with 17.12.1. They don’t seem to list any fixes, although there’s a handful of known issues, like FreeSync coming online during Google Chrome video playback, refreshing incorrectly and causing flicker. There’s also a system hang that could occur when twelve GPUs are performing a compute task. I WONDER WHAT CONDITIONS WOULD CAUSE THAT.

You can pick up the latest driver from AMD’s website.

Source: AMD

Valve Supporting AMD's GPU-Powered TrueAudio Next In Latest Steam Audio Beta

Subject: General Tech, Graphics Cards | February 7, 2018 - 09:02 PM |
Tagged: VR, trueaudio next, TrueAudio, steam audio, amd

Valve has announced support for AMD's TrueAudio Next technology in its Steam Audio SDK for developers. The partnership will allow game and VR application developers to reserve a portion of a GCN-based GPU's compute units for audio processing and increase the quality and quantity of audio sources as a result. AMD's OpenCL-based TrueAudio Next technology can run CPUs as well but it's strength is in the ability to run on a dedicated portion of the GPU to improve both frame times and audio quality since threads are not competing for the same GPU resources during complex scenes and the GPU can process complex audio scenes and convolutions much more efficiently than a CPU (especially as the number of sources and impulse responses increase) respectively.

AMD True Audio Next In Steam Audio.jpg

Steam Audio's TrueAudio Next integration is being positioned as an option for developers and the answer to increasing the level of immersion in virtual reality games and applications. While TrueAudio Next is not using ray tracing for audio, it is physics-based and can be used to great effect to create realistic scenes with large numbers of direct and indirect audio sources, ambisonics, increased impulse response lengths, echoes, reflections, reverb, frequency equalization, and HRTF (Head Related Transfer Function) 3D audio. According to Valve indirect audio from multiple sources with convolution reverb is one of the most computationally intensive parts of Steam Audio, and TAN is able to handle it much more efficiently and accurately without affecting GPU frame times and freeing the CPU up for additional physics and AI tasks which it is much better at anyway. Convolution is a way of modeling and filtering audio to create effects such as echoes and reverb. In the case of indirect audio, Steam Audio uses ray tracing to generate an impulse response (it measures the distance and path audio would travel from source to listener) and then convolution is used to generate a reverb effect which, while very accurate, can be quite computationally intensive with it requiring hundreds of thousands of sound samples. Ambisonics further represent the directional nature of indirect sound which helps to improve positional audio and the immersion factor as sounds are more real-world modeled.

GPU Convolution Performance versus CPU.png

GPU versus CPU convolution (audio filtering) performance. Lower is better.

In addition to the ability of developers to dedicate a portion (up to 20 to 25%) of a GPU's compute units to audio processing, developers can enable/disable TrueAudio processing including the level of acoustic complexity and detail on a scene-by-scene basis. Currently it appears that Unity, FMOD Studio, and C API engines can hook into Steam Audio and the TrueAudio Next features, but it remains up to developers to use the features and integrate them into their games.

Note that GPU-based TrueAudio Next requires a GCN-based graphics card of the RX 470, RX 480, RX 570, RX 580, R9 Fury, R9 Fury X, Radeon Pro Duo, RX Vega 56, and RX Vega 64 variety in order to work, so that is a limiting factor in adoption much like the various hair and facial tech is for AMD and NVIDIA on the visual side of things where the question of is the target market large enough to encourage developers to put in the time and effort to enable X optional feature arises.

I do not pretend to be an audio engineer, nor do I play a GPU programmer on TV but more options are always good and I hope that developers take advantage of the resource reservation and GPU compute convolution algorithms of TrueAudio Next to further the immersion factor of audio as much as they have the visual side of things. As VR continues to become more relevant I think that developers will have to start putting more emphasis on accurate and detailed audio and that's a good thing for an aspect of gaming that has seemingly taken a backseat since Windows Vista. 

What are your thoughts on the state of audio in gaming and Steam Audio's new TrueAudio Next integration?

Also read:

Source: Valve

MSI Launches Radeon RX 580 Armor MK2 Graphics Cards

Subject: Graphics Cards | February 3, 2018 - 05:00 PM |
Tagged: RX 580, msi, GDDR5, factory overclocked, amd, 8gb

MSI is updating its Radeon RX 580 Armor series with a new MK2 variant (in both standard and OC editions) that features an updated cooler with red and black color scheme and a metal backplate along with Torx 2.0 fans.

MSI Radeon RX 580 Armor MK2 OC Graphics Card.png

The graphics card is powered by a single 8-pin PCI-E power connection and has two DisplayPort, two HDMI, and one DVI display output. MSI claims the MK2 cards use its Military Class 4 hardware including high end solid capacitors. The large heatsink features three copper heatpipes and a large aluminum fin stack. It appears that the cards are using the same PCB as the original Armor series but it is not clear from MSI’s site if they have dome anything different to the power delivery.

The RX 580 Polaris GPU is running at a slight factory overclock out of the box with a boost clock of up to 1353 MHz (reference is 1340) for the standard edition and at up to 1366 MHz for the RX 580 Armor MK2 OC Edition. The OC edition can further clock up to 1380 MHz when run in OC mode using the company’s software utility (enthusiasts can attempt to go beyond that but MSi makes no guarantees). Both cards come with 8GB of GDDR5 memory clocked at the reference 8GHz.

MSI did not release pricing or availability but expect them to be difficult to find and for well above MSRP when they are in stock  If you have a physical Microcenter near you, it might be worth watching for one of these cards there to have a chance of getting one closer to MSRP.

Source: MSI

Up next on things you can't have, the GIGABYTE AORUS GTX 1080 Ti Waterforce Xtreme

Subject: Graphics Cards | January 30, 2018 - 04:57 PM |
Tagged: gigabyte, aorus, gtx 1080 ti, waterforce extreme edition, watercooling, factory overclocked

On the odd occasion it is in stock, the GIGABYTE AORUS GTX 1080 Ti Waterforce Xtreme will cost you $1300 or more, about twice what the MSRP is.  The liquid cooled card does come with overclocking, Gaming mode offers 1607MHz Base and 1721MHz Boost Clock, OC mode is 1632MHz Base and 1746MHz Boost Clock.  [H]ard|OCP managed to hit an impressive 2038MHz Base, 2050MHz Boost with 11.6GHz VRAM.  Check out the full review to see what that did for its performance.

1516441630zbp5jc8wvb_1_3_l.png

"GIGABYTE has released a brand new All-In-One liquid cooled GeForce GTX 1080 Ti video card with the AORUS Waterforce Xtreme Edition video card. This video card gives the Corsair Hydro GFX liquid cooled video card some competition, with a higher out-of-box clock speed we’ll see how fast this video card is and if there is any room left for overclocking."

Here are some more Graphics Card articles from around the web:

Graphics Cards

 

Source: [H]ard|OCP

SK Hynix Launches Its 8Gb GDDR6 Memory Running at 14 Gbps

Subject: Graphics Cards, Memory | January 24, 2018 - 11:04 PM |
Tagged: SK Hynix, graphics memory, gddr6, 8gb, 14Gbps

SK Hynix recently updated its product catalog and announced the availability of its eight gigabit (8 Gb) GDDR6 graphics memory. The new chips come in two SKUs and three speed grades with the H56C8H24MJR-S2C parts operating at 14 Gbps and 12 Gbps and the H56C8H24MJR-S0C operating at 12 Gbps (but at higher voltage than the -S2C SKU) and 10 Gbps. Voltages range from 1.25V for 10 Gbps and either 1.25V or 1.35V for 12 Gbps to 1.35V for 14 Gbps. Each 8 Gb GDDR6 memory chip holds 1 GB of memory and can provide up to 56 GB/s of per-chip bandwidth.

SK Hynix logo.jpg

While SK Hynix has a long way to go before competing with Samsung’s 18 Gbps GDDR6, its new chips are significantly faster than even its latest GDDR5 chips with the company working on bringing 9 Gbps and 10 Gbps GDDR5 to market. As a point of comparison, its fastest 10 Gbps GDDR5 would have a per chip bandwidth of 40 GB/s versus its 14 Gbps GDDR6 at 56 GB/s. A theoretical 8GB graphics card with eight 8 Gb chips running at 10 Gbps on a 256-bit memory bus would have maximum bandwidth of 320 GB/s. Replacing the GDDR5 with 14 Gbps GDDR6 in the same eight chip 256-bit bus configuration, the graphics card would hit 448 GB/s of bandwidth. In the Samsung story I noted that the Titan XP runs 12 8 Gb GDDR5X memory chips at 11.4 Gbps on a 384-bit bus for bandwidth of 547 GB/s. Replacing the G5X with GDDR6 would ramp up the bandwidth to 672 GB/s if running the chips at 14 Gbps.

Theoretical Memory Bandwidth
Chip Pin Speed Per Chip Bandwidth 256-bit bus 384-bit bus 1024-bit (one package) 4096-bit (4 packages)
10 Gbps 40 GB/s 320 GB/s 480 GB/s    

12 Gbps

48 GB/s 384 GB/s 576 GB/s    
14 Gbps 56 GB/s 448 GB/s 672 GB/s    
16 Gbps 64 GB/s 512 GB/s 768 GB/s    
18 Gbps 72 GB/s 576 GB/s 864 GB/s    
HBM2 2 Gbps 256 GB/s     256 GB/s 1 TB/s

GDDR6 is still a far cry from High Bandwidth Memory levels of performance, but it is much cheaper and easier to produce. With SK Hynix ramping up production and Samsung besting the fastest 16 Gbps G5X, it is likely that the G5X stop-gap will be wholly replaced with GDDR6 and things like the upgraded 10 Gbps GDDR5 from SK Hynix will pick up the low end. As more competition enters the GDDR6 space, prices should continue to come down and adoption should ramp up for the new standard with the next generation GPUs, game consoles, network devices, ect. using GDDR6 for all but the highest tier prosumer and enterprise HPC markets.

Also read:

AMD Hires Two Graphics Execs to Help Tackle NVIDIA

Subject: Graphics Cards | January 23, 2018 - 05:10 PM |
Tagged: amd, radeon, radeon technologies group, rtg

The following story was originally posted on ShroutResearch.com.

AMD announced today that it has hired two new executives to run its graphics division after the departure of Radeon Technologies Group’s previous lead. Raja Koduri left AMD in November to join Intel and launch its new Core and Visual Computing group, creating a hole in the leadership of this critical division at AMD. CEO Lisa Su filled in during Koduri’s sabbatical and subsequent exit, but the company had been searching for the right replacements since late last year.

Appointed as the senior vice president and GM of the Radeon Technologies Group, Mike Rayfield comes to AMD from previous stints at both Micron and NVIDIA. Rayfield will cover all aspects of the business management of AMD’s graphics division, including consumer, professional, game consoles, and the semi-custom division that recently announced a partnership with Intel. At Micron he served as the senior vice president of the Mobile Business Unit, responsible for company’s direction in working with wireless technology providers (smart phones, tablets, etc.) across various memory categories. While at NVIDIA, Rayfield was the general manager of the Mobile Business Unit helping to create the Tegra brand and products. Though in a different division at the time, Rayfield’s knowledge and experience in the NVIDIA organization may help AMD better address the graphics markets.

11471-amd-logo-1260x709.jpg

David Wang is now the senior vice president of engineering for the AMD Radeon Technologies Group and is responsible for the development of new graphics architectures, the hardware and software that integrate them, and the future strategy of where AMD will invest in graphics R&D. Wang is an alumni of AMD, working as corporate vice president for graphics IP and chip development before leaving in 2012 for Synaptics. David has more than 25 years of graphics and silicon experience, starting at LSI Logic, through ArtX, then ATI, before being acquired by AMD.

The hires come at a critical time for AMD. Though the processor division responsible for the Zen architecture and Ryzen/EPYC processors continues to make strong movement against the Intel dominated space, NVIDIA’s stranglehold on the graphics markets for gaming, machine learning, and autonomous driving are expanding the gap between the graphics chip vendors. The Vega architecture was meant to close it (at least somewhat) but NVIDIA remains the leader in the space by a not insignificant margin. Changing that is and should be AMD’s primary goal for the next few years.

AMD is hoping that by creating this two-headed spear of leadership for its Radeon graphics division it can get the group back on track. Rayfield will be taking over all business aspects of the graphics portion of AMD and that includes the addition of the semi-custom segment, previously a part of the EESC (Enterprise, Embedded, and Semi-Custom) group under senior vice president Forrest Norrod. AMD believes that with the growth and expansion of the enterprise segment with its EPYC processor family, and because the emphasis on the semi-custom group continues to be the advantage AMD holds in its graphics portfolio, the long-term strategy can be better executed with that group under the Radeon Technologies umbrella.

The return of Wang as the technical lead for the graphics division could bring significant positive momentum to the group that has struggled in the weeks leading up to the release of its Vega architecture. The product family based on that tech underwhelmed and had concerns over availability, pricing, and timing. Wang has a strong history in the graphics field, with experience as far back as any high-level graphics executive in the business. While at ATI and AMD, Wang worked on architectures from 2002 through 2012, with several periods of graphics leadership under his belt. Competing against the giant that NVIDIA has become will be a challenge that requires significant technical knowledge and risk-taking and Wang has the acumen to get it done.

AMD CEO Lisa Su expressed excitement and trust in the new graphics executives. “Mike and David are industry leaders who bring proven track records of delivering profitable business growth and leadership product roadmaps,” she says. “We enter 2018 with incredible momentum for our graphics business based on the full set of GPU products we introduced last year for the consumer, professional, and machine learning markets. Under Mike and David’s leadership, I am confident we will continue to grow the footprint of Radeon across the gaming, immersive, and GPU compute markets.”

Samsung Begins Mass Production Of 18 Gbps 16-Gigabit GDDR6 Memory

Subject: Memory | January 18, 2018 - 12:34 AM |
Tagged: Samsung, graphics memory, graphics cards, gddr6, 19nm

Samsung is now mass producing new higher density GDDR6 memory built on its 10nm-class process technology that it claims offers twice the speed and density of its previous 20nm GDDR5. Samsung's new GDDR6 memory uses 16 Gb dies (2 GB) featuring pin speeds of 18 Gbps (gigabits-per-second) and is able to hit data transfer speeds of up to 72 GB/s per chip.

Samsung GDDR6_PhotoFs.png

According to Samsnug, its new GDDR6 uses a new circuit design which allows it to run on a mere 1.35 volts. Also good news for Samsung and for memory supply (and thus pricing and availability of products) is that the company is seeing a 30% gain in manufacturing productivity cranking out its 16Gb GDDR6 versus its 20nm GDDR5. 

Running at 18 Gbps, the new GDDR6 offers up quite a bit of bandwidth and will allow for graphics cards with much higher amounts of VRAM. Per package, Samsung's 16Gb GDDR6 offers 72 GB/s which is twice the density, pin speed, and bandwidth than that of its 8Gb GDDR5 running at 8Gbps and 1.5V with data transfers of 32 GB/s. (Note that SK Hynix has announced it plans to produce 9Gbps and 10Gbps dies which max out at 40 GB/s.) GDDR5X gets closer to this mark, and in theory is able to hit up to 16 Gbps per pin and 64 GB/s per die, but so far the G5X used in real world products has been much slower (the Titan XP runs at 11.4 Gbps for example). The Titan XP runs 12 8Gb (1GB) dies at 11.4 Gbps on a 384-bit memory bus for maximum memory bandwidth of 547 GB/s. Moving to GDDR6 would enable that same graphics card to have 24 GB of memory (with the same number of dies) with up to 864 GB/s of bandwidth which is approaching High Bandwidth Memory levels of performance (though it still falls short of newer HBM2 and in practice the graphics card would likely be more conservative on the memory speeds). Still, it's an impressive jump in memory performance that widens the gap between GDDR6 and GDDR5X. I am curious how the GPU memory market will shake out in 2018 and 2019 with GDDR5, GDDR5X, GDDR6, HBM, HBM2, and HBM3 all being readily available for use in graphics cards and where each memory type will land especially on the mid-range and high-end consumer cards (HBM2/3 still holds the performance crown and is ideal for the HPC market).

Samsung is aiming its new 18Gbps 16Gb memory at high performance graphics cards, game consoles, vehicles, and networking devices. Stay tuned for more information on GDDR6 as it develops!

Also read:

Source: Samsung

Samsung Mass Producing Second Generation "Aquabolt" HBM2: Better, Faster, and Stronger

Subject: Memory | January 12, 2018 - 05:46 PM |
Tagged: supercomputing, Samsung, HPC, HBM2, graphics cards, aquabolt

Samsung recently announced that it has begun mass production of its second generation HBM2 memory which it is calling “Aquabolt”. Samsung has refined the design of its 8GB HBM2 packages allowing them to achieve an impressive 2.4 Gbps per pin data transfer rates without needing more power than its first generation 1.2V HBM2.

SAMSUNG-HBM2_C.jpg

Reportedly Samsung is using new TSV (through-silicon-via) design techniques and adding additional thermal bumps between dies to improve clocks and thermal control. Each 8GB HBM2 “Aquabolt” package is comprised of eight 8Gb dies each of which is vertically interconnected using 5,000 TSVs which is a huge number especially considering how small and tightly packed these dies are. Further, Samsung has added a new protective layer at the bottom of the stack to reinforce the package’s physical strength. While the press release did not go into detail, it does mention that Samsung had to overcome challenges relating to “collateral clock skewing” as a result of the sheer number of TSVs.

On the performance front, Samsung claims that Aquabolt offers up a 50% increase in per package performance versus its first generation “Flarebolt” memory which ran at 1.6Gbps per pin and 1.2V. Interestingly, Aquabolt is also faster than Samsung’s 2.0Gbps per pin HBM2 product (which needed 1.35V) without needing additional power. Samsung also compares Aquabolt to GDDR5 stating that it offers 9.6-times the bandwidth with a single package of HBM2 at 307 GB/s and a GDDR5 chip at 32 GB/s. Thanks to the 2.4 Gbps per pin speed, Aquabolt offers 307 GB/s of bandwidth per package and with four packages products such as graphics cards can take advantage of 1.2 TB/s of bandwidth.

This second generation HBM2 memory is a decent step up in performance (with HBM hitting 128GB/s and first generation HBM2 hitting 256 GB/s per package and 512 GB/s and 1 TB/s with four packages respectively), but the interesting bit is that it is faster without needing more power. The increased bandwidth and data transfer speeds will be a boon to the HPC and supercomputing market and useful for working with massive databases, simulations, neural networks and AI training, and other “big data” tasks.

Aquabolt looks particularly promising for the mobile market though with future products succeeding the current mobile Vega GPU in Kaby Lake-G processors, Ryzen Mobile APUs, and eventually discrete Vega mobile graphics cards getting a nice performance boost (it’s likely too late for AMD to go with this new HBM2 on these specific products, but future refreshes or generations may be able to take advantage of it). I’m sure it will also see usage in the SoCs uses in Intel’s and NVIDIA’s driverless car projects as well.

Source: Samsung