New Upcoming ATI/AMD GPU's Thread: Leaks, Hopes & Aftermarket GPU's

Intel’s 10nm Cannonlake CPU Based Crimson Canyon NUCs Will Be Shipping With AMD Radeon RX 500 Discrete GPUs

Intel’s 10nm process has been a disaster of epic proportions with countless delays and still no product in sight. Promised to appear in 2017, there’s still no consumer grade product that was meant to launch under the Cannonlake family but that may change soon as Liliputing (via Win Future) has got details regarding a specific Intel NUC featuring the Cannonlake 10nm processor.

Intel’s 10nm Cannonlake CPU Based NUC Launching Soon – Allegedly Features Dual-Core CPU With AMD Radeon RX 500 Series Discrete Graphics
The Intel Cannonlake processors once appeared to be a generational leap with the latest 10nm process node but difficulties with the 10nm process led to huge delays for a product that was meant to launch years ago.
During this time, rivals such as TSMC and Samsung have managed to catch up with Intel’s own Fabs and the blue team is no longer looking like the leader of this chip department anymore.

However, Intel will be launching 10nm Cannonlake processors soon and while there are no details on when these CPUs will actually launch, some data from various benchmark databases confirm that OEMs are already testing 10nm samples within their upcoming products.
We previously saw Lenovo based notebooks that appeared with a dual-core Core i3-8121U chip rocking a clock speed of 2.20 GHz and TDP of 15W. Performance in can be seen -> Geekbench 4 benchmark

The 3DMark performance benchmark was also posted by TUM APISAK where the Core i3-8121U based NUC along with the discrete Radeon RX 550 graphics chip posted a score of 4434 points (4553 points for graphics and 4209 points for physics).





There will be two variants of the Intel Crimson Canyon NUCs starting at 450 Euros and will launch around May with a possibility of the launch being moved to June for Computex 2018. The models will include:

• NUC8I3CYSM2 (2.20 GHz / 2C/4T / 2 GB Radeon 500 Series GPU)
• NUC8I3CYSM3 (2.40 GHz / 2C/4T / 2 GB Radeon 500 Series GPU)

This will be the latest addition to the NUC family feature graphics from AMD Radeon. The Hades Canyon NUC already rocks custom-built RX Vega graphics with HBM2 memory.
The Crimson NUCs will support 2 GB of GDDR5 memory and that will definitely be a good update in graphics performance over Intel’s integrated graphics chipsets.
We are working on our own Hades Canyon review at the moment and can’t wait to show you what kind of performance it offers to the average gamer in a nice sleek package.

THX to WCCFtech

 

Jump into Square Enix Collective with AMD Radeon™ RX graphics.
Access the Collective with your selection of two FREE games when you purchase an AMD Radeon™ RX 560 or one FREE game when you purchase an AMD Radeon™ RX 550 graphics card!

-> https://www.amd.com/en/gaming/AMD-SECollective?sf187421562=1

 

Download: Radeon tech demos from 2001 until today

As expected those tech demos are supposed to show the advantages of the according Radeon graphics card.
PC Games Hardware has gathered the demos and offers them for download. In the picture gallery below you can get an impression of what the individual titles look like.
Of course there are more tech demos, but they usually belong to third parties.

-> http://www.pcgameshardware.de/Grafi...DirectX-101-AMDs-Radeon-graphics-cards-683297

From the first Radeon girl to the modern Ruby and the Froblins: They are all there.
We especially recommend the HDR Rendering Demo Debevec RNL of the Radeon 9700, the Toyshop Demo of the X1000 series and the March of the Froblins Demo of the HD 4800 series, because all of them set new technological standards in their time.

Please bear in mind that for most of the tech demos a Radeon graphics card is required and sometimes even a card of the individual series.
Furthermore some demos are available as video files only - because of licensing reasons - so we didn't add them to our list.

If you are interested in the tech demos of graphics cards, you might want to take a look at the following articles:
• DirectX 10.1: AMD tech demo March of the Froblins with A.I. calculation on the GPU
• Ruby 2.0: Screenshots and video of the new Radeon tech demo

 

AMD add-in-board manufacturers including Gigabyte, MSI and TUL are expecting their shipments to drop by as much as 40% in April according to a report by DigiTimes.
All of these Taiwan based manufacturers have established their forecasts based on the assumption that cryptocurrency mining demand is going to continue to fall.
Keeping in mind the fact that there is usually a bureaucratic delay associated with such assumptions, and the fact that the crypto market just started rebounding, it remains to be seen how accurate these predictions are going to be.

This, of course, has further reaching implications than just a fall in demand. Due to the high ROI offered by mining, graphics card manufacturers were able to sustain insanely high margins of up to 50% with prices following a supranatural trend.

With the fall from grace od graphics card as the darling of the cryptocurrency mining world, while the demand will fall 40%, the demographic will change even more materially.
From rich cryptocurrency miners willing to pay 3-4x the MSRP, we will go back to gamers being the primary audience and with it a shrink in the Total Available Market for AIBs.

To put it simply the high end of the market segment has started declining, and the channel distributors will be forced to go back into sales promotion mode for gamers.
This means that we will converge towards the old margins of roughly 20-25% for AIBs IF bitmain does indeed release its ethereum miner as promised and/or the cryptocurrency market continues to stay in its present ditch.
Gigabyte posted record shipments of 4.5 million graphic cards in 2017, up one million on a year, with profits from the shipments doubling to over $67.37 million, it’s a similar story for MSI and TUL.

-- As i said earlier, im waiting for summer to go shoping (But im more into new Vega batch tho)

500-600 USD for V64 is OK IMO.
I mean Sapphire/XFX or PowerColor/AsRock

 

lol, and that's true
Gamers

Next Gen PS5/XboX X2 will have: Vega56 + Zen 8/16 with 8GB HBM_2

 

AMD Has 7nm GPUs Running in Labs, Sampling Later This Year !

AMD’s President and CEO Dr. Lisa Su made an exciting revelation yesterday during the company’s 1st quarter 2018 earnings conference call indicating that 7nm GPUs are much closer than we might have thought.


Dr. Su announced that AMD has working 7nm GPUs running in the Radeon Technologies Group labs right now. <- This one is for Kajit
Specifically, 7nm Vega which is expected to begin sampling later this year.


AMD President & CEO Dr. Lisa Su – Q1 2018 Earnings Report

“I’m also happy to report that our next-generation 7nm Radeon Instinct product, optimized for machine learning workloads, is running in our labs and we remain on track to provide samples to customers later this year.”



7nm Vega in the House
Earlier this year at CES AMD announced that it’s working on a 7nm Vega based graphics processor optimized for Artificial Intelligence (AI) and Machine Learning workloads.
Dr. Su’s remarks yesterday evening confirm that the company has gotten silicon back from the fab and is currently testing the chips.

later in a question and answer segment of the earnings call, Su also confirmed that 7nm Vega is manufactured at TSMC, rather than Globalfoundries.
A decision that was likely influenced by the maturity of TSMC’s 7nm manufacturing process.
AMD Q1 2018 Earnings Call

Timothy Arcuri – UBS Securities
“I think you said that the 7nm product is in the lab and it’s going to launch later this year. That’s the product at TSMC, correct?
And I’m just wondering on that front do you feel comfortable that you can get capacity from that vendor?”

AMD President & CEO Dr. Lisa Su
“Our foundry strategy is to use both TSMC and Globalfoundries on the first 7nm product, we are using TSMC for that product and we have a very strong relationship with them.
And so, we do see good momentum on it from what we see and I’m not concerned about capacity.”

AMD’s Infamous 7nm Vega 20
Earlier this month we spotted a peculiar piece of code in a Linux patch that makes reference to Vega 20 by name.
The patch introduced support for more than 50 new Vega specific hardware-level features and we stated that this indicates that AMD could be approaching final post-silicon testing and validation of Vega 20. Thanks to Su’s announcement yesterday, it turns out that’s exactly right.

We brought you news of Vega 20 in as early as January of 2017. So we have known about AMD’s plans to introduce a high performance Vega chip based on the 7nm process for quite a while.
Confirmation of such a project, however, did not come until exactly a year later at this year’s CES back in January.

Although the company still does not refer to its new 7nm Vega chip as Vega 20, previously leaked internal AMD slides clearly describe a 7nm based “Vega 20” GPU that’s specifically designed for AI and intended for market entry in 2018.

Vega 20 differs from Vega 10 in several key aspects. In that it’s designed to deliver 8 times the double precision compute performance, double the memory interface width and support for up to 4 HBM2 stacks for a total of 32GB of HBM2 vRAM and 1TB/s of memory bandwidth.

Vega 20 will be AMD’s first chip designed specifically for AI and machine learning workloads.

 

Guessing that's going to be a workstation GPU but who knows, maybe we'll see consumer variants too down the line unless Navi is going to be the next GPU for this.

16 to 32 GB HBM2 is not going to be cheap (4x 4GB chips and 4x 8 GB for stacks I assume.) and availability hasn't gotten better either far as I know and the focus on AI, machine learning and double precision seem aimed more towards workstation too.

EDIT: Perhaps it'll be similar to the Titan V or more like the Vega Frontier that could do a bit of both, though it might not be all that good for a desktop system and gaming at least for what it costs and what it will offer in terms of performance compared to current cards and competition from NVIDIA.

 

7mm Vega for gaming is possible (October-december '18) but is unlikely IMO
More like Refresh Vega 12nm in Summer time.
We shall see, i hope my Hong-Kong source is right
if so we will have 2x Fury-X Performance with 230tW
up to 499-549USD (similar to today prices)

 

Would be nice, die shrink could improve the power usage though it is also affected by AMD over-volting both memory and GPU core to achieve the target clock speeds although this does result in many GPU's being able to under volt heavily and thus power usage is somewhat in check again.

Double the performance of the AMD Fury X is going to be tough though, clock speed alone isn't enough for GCN it seems with Vega pulling 30 to 40% ahead (Under ideal conditions.) but retaining some of the hardware bottlenecks though resolving a few others (Tessellation for example similar to Polaris.) good overclock on memory and core adds some 10% or so if stability and the dynamic clock or boost frequency can be maintained.

Anything is possible though but as a refinement of Volta and not a new architecture I'm guessing more along the lines of perhaps another 20% performance gain though more is certainly possible depending on how extensively they've tweaked the GPU from Vega 10 and what other improvements could be had.
(Better HBM2 chips allowing for 1 Ghz for example as the stock speed though clock speeds alone won't do enough though it will pull ahead of Vega64 a bit at least though it could also be clocked lower for GPU core.)

Will be interesting to see what AMD announces, for now it's mostly speculation and there's a lot of factors in play.
(Costs and pricing is going to be a big one again I'm guessing.)

 

Gaming Monitors Explained: Panel Types, Refresh Rates, Resolution, And More

Here Lecture

-> https://www.gamespot.com/gallery/ga...s-refresh-rate/2900-1555/?ftag=GSS-05-10aab8e

 

AMD at Work:
DDR5?
HBM2 for CPUs?

Spoiler: NAVI in PS5/XboX X2?

As for when the PS5 will arrive, late 2019 is probably the earliest you can expect the machine, with 2020 being more likely. Leadbetter bases this prediction on the tech currently available – Sony simply wouldn’t be able to offer a true generational leap in power in compact console form at an affordable price with what’s out there right now. The PS5 will most likely use the smaller, denser 7nm chip manufacturing process and super-speedy HBM2 or GDDR6 memory, neither of which will be ready for big-time mainstream production until 2019.

But hey, it sounds like the PS5 will be worth the wait. Leadbetter speculates the PS5 will be powered by an APU composed of an AMD 8-core Ryzen CPU and a Navi-based GPU (which is similar to what we heard before) and should deliver true high-end desktop-level performance. Of course, take all this with a grain of salt, but again, Eurogamer is a very reliable source when it comes to gaming hardware rumors – do check out their full article for a lot more juicy tech talk.

New Vega 2.0 in 2d Half of 2018?
Multi Stacked GPU Chips on Infinity Fabric this YAS? -> OK, i can wait

But we will Hype Again -> so Make the Hype Great Again

 

https://videocardz.com/76076/early-amd-vega-20-3dmark11-benchmark-result-emerges

https://cdn.videocardz.com/1/2018/04/AMD-Vega-20-vs-Vega-Frontier-768x952.jpg

So 32 GB HBM2 and running on a newer driver and as the article mentions probably a engineer sample and someone forgot to uncheck the results being uploaded thus this leaked benchmark.

Doesn't mean much for now though, ES meaning it could differ from the final in various ways (Clock speeds for example and those might not be reporting right anyway in this chart.) plus it's a early driver too so the final product could be faster but we'll see if they make a desktop variant of Vega 20 at all or if it's for workstations only.

EDIT: And yeah VcardZ but since it's a graph from 3DMark 11 that should be valid.
(Or I'm mixing it up with some other website like FUD being known for posting unconfirmed rumors and what not.)


EDIT: Ah the hotlinking of the image didn't work for very long, well the URL is still there as is the article itself.

 

"Now, first things first. I should give a shout out to our friends over at VCZ who have also managed to track down this result.
The database entry is for 3DMark 11 “Performance” and shows a yet unidentified “Generic” AMD GPU with 32GB of HBM2 vRAM clocked at a whopping 1.25GHz, the highest clock we’ve seen on HBM to date.

Now, the most interesting part here is the GPU clock speed. If the reported figure is to be believed this early engineering sample of Vega 20 is running at exactly 1GHz. It’s possible that this is because 3DMark is misreporting the clock speed.
However, if we go back to really early Vega Frontier Edition database entries we find that when 3DMark fails to read the clock speed it simply reports it as 0, rather than 1GHz. And misreadings of the UEFI data have also led to misreadings of the HBM2 vRAM by 3DMark.
We don’t see any of that with this particular Vega 20 entry.

Additionally, early RX 480 engineering samples that we spotted back in the day were also running at 1GHz, with later production units hitting around ~1.2GHz.
1GHz is just a nice, round and easy frequency figure to run these very early engineering samples on. That’s why they go for it with these super early engineering samples.
Remember, we’re potentially looking at the first operational 7nm Vega 20 GPU silicon back from the TSMC fabs.

Vega 20 70% Faster Per Clock vs RX Vega 64 Liquid, if the Numbers Aren’t Lying

With that being said, if Vega 20 manages to deliver similar performance to the Vega 64 Liquid at 1GHz we’re looking at no less than 70% faster performance per clock, which seems absolutely bonkers.
The alternative hypothesis is that Vega 20 is running at a modest but unknown boost clock speed that 3DMark is just not picking on, although we’ve never seen boost functionality operational on any of AMD’s engineering samples in the past.

The truth most likely lies somewhere in the middle. We’re no doubt looking at a very early silicon here with nowhere near final clock speeds.
Vega 20 isn’t expected to begin sampling until the end of the year and won’t hit shelves until next year."

THX to WCCFtech

 

New patches for AMDGPU llvm backend to support new instructions for vega 20 and vega 12
https://www.phoronix.com/scan.php?page=news_item&px=Vega-12-20-AMDGPU-LLVM

 

Free on Origin ! Just add to you games:

BF1 - They Shall Not Pass DLC
Battlefield 4 - Dragon's Teeth DLC

 

Awesome, thanks for the heads up.

 

I have a feeling the core clocks have been read wrong. I can't see it being 70% faster than Vega 64 with a die shrink and some ipc optimizations. It would be sensational if true, but in reality take this with a huge grain of salt.

 

I assume that this was already with some degree of "boost clock" so the reading is false. However we all know that the Vega is bandwidth-starved and I assume that with the full 1024 bandwidth it could reach maybe 20-30% improvement over Vega "1". Plus maybe some internal small improvements and we are talking about "30-35%" improvement in the performance plus improvement in the power usage. This is my humble opinion to "what to expect next". Would be really sexy if they could connecty two or three smaller chips with infinity fabric already, but I've heard that's not the case as vega was not designed for it... Navi is the next "small" thing that will be bigger in the end ;-)

 

30-40% it's just Great ! and also enough for me
+ it's strongly possible...

 

That's about 1080Ti territory. Considering just how powerful that card is, I'd be more than happy with that performance.