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

There was some developer from IsraEl that have some really good A.I. solution for Games - RPG in mind (im sure Google will find it) as i remember it was Hardware (like AGEIA Physix)
Living world of A.I. NPC's changing and evolving all the time.

Now we have affordable 8/16 and this is the Power to drive some really good A.I in games.

I'm really into Next Gen DX12 SM 6.0 Games that will pop up when ZEN+VEGA Consoles are ready.
AMD has a point to prioritize this over Vega Drivers (Look at those +100mln ppl)
Then is really easy to pick this for Us PC gamers, we are all A Family of Gamers.

Do you know that when True DX12 game hit us, it will be Big like Tesselation back in the days.
Just look at DX12/VLK real capabilities

 

According to Expreview AMD could reveal 7nm Ryzen Raven Ridge APUs later this year. The story is allegedly based on Citigroup analyst report, which describes TSMC production capacity for different nodes.
He claims that AMD will launch 7nnm APU at the end of this year.
Interestingly enough, yesterday at Hotchips, AMD also confirmed that more power efficient APUs will be unveiled in 2018.
They specifically referred to them as Raven Ridge, so we might be looking at a die shrink (otherwise it would probably carry a different codename).

AMD clarified that 7nm Vega and 7nm EPYC processors will have the priority for 7nm production.
Consumer 7nm products were not expected this year.
The Ryzen 3000G series would, therefore, follow the path of Ryzen 2000G series, which were released before Ryzen desktop 2000 series.

1:
Anandtech’s coverage of HotChips reveals the slide mentioned in this post.
This slide only mentions Raven Ridge 2018. The 2000G series were released in 2018, so the power efficiency thing may actually refer to existing products.

2.
AMD has their own dedicated website to 25X20 energy efficiency goal.
The footnotes reveal that Raven Ridge 2018 is actually Ryzen 7 2800H, the high-end mobile CPU, which was rumored for a few months.

The efficiency goal is likely referring to 12nm refresh, rather than the 7nm APU.

Spoiler

-


-

 

Nice findings, as always.

Just to be sure, though:

Can I flash my V56Pulse-Samsung mem with a 64Nitro+ VBIOS, more or less safely/successfully (my responsibility, of course)?

Last time I checked around, I couldn't find a definite answer.

Do I need some "special" s/w, or just good-old AtiFlash/AtiWinFlash?

TY!

 

The latest version of AtiFlash (NOT AtiWinFlash.) supports Vega GPU's but I would be careful about the Pulse GPU flash.

The tri-fan (Non-limited edition so the 2x 8-pin model.) Sapphire Nitro+ GPU seems to be the best but results vary so first you should use GPU-Z and dump the existing bios from the GPU and save that, then flip the GPU bios switch and dump the other bios this way (Can be done without a reboot.) and then flip back to the default again so you now have the default bios as active and dumps of both bios files.

After that you can try to use AtiFlash and the latest version from TechPowerUp and their bios database and tool download site and attempt to flash from the Sapphire Vega 64 bios but it might error out depending on checks and such that the card has at which point there's a option to force it and that could work but if the GPU exhibits any issues at all either immediately following a reboot or in-game then flip the bios switch, reboot, flip the bios switch back to the flashed position and then apply the backup of the original bios.


The protection on the Vega GPU's requiring a signed bios files means you can't just edit the memory voltage and apply a modified bios file but the bios also contains voltage and clock speed info and then fan curve and settings for how the fan ramps up and also the thermal and power information all of which differs between Vega 64 and Vega 56 even for the air models and then between various third party GPU models such as the fan settings.

In addition to all that which makes flashing like this a risk the Pulse is also a Nano style PCB having a 7+1 type voltage setup instead of 12+1 and maybe other component differences on the PCB so there are risks due to these sorts of changes too.


Personally I'd see what the memory overclock can go up to without flashing the bios, there's more stories about flashing having failed and having to re-flash the GPU to the original bios and then there's a few that state it's been working but they might have replied immediately following the reboot and not done any longer term testing or testing under load where the card is a bit more stressed and pushing more power and such.

900 - 950 Mhz is still really good though the GPU still scales well up to at least 1 Ghz before it trails of slightly and might require better cooling or more voltage for full stability for a small performance gain that might not be worth the hassle though from 800 Mhz stock to 900 Mhz still sees a nice gain especially if you can also keep the core clock around 1550 - 1600 Mhz which the Pulse can do even at stock settings if cooling allows for it though the default 1.2v might see the GPU run fairly hot after a while so it will go back and forth between max clock speeds and the state just below it more instead of sticking to the max clock state.


Though most results with Vega even the water model has the GPU do this anyway even if you seriously undervolt and underclock the GPU so it's just how it behaves with current drivers, boosting the 3D clock speeds so it keeps to 1590 - 1650 Mhz is one way but pushes more demand on the card and might need testing for thermal and stability particularly since the GPU can spike a bit seeing a increase in voltage or clock speeds pushing above the p7 state specified values and that can cause a GPU crash though it's harmless but might need a reboot and then re-applying the Wattman settings or other utility since the drivers reset these to default if it detects that the GPU has reset for any reason such as from a driver crash.
(Or the current issue with fast startup and sleep mode which AMD has yet to solve I believe.)

 

There is chance that you'll not be able to flash if you have newer build of W10. So, you may need boot media/older build. I have live USB with older W10 build just for flashing.

 

I know...already had this unfortunate experience while trying to reflash the original BIOS on my 290. I think that Rufus will be quite useful.

@JonasBeckman: I know all the steps, been flashing since my 5770s.

Still, thank you for your response and for clarifying every little detail.

I don't know if I'll give it a try for the time being, but it's good to know that options are always available.

 

Also gents, if taking about Flash -> go to BIOS Mods Threads.

 

AMD 7nm "Vega" by December, Not a Die-shrink of "Vega 10"

AMD is reportedly prioritizing its first 7 nanometer silicon fabrication allocations to two chips - "Rome," and "Vega 20." Rome, as you'll recall, is the first CPU die based on the company's "Zen 2" architecture,
which will build the company's 2nd generation EPYC enterprise processors. "Vega 20," on the other hand, could be the world's first 7 nm GPU.

"Vega 20" is not a mere die-shrink of the "Vega 10" GPU die to the 7 nm node. For starters, it is flanked by four HBM2 memory stacks, confirming it will feature a wider memory interface, and support for up to 32 GB of memory.
AMD at its Computex event confirmed that "Vega 20" will build Radeon Instinct and Radeon Pro graphics cards, and that it has no plans to bring it to the client-segment.
That distinction will be reserved for "Navi," which could only debut in 2019, if not later.

 

"All road leads to Rome."
It was known for a while that 7nm Vega is not just die shrink. There was a lot of unfinished work put into Vega which has to be either completed or undone to either reap benefit or save on transistor count.

Now, 4 HBM2 stacks... Did they break 4096SP limitation? Good.

 

Hmm suppose this works here, it is technically a AMD GPU news as it uses their functionality.
(EDIT: Nope it works on Intel and NVIDIA too it seems, neat!)

https://www.reddit.com/r/Amd/comments/99rmzx/steam_audio_beta_15_radeon_rays_support/

https://steamcommunity.com/games/596420/announcements/detail/1681419156989664451

EDIT: And in other news the open source drivers have more Vega 12 entries now.
https://www.reddit.com/r/Amd/comments/99wcxu/amd_posts_opensource_vulkan_driver_code_for_vega/

https://www.phoronix.com/scan.php?page=news_item&px=AMD-Vulkan-Vega-12-Posted

 

RX 580 MSI ARMOR OC 8GB $229.99 on Amazon ! A great price indeed

 

Here some interesting read

-> https://community.amd.com/community...-our-25x20-energy-efficiency-goal-report-card

-> https://www.amd.com/en/technologies/25x20

 

GPUOpen Libraries & SDKs

Radeon Rays is ray intersection acceleration library for hardware and software multiplatforms using CPU and GPU

-> https://github.com/GPUOpen-LibrariesAndSDKs

 

AMD Radeon Pro V340 Dual Vega 32GB VDI Solution Launched

The new Radeon Pro V340 is a dual-GPU accelerator that uses the Vega architecture with 56 Compute Units per GPU, and a huge 32GB of HBM2 with Error Correcting Code (ECC). This is not a gaming card and something destined for workstations and servers.

Radeon Pro GM Ogi Brkic explains: "As the flagship of our new Radeon Pro V-series product line, the Radeon Pro V340 graphics card employs advanced security features and helps to cost effectively deliver and accelerate modern visualization workloads from the datacenter". AMD is passively cooling the card which is thge most interesting thing here, and I'm actually a little blown away by this to be honest.
The consumer Radeon RX Vega 56 and RX Vega 64 graphics cards have fans that even when spinning at 100% can't keep the card cool, as it drives past 80C even in a cool-operating environment.
A passively cooled dual-GPU card is quite the engineering feat.
AMD has most likely clocked the Vega GPUs down heavily on the card,
and I'm sure the ECC HBM2 has been down clocked as well.
The company hasn't revealed any GPU or HBM2 clocks, nor is the TDP available.
We are probably looking at 400-500W at the very least, considering Vega uses 300W without a problem in single gaming form with RX Vega 56/64.

Sheldon D'Paiva, director of Product Marketing at VMware chimed in with the announcement of the Radeon Pro V340, where he said:
"The AMD Radeon Pro V340 graphics card will enable our customers to securely leverage desktop and application virtualization for the most graphically demanding applications. With Radeon Pro for VMware, admins can easily set up a VDI environment, rapidly deploy virtual GPUs to existing virtual machines and enable hundreds of professionals with just a few mouse clicks".

Calvin Hsu, the VP of Product Marketing at Citrix said:
"With increased density, faster frame buffer and enhanced security, the AMD Radeon Pro V340 graphics card delivers a powerful new choice for our customers to power their Citrix Workspace, even for the most demanding applications".

Read more: https://www.tweaktown.com/news/6293...pu-card-32gb-hbm2-passively-cooled/index.html


https://www.servethehome.com/amd-radeon-pro-v340-dual-vega-32gb-vdi-solution-launched/

 

I hope that GlobalFoundries wasn't doing the 7nm fab...

https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development

 

AMD accelerates 7nm process adoption

In an interview with IT industry journal CRN, AMD's Mark Papermaster has spoken of his firm's advances in 7nm process adoption.
The AMD Chief Technical Officer (CTO) claimed that, though 7nm has been a big challenge it decided to go all-in, and thanks to its partners is now able to bring ahead the launch date for the technology.



"We knew 7nm would be a big challenge, so we made the bet, we shifted our resources onto the new node," Papermaster said in an exclusive interview with CRN.
"We didn't just dip our toe in the water. We went all in."

AMD had originally planned to release the first 7nm Vega GPU in 2019 but, thanks to "immense focus" has been able to move the date forward to later this year.
Thus we will see the next-generation Radeon Vega Instinct GPU, demonstrated at Computex with 32GB HBM2, launched before the year is out.



AMD will also transition its CPUs to 7nm in the not-too-distant future. Papermaster said that the first 7nm EPYC server CPU will be released next year, followed by 7nm Ryzen desktop PC processors in the wake of the server products.
Significantly, the AMD CTO claimed that the advance to 7nm processors will provide "double the performance per watt of energy expended in a previous node."



The progress from AMD sounds quite different in tone to the news we have heard from Intel, regarding process advancements lately.
AMD with its design team, foundry partners, and electronic design automation partners has managed perform "much stronger than we originally anticipated," Papermaster told CRN, allowing AMD to bring plans forward.
Papermaster explained that the original AMD roadmap was already strong but that the delays from a competitor (Intel) simply strengthened AMD's position.

Intel has released some 10nm parts for a limited range of laptops and NUCs but has signalled that volume production is still some way away. In its most recent investor conference call, a month ago, Intel indicated that it was having difficulties transitioning from its 14nm process.
Thus it will keep tweaking the 14nm process and chip architectures to maintain 'system leadership' until 'Holiday 2019'.

 

AMD hands TSMC its entire 7nm portfolio: Vega 20, Zen 2 and Navi right around the corner

Why it matters?
This is the last piece of the puzzle in AMD's long journey to put its manufacturing strategy in order. Unfortunately this meant dropping Globalfoundries, and taking on TSMC as the go-to partner for its current portfolio.
TSMC's manufacturing lead over Intel is substantial and AMD aims to capitalize on that.
Mark Papermaster, CTO of AMD, has announced that the company has finally leveled out its GPU and CPU development and unified everything under a single manufacturing node, at 7nm, handing TSMC the keys to the kingdom.

In a blog post, the exec explains how the company's long term vision has finally paid off and AMD is now in a position to put its competition between a rock and a hard place with advanced 7nm designs in both GPU and CPU segments, thanks exclusively to the help of TSMC.
This announcement sets the stage for the departure of GlobalFoundries, AMD's other, oldest, foundry partner.

The announcement had already been hinted at during several corporate communications, events and calls, however, it is the first time AMD comes forward and effectively drives this narrative.
Papermaster said "Our work with TSMC on their 7nm node has gone very well and we have seen excellent results from early silicon."

Papermaster goes on to report that multiple 7nm products have taped out at TSMC. This includes the workstation-class GPU, Vega 20, and the next generation CPU architecture, Zen 2, all manufactured by TSMC.
Going forward, AMD will focus its entire product portfolio on TSMC's 7nm process, and that includes this year's Vega 20, followed by an early 2019 launch of the Zen 2-based server CPU, Epyc, and later that year, the Navi GPU architecture, the long-awaited 7nm follow-up to Polaris which AMD is now beginning to talk up.

For more time than most AMD fans would care to remember, the company had relied on GlobalFoundries to cater to its semiconductor manufacturing needs. The results, however, have been a mixed bag.
GlobalFoundries has tripped itself up more than once while playing catch-up with Intel, having twice nixxed its developing processes.
It trashed its 14nm development in favor of licensing it off Samsung, and now it has announced it has ended its 7nm node. In the meantime, TSMC, the backup fab for AMD, not only delivered where GlobalFoundries failed, but managed to overtake Intel in manufacturing technology, and this has paid off.

-> https://community.amd.com/community...mance-leadership-with-focused-7nm-development

 

I wish they were releasing a 7nm gaming part this year or early next year. Just a die shrunk Vega would probably do pretty well against what the green team is offering.

 

With the overwhelming strength of Turing, I can only hope for a stronger RX680, in top of upper mid-range vegas. They aren't for any reason high end anymore. We need a new red leader. So vegas could stop being so expensive, or give us a new kind of juice to drink.

All RTX series are gonna be stronger than any AMD gpu. At least we need an RX680 to beat 2070. As it seems vega gpus are gonna be between 2060 and 2070.