• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD 600 Series GPU Project “Zen” Detailed – Radeon on Steroids to Amp Clock Speeds & Efficiency

Caporegime
Joined
8 Jul 2003
Posts
30,063
Location
In a house
Radeon Project “Zen” & RTG’s Zen Team Supercharging AMD’s GPU Roadmap Starting With RX 600 Series

The new team has been put in place to inject the company’s upcoming GPU products with a dose of engineering steroids so to speak and leapfrog the company’s original GPU roadmap. It’s tasked with bringing expertise from the original Zen CPU project to the GPU side as well as to innovate in two key areas in an effort to make upcoming GPU products more competitive than originally planned.

The two key areas that the team is focused on is significantly improving the clock speeds of AMD’s GPU designs and pushing them to be more power efficient. We’re told that this project was put in motion under the direct orders of CEO Su, who expects results in as early as this year. Although the team is also involved with engineering efforts around the Navi architecture which is expected in 2019 as well as the company’s brand new post-GCN architecture that’s expected in 2020.

We’ve been hearing a lot of chatter in private channels about how the company’s post-GCN, all new architecture is going to be Radeon’s Zen so to speak. But if what we’ve been hearing about Radeon Project Zen is also accurate, we might see some of the new team’s results with the Radeon 600 series, much earlier than 2020.

https://wccftech.com/amds-secret-ra...locks-efficiency-in-2018-navi-in-2019-beyond/
 
A good start would probably be ditching GF who don't seem to be able to produce nodes that have good synergy with GPUs. Unfortunately the wafer agreement probably makes that very costly.
 
A good start would probably be ditching GF who don't seem to be able to produce nodes that have good synergy with GPUs. Unfortunately the wafer agreement probably makes that very costly.

They don't need to, they already will be using TSMC and GloFo next year for 7nm. They can pick and choose whichever products to use for each one.
 
They don't need to, they already will be using TSMC and GloFo next year for 7nm. They can pick and choose whichever products to use for each one.

AMD have a wafer supply agreement with GF until 2020 - they can't just pick and choose there will be limitations in what products are covered by it and profitability versus penalties.
 
I don't think many have much faith in AMD on the gpu-side now, they're on another 5 year journey imo. As long as they can produce decent low/medium/upper medium products then they're still be fine for a while. Just don't be expecting them to set the gpu world alight!
 
AMD have a wafer supply agreement with GF until 2020 - they can't just pick and choose there will be limitations in what products are covered by it and profitability versus penalties.

No, they have already confirmed that they will be using both TSMC and GloFo for 7nm.

"LS: So in 7nm, we will use both TSMC and GlobalFoundries. We are working closely with both foundry partners, and will have different product lines for each. I am very confident that the process technology will be stable and capable for what we’re trying to do."

https://www.anandtech.com/show/1231...n-exclusive-interview-with-dr-lisa-su-amd-ceo
 
No, they have already confirmed that they will be using both TSMC and GloFo for 7nm.

"LS: So in 7nm, we will use both TSMC and GlobalFoundries. We are working closely with both foundry partners, and will have different product lines for each. I am very confident that the process technology will be stable and capable for what we’re trying to do."

https://www.anandtech.com/show/1231...n-exclusive-interview-with-dr-lisa-su-amd-ceo

They still have a WSA with GF - they can use other foundries but there are restrictions, penalties and minimum supply agreements which means depending on desire for profitability, etc. they can't just pick and choose.
 
They still have a WSA with GF - they can use other foundries but there are restrictions, penalties and minimum supply agreements which means depending on desire for profitability, etc. they can't just pick and choose.
It would be interesting to see the details of that supply agreement, it must be instrumental in allowing AMD to produce chips as big the ones found on the 570/580 which are competition against the much smaller and efficient 1060. It probably also gives AMD first stab a new process as well as having a secure and reliable supply source of wafers rather verses TSMC were everyone is behind Apple and Qualcomm. In likelihood the 7nm arrangement will the be same as 14nm, GF produces AMD's CPU's and Radeon graphics cards and TSMC does all the semi custom staff for Intel, Microsoft and Sony etc.
 
Ryzen is amazing, Fantastic chip that can run with the big boys but I've been burned too many times now with AMD GPU purchases, Fury X was garbage, Vega 64 was garbage, I'll believe the hype when they actually produce something worth buying.
 
Ryzen is amazing, Fantastic chip that can run with the big boys but I've been burned too many times now with AMD GPU purchases, Fury X was garbage, Vega 64 was garbage, I'll believe the hype when they actually produce something worth buying.

Obviously they do.

Statements such as all products are worthless (because personal demands are not met) are what makes GPU and CPU sections so toxic.

It's like a reverse hype, everything current and in the future is trash unless it is the best (or at least meets personal demands).

The hype you think you're above, you encourage in the opposite direction. And we look around and see others just the same.
 
It is going to be interesting to see the AMD 2018 Q1 financials. I suspect that with the Cyrpto situation that volumes YOY for RTG will be up anyway - that will have been protecting the whole division to some degree. That boardroom conversation where they show the actual financial graph and think that they have done enough to be safe.

I think that what we are seeing is a mix of reaction from the 12nm process, ASIC ethereum development, Koduri leaving and Lisa having a rock solid reputation following Ryzen within the business.

Intel's security issues, Nvidia's GPP decisions and the success of Ryzen's modular design has got to be a really strong opportunity to launch new technology ASAP.

I predict that there are two things happening here. First that GPU's are going to the 12nm+ process and they have brought Engineers from the transition project from Ryzen making that change to GPU. That gives them an opportunity for a 10% clock bump. I hope they will offer a re-engineered Vega with a lower core count card to replace the Polaris 580. Vega does appear to be doing well in the 2400g and obviously plays well with DDR4 so I hope we see a Vega 52 with GDDR that hopefully they can produce in enough quantity to actually hold stock of and has performance 10-15% higher than a 580.

Second - Navi has already been rumored to be a modular design. That would directly leverage from the Ryzen marketing and success and get them away from large graphics cores. I suspect that the team that led the development of IF for Ryzen has been tasked with making sure that whatever they are using to do that in Graphics actually delivers.
 
Last edited:
I've been in these types of threads/discussions before. We all get hyped up thinking this will finally be the time that AMD get's back in the game and then we get disappointed when the GPUs launch. As such, my expectations are super low right now for GPUs.

Also pretty funny that they decided to call these GPU's 'Zen'. Are they hoping that these GPUs will do for AMD graphics what Ryzen did for the CPU side? I wouldn't count eggs before they hatch...

Edit: Also, having recently watched DigitalFoundry's PS5 speculation video, it seems that AMD have a limitation on CUs, maximum of 64. The only way they can squeeze more performance out of GCN appears to be only via clockspeed (and drivers). So they need a need to come up with something post-GCN to really push forward.
 
2010: Our CPUs are giving Intel a run for their money, but our GPUs suck, quick divert resources to the GPU division!
2012: Our GPUs are giving Nvidia a run for their money, but our CPUs suck, quick divert resources to the CPU division!
2018: Our CPUs are giving Intel a run for their money, but our GPUs suck, quick divert resources to the GPU division!

I see no way this plan can end badly...
 
Obviously they do.

Statements such as all products are worthless (because personal demands are not met) are what makes GPU and CPU sections so toxic.

It's like a reverse hype, everything current and in the future is trash unless it is the best (or at least meets personal demands).

The hype you think you're above, you encourage in the opposite direction. And we look around and see others just the same.

I'm speaking about my own experiences with recent AMD GPU's, I'm not speaking on behalf of anyone else.
 
GF has been the thorn in AMD's side for quite a while, The sooner RTG can ditch them the better.

2010: Our CPUs are giving Intel a run for their money, but our GPUs suck, quick divert resources to the GPU division!
2012: Our GPUs are giving Nvidia a run for their money, but our CPUs suck, quick divert resources to the CPU division!
2018: Our CPUs are giving Intel a run for their money, but our GPUs suck, quick divert resources to the GPU division!

I see no way this plan can end badly...

It reminded me of this for some reason :D

 
Back
Top Bottom