• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

[Computerbase] RDNA 2 CU Scaling

Tripolygon

Banned
cerny said developers basically see some kind of expected performance profile or something along these lines, and that it's basically always easy to know how the hardware will behave at any given moment.

You are mixing a bunch of things together. Smartshift is simply just a way to share power budget and move power where it is needed instead of having separate power for CPU and GPU. How it is used on a Laptop is definitely different from how it is used on PS5 because their thermal design and power needs are different, but it is still the same underlying AMD power management technology.
 

01011001

Banned
You are mixing a bunch of things together. Smartshift is simply just a way to share power budget and move power where it is needed instead of having separate power for CPU and GPU. How it is used on a Laptop is definitely different from how it is used on PS5 because their thermal design and power needs are different, but it is still the same underlying AMD power management technology.

yeah but it's not really the same thing if it behaves differently... but then again, AMD calls FSR 1.0 and FSR 2.0 the same thing only with a version number change... so I guess there's precedent for calling completely different things the same name
 

Panajev2001a

GAF's Pleasant Genius
cerny said developers basically see some kind of expected performance profile or something along these lines, and that it's basically always easy to know how the hardware will behave at any given moment.
That is just a mode you can select only for development while you are optimising.
 

assurdum

Banned
yeah but it's not really the same thing if it behaves differently... but then again, AMD calls FSR 1.0 and FSR 2.0 the same thing only with a version number change... so I guess there's precedent for calling completely different things the same name
It doesn't behave differently. Learn to read. And why persist to say it's a different thing.
 
Last edited:

Tripolygon

Banned
yeah but it's not really the same thing if it behaves differently... but then again,
It’s the same thing. Smartshift on an Asus laptop will behave differently from an HP laptop because they are designed by different companies and have different power and thermal design and needs. Smartshift on a thin and light laptop will behave differently from a gaming laptop because they again have different design goals.

Fundamentally they all use the same AMD power management technology that is on the second generation now with newer laptops coming out AMD added more features to it.

AMD calls FSR 1.0 and FSR 2.0 the same thing only with a version number change... so I guess there's precedent for calling completely different things the same name
This is not true. AMD delineates both with 1.0 and 2.0 and makes a big note on talking about how they are different, even making lots of technical talks on how they are different. They are both part of AMDs upscaling (super resolution) tools but they work differently. There is a whole presentation talking about how they are different.
 
Last edited:

01011001

Banned
It doesn't behave differently. Learn to read. And why persist to say it's a different thing.

it can't behave like on PC, it simply can't. and we know it doesn't as on PS5 the hardware always rund max speed until it hits a power limit no matter what is happening on the hardware
 

01011001

Banned
This is not true. AMD delineates both with 1.0 and 2.0 and makes a big note on talking about how they are different, even making lots of technical talks on how they are different. They are both part of AMDs upscaling (super resolution) tools but they work differently. There is a whole presentation talking about how they are different.

they still use the same name, and they aren't comparable things whatsoever.

one is a shitty upscaling algorithm that uses no temporal information or anything.
the other is a reconstruction method that uses temporal data and motion vectors

and the naming scheme confused a lot of people and still does
 
Last edited:

Tripolygon

Banned
they still use the same name, and they aren't comparable things whatsoever.
FSR stands for Fidelity Fx Super Resolution. 1.0 and 2.0 differentiates them. It’s not 1.1 or 1.2 to show progression of the same.

Definition of 2.0​


—used postpositively to describe a new and improved version or example of something or someone
one is a shitty upscaling algorithm that uses no temporal information or anything.
the other is a reconstruction method that uses temporal data and motion vectors
Yea I know what they are and again there are technical talks that talk about them in detail.
and the naming scheme confused a lot of people and still does
The naming was not confusing. People are just generally confused by a lot of things they don’t understand. Like you for instance who seem to be confused about what Smartshift is. This is why Digital Foundry and NXGamer serve as a good source of explaining things so everybody can understand.
 

winjer

Gold Member
they still use the same name, and they aren't comparable things whatsoever.

one is a shitty upscaling algorithm that uses no temporal information or anything.
the other is a reconstruction method that uses temporal data and motion vectors

and the naming scheme confused a lot of people and still does

But DLSS has the same issue then.
DLSS 1.x was even worse than FSR 1.0
And it used a different technique to DLSS 2.x
But o one complains about having the same name.
 
Dont think you can apply the findings of this particular test that normalizes clocks to the PS5 and XSX archs which have a pretty significant difference in clocks.

Now if the test measured the 40 CU 6700xt with the 40 CU 5700xt at 2.23 GHz and 1.825 GHz then we might be able to make a proper comparison between the two consoles. Maybe DemonCleaner DemonCleaner can do the comparison because DF doesnt seem to be particularly interested.

don't have no rdna GPU at the moment.

but im pretty sure i published this data in my graphs already


Preface: THIS THREAD IS FOR THOSE INTERESTED IN TECH ANALYSIS/DISCUSSION OF RDNA 2 AND THE DESIGN CHOICES FOR PS5 AND SERIES X. NO CONSOLE WARS PLEASE!

https://www.computerbase.de/2021-03/amd-radeon-rdna2-rdna-gcn-ipc-cu-vergleich/2/

Awesome analysis that I think is the most comprehensive in comparing RDNA 2 CU scalability between AMD 6000 series cards, all fixed at 2ghz clock frequencies. Results posted below, but I wanted to point out a crucial point from a different but related test they conducted that determined RDNA 1 CUs are actually faster than RDNA 2 CU due to the shorter ALU pipeline (this should serve as a reminder to some that RDNA 2 isn't inherently better than RDNA 1 across the board AMD) .



Applying the findings below to the premium consoles, there are a few interesting facts from my perspective:

1. Cerny's argument of CU utilization doesn't pan out in the games tested below. CU scalability remains relatively constant from 1080p-4k. Although it is possible that current gen games could yield a different outcome.

2. Series consoles chose RDNA 2 design with inherent CU latency increase compared to RDNA 1, without offsetting via higher clock frequency (2ghz+) as AMD and Sony have done with 6000 series and PS5, respectively. Based on the testing, it doesn't appear as though Series X 560gb/s throughput would be enough to compensate to achieve the levels AMD was able to with faster clocks/cache. As shown below, A 50% CU advantage AND 33% bandwidth advantage for the 6800 over the 6700xt resulted in only 36% performance increase at 4k (again, both GPU clock frequencies fixed at 2ghz). Series X has 44% CU advantage and 25% bandwidth advantage ceiling over PS5 with its fastest memory segment. If anyone has information as to why Microsoft deviated from AMD RDNA 2 strategy, I would be interested to learn about this (proprietery hw/sw, etc.).

while i feel MS played it a bit safe with the conservative clock rates in XSX (probably because of yield projections) i need to stress that they also benefit from the reworked ALU pipe as that not only enabled higher clock speeds but also a much improved (meaning less power hungry) voltage/frequency curve.
 
Top Bottom