• 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.

Alder lake + 6900xt and RTX3090, heaven or hell?

Mesai said:
Rather just replace Intel with AMD and then you're more power efficient regardless :cry:
Click to expand...


Yes but you lose performance so there would be no point - this test shows that Intel can be a better option for high end AMD GPUs (indirectly), the main objective of this data I guess was so Igor could point out how much more inefficient An Nvidia setup is in CPU bottlenecked scenarios, but it indirectly also showed that Intel CPUs make for a better match when paired with an AMD GPU (at least a high end one)
 
Last edited:
Grim5 said:
Yes but you lose performance so there would be no point - this test shows that Intel can be a better option for high end AMD GPUs (indirectly), the main objective of this data I guess was so Igor could point out how much more inefficient An Nvidia setup is in CPU bottlenecked scenarios, but it indirectly also showed that Intel CPUs make for a better match when paired with an AMD GPU (at least a high end one)
Click to expand...

Sure, I can understand the reason behind the test. On a more practical level, it won't matter to 99% of users.
 
Mesai said:
1080p and 1440p on Steam account for 75% of the surveyed userbase. Not sure which market segment games at 720p on an recent hardware, not that this test is that relevant to gaming.
Click to expand...

No arguments here, except to say that in terms of efficiency while gaming I believe the test is relevant.

There is a lot of talk on here about efficiency with every new CPU release, so knowing why review results conflict when it comes to efficiency has value. Alder Lake especially has been the subject of a lot of efficiency related talk, so knowing that using an AMD GPU has the potential to be more efficient with Alderlake while gaming than a Nvidia one has value.
This is particularly relevant for anyone buying a new system.
If there wasn't so much talk about efficiency, and Alderlake wasn't the subject of so much scrutiny when it comes to power draw, then I agree the test wouldn't be that relevant. But people do seem to care.
The test was conducted at 720p to highlight the issue and ensure the CPUs were being tested, but the reviewer notes that the trend continues at 1080p which is the resolution the overwhelming majority of gamers use, if the Steam hardware survey is to be believed.

As for the statement "No one across three quarters of the world games at 720p"
I checked the steam hardware and software survey as you did, and I noted that for every two people playing at 4K there is at least one person playing at 720p.
A not insignificant amount of people.
I can't see any breakdown for where the survey results are coming from, so being able to separate the results by region isn't possible as far as I can tell. For both of these reasons I ruled Steam out as the source, and I'm assuming the statement is just incorrect.
 
mickyflinn said:
720p :cry:
Click to expand...

I wish people would stop laughing at this, if we only test at the GPU limit we never learn anything about these CPU's, this is something we have learned because of this sort of testing, Nvidia's GPU's have a performance hard limit, at least Ampere on ADL in this game.

3090 + 12900K topped out at 127 FPS.

2070S + 5800X. 149 FPS. What's going on here?

tNQtL9A.png
 
Distracted said:
No arguments here, except to say that in terms of efficiency while gaming I believe the test is relevant.
Click to expand...

I agree, and probably something that will become more relevant as everyone gets more creative with the various architectures.

humbug said:
2070S + 5800X. 149 FPS. What's going on here?
Click to expand...

Wonder whether that's related to the hardware scheduler that Intel are using - potentially not properly allocating the hardware-bound Nvidia processes, which the AMD units don't have to deal with?
 
humbug said:
I wish people would stop laughing at this, if we only test at the GPU limit we never learn anything about these CPU's, this is something we have learned because of this sort of testing, Nvidia's GPU's have a performance hard limit, at least Ampere on ADL in this game.

3090 + 12900K topped out at 127 FPS.

2070S + 5800X. 149 FPS. What's going on here?
Click to expand...

He is running 720p Ultra, you are doing low
 
Mesai said:
I agree, and probably something that will become more relevant as everyone gets more creative with the various architectures.



Wonder whether that's related to the hardware scheduler that Intel are using - potentially not properly allocating the hardware-bound Nvidia processes, which the AMD units don't have to deal with?
Click to expand...
Who knows but it is true that AMD's Thread Scheduler is hardware based, its physically on the silicon, Nvidia use a software approach which is good for Main Thread limited DX11 but for DX12 and Vulkan this can be a problem, because Nvidia is using resources of the CPU that is done by a dedicated part of the silicon on AMD's GPU, and that can put up to 30% more load on the CPU vs AMD's solution, it can get to the point where a slower AMD card actually gets higher frame rates than a faster Nvidia card, Even in DX11 where the Main Thread is not the bottleneck which in modern DX11 titles the Main Thread is less of an issue.

That's the background on it, so given that Nvidia rely on the CPU like that Intel's Thread Director may not be in agreement with Nvidia Drivers, the somewhat more cynical view is Intel could be beating Nvidia over the head with this by refusing them access to optimise, Intel have their own GPU's coming out.
 
You are rendering entirely different things, which will lead to the CPU doing different setup work, it may not be that of course I don't have the game to test but seems like an apples vs oranges comparison.
 
sandys said:
You are rendering entirely different things, which will lead to the CPU doing different setup work, it may not be that of course I don't have the game to test but seems like an apples vs oranges comparison.
Click to expand...

You could be right, certainly its not an Apple to Apples comparison, my problem is i don't have the GPU muscle to make it CPU bound at Ultra Settings, even at 720P.

720P Ultra, but this is almost entirely GPU locked.

25kaK8n.png
 
humbug said:
I wish people would stop laughing at this, if we only test at the GPU limit we never learn anything about these CPU's, this is something we have learned because of this sort of testing, Nvidia's GPU's have a performance hard limit, at least Ampere on ADL in this game.

3090 + 12900K topped out at 127 FPS.

2070S + 5800X. 149 FPS. What's going on here?

tNQtL9A.png
Click to expand...
Sure there testing the cpu but how releavent is it for most users ?
 
You must log in or register to reply here.
Back
Top Bottom