Jump to content

Labs data issue?

SwiftCoderJoe
By SwiftCoderJoe
in General Discussion
 Share
Posted

I believe I’ve found a performance data issue on the labs site on the Arc A770 review. It seems the RTX 4060’s performance in Rocket League, 1440p, is far too high.

image.thumb.png.8ee2b38f551063b355bd39324d819312.png
Compare it with the RTX 4060 at 1080p, it is almost exactly the same:

image.thumb.png.7990afbdc2e37bf8238776b386d37c9c.png

Also compare it to the RTX 4060 Ti in the  RC 7600 review, where the 4060 somehow outperforms the 4060 Ti:

image.thumb.png.399449ed897164e9cb5120c6f470e1de.png

 

It looks to me like the test was accidentally run at 1080p instead of 1440p. What do you guys think?

Link to comment
https://linustechtips.com/topic/1575897-labs-data-issue/
Share on other sites
Link to post
Share on other sites
Posted

Looking at the review from a year ago (https://www.youtube.com/watch?v=6T9d9LM1TwY), it looks like the A770 was trading blows with the RTX 3060. Knowing that the RTX 4060 pretty much outperforms the 3060 in just about every way (I think there was a fluke at 1 point where the 3060 beat the 4060, but I dont remember in what).

This honestly seems correct (though I could just be basing this on the data from 1 year ago, and I know there have been MASSIVE leaps and bounds in driver development since then).

We could even be seeing engine limitations for rocket league. Looking at other GPUs on just rocket league, we see this (taken from the 4070 page):
image.thumb.png.8a22e6a06b3d510498187a004e664831.png

image.thumb.png.59518dd190eb5d9d42524ac83525ade4.png


It really seems more like an engine limitation more than anything (and you can kind of tell since multiple GPUs from various ranges are grouped together very tightly in performance).

 

Link to comment
https://linustechtips.com/topic/1575897-labs-data-issue/#findComment-16459795
Share on other sites
Link to post
Share on other sites
Posted
  • Author

Hmm. You make some good points, but I still don’t think so. 
 

The data you showed puts the RTX 4070 (Rocket League, 1440p) at 484 FPS avg, while the A770 page puts the RTX 4060 (again 1440p) at 458 FPS avg. That’s only a  5% decrease! The 4060’s MSRP is $299, while the 4070’s is $599. No shot they have a 5% difference.

 

Even the 4060 Ti usually performs 20% worse than the 4070, and the 4060 Ti will absolutely outperform the 4060. 
 

I doubt we’re hitting any bottlenecks or engine limitations anyway since certain cards well exceed 600fps at 1440p and 700fps at 1080p.

 

I still believe this is an error. Maybe framegen or DLSS was left on, somehow making the render resolution 1080p?

Link to comment
https://linustechtips.com/topic/1575897-labs-data-issue/#findComment-16459808
Share on other sites
Link to post
Share on other sites
Posted
On 7/7/2024 at 10:28 AM, SwiftCoderJoe said:

It looks to me like the test was accidentally run at 1080p instead of 1440p. What do you guys think?

You're on the right track. It's pulling 1080p data for that graph instead of 1440p data. We'll be fixing this within the next day or so. Thanks for pointing this out.

Link to comment
https://linustechtips.com/topic/1575897-labs-data-issue/#findComment-16460825
Share on other sites
Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share
Go to topic listing General Discussion

×