Performance ranking oddity

[Tomorrow]

I was comparing 2080 Ti (my current card) to 9070 and 9070 XT.
Perhaps im simply reading percentages wrong but:

When selecting 2080 Ti as baseline 100%, then 9070 is 69% faster and 9070 XT is 87% faster. The cap between the 9070 models is 18%.
However when selecting 9070 as baseline 100% then 2080 Ti is 59% of it's performance which i interpret as being only 41% slower, not 69%.
Also the cap between the 9070 models is now 11%, not 18%.

However when selecting 9070 XT as baseline 100% then 2080 Ti is 53% of it's performance which i interpret as being only 47% slower, not 87%.
Now the cap between 9070 models shirks to 10%.

What gives?

 

[DarkDreams]

A percentage is always relative. If you take the 2080 Ti as baseline, 100% is less than if you take the 9070 xt as baseline.
Think of it in terms of fps. Lets assume the 2080 Ti gets 50 fps in a game while the 9070 xt gets 100 fps. 1% of the 2080 Ti is 0.5 fps while for the 9070 xt 1% equals 1 fps.
Looking down from a higher point will always result in smaller percentage numbers. Hope that was understandable.
Basically always think 20% of X. So 20% of the performance of a 2080 Ti is different to 20% of the preformance of a 9070 xt.

 

[Hyderz]

judging from the benchmarks all around the 9070xt at least doubles in performance, game dependant sometimes more than double it is a sensible upgrade

 

[londiste]

However when selecting 9070 XT as baseline 100% then 2080 Ti is 53% of it's performance which i interpret as being only 47% slower, not 87%.
Now the cap between 9070 models shirks to 10%.

9070XT is 87% faster than 2080Ti. 2080Ti is the baseline.
2080Ti is 53% slower than 9070XT. 9070XT is the baseline.

9070 is 69% faster than 2080Ti. 2080Ti is the baseline.
2080Ti is 41% slower than 9070. 9070 is the baseline.

Using 2080Ti as baseline, 187% to 169% is 110%.

When selecting 2080 Ti as baseline 100%, then 9070 is 69% faster and 9070 XT is 87% faster. The cap between the 9070 models is 18%.
However when selecting 9070 as baseline 100% then 2080 Ti is 59% of it's performance which i interpret as being only 41% slower, not 69%.
Also the cap between the 9070 models is now 11%, not 18%.

Keep the baseline in mind. Using 2080Ti as baseline and looking at that 18%, the difference is 18% of 2080Ti performance.
Using 9070 as baseline the difference is 11% of 9070 performance.
2080Ti and 9070 performance is quite obviously not the same.

 

[JustBenching]

However when selecting 9070 XT as baseline 100% then 2080 Ti is 53% of it's performance which i interpret as being only 47% slower, not 87%.
Now the cap between 9070 models shirks to 10%.

What gives?

Math gives. 53% of it's performance means that the 9070xt is 88% faster. 53 + (88% * 53) = 53 + 47 = 100.

Percentages are relative. If you go from 1 to 2, that's a 100% increase. If you go from 2 to 1 that's a 50% decrease.

 

[freeagent]

I hate math.. nice one guys.

 

[Tomorrow]

Thank you. I think i understand these numbers better now. As for fps performance - in the games i play or im interested in 2080 Ti does around 60fps at 1440p.
9070 would raise this to 116-122fps based on reviews so that look good. Of course im not running the exact same custom scene and im using 5800X3D, not 9800X3D so i can expect lower than that but still around 100 or so.

 

[oxrufiioxo]

The database rankings are just to be used as a guide and don't always represent actual differences anyways.


2080ti vs 9070XT

73%


79.5%


81.7%


100fps should be doable in most games at 1440p there will be a few where you will need to use upscaling or turn down settings but either way very good uplift over the 2080ti.