This is how wrong it is to use Userbenchmark for any GPU comparisons.

[Wheresmehammer]

GTX 1070 vs RTX 2070 SUPER  RTX is +34% faster.

UserBenchmark: Nvidia GTX 1070 vs RTX 2070S (Super)

 

GTX 1070 vs RX 5700-XT  RX is +18% faster.

 

UserBenchmark: AMD RX 5700-XT vs Nvidia GTX 1070

 

Then we have a real review.

 

 

 

 

Say no more...

[Coolmaster]

12 minutes ago, Wheresmehammer said:

GTX 1070 vs RTX 2070 SUPER  RTX is +34% faster.

UserBenchmark: Nvidia GTX 1070 vs RTX 2070S (Super)

 

GTX 1070 vs RX 5700-XT  RX is +18% faster.

 

UserBenchmark: AMD RX 5700-XT vs Nvidia GTX 1070

 

Then we have a real review.

 

 

 

 

Say no more...

While I don't like UserBenchmark, I think it factors in productivity into its score too, which Nvidia is much better than AMD at.

[Wheresmehammer]

Just now, Coolmaster said:

While I don't like UserBenchmark, I think it factors in productivity into its score too, which Nvidia is much better than AMD at.

It definitely does not factor anything of the sort in and if an application suits CUDA more then Nvidia will be better for it, +34% is quite low even for the 2070 SUPER vs the 1070.

[Letgomyleghoe]

1 minute ago, Coolmaster said:

While I don't like UserBenchmark, I think it factors in productivity into its score too, which Nvidia is much better than AMD at.

That's not always true, it depends completely on what you're doing.

and the percentages that are used in the OP are the gaming EFPS.

[Mark Kaine]

if you're comparing gpus you should at least get the numbers / names right o.o

 

 

1070 ≠ 2070

 

Also everyone here knows userbench is trash, and I couldn't care less about "hardware unboxed" every single yt reviewer has some kind of agenda.

 

only way to test a gpu reliably is to have it in my pc... 

[Wheresmehammer]

Just now, Mark Kaine said:

if you're comparing gpus you should at least get the numbers / names right o.o

 

 

1070 ≠ 2070

 

Also everyone here knows userbench is trash, and I couldn't care less about "hardware unboxed" every single yt reviewer has some kind of agenda.

 

only way to test a gpu reliably is to have it in my pc... 

The point was the percentage difference between the 1070 & 2070 SUPER to the 5700-XT percentage difference from the 1070 and the real review only showing 1-2% between the 2070 Super & 5700-XT.