I think the last patched Prey can be usefull as benchmark to really discover what are the real Jaguar performances... What I know is that not automatically a stronger CPU is overall better in an unified RAM environment... Aka is better a 16xjaguar@3ghz or 8xryzen ?!? Transistor, power... Performances... Backward & Forward compatibility
high preset is what is seen on ps4pro @ 30 fps... so lets see the fps of the i5-6500 (that runs 90%)... with a comparable gpu...
http://www.tomshardware.com/reviews/prey-2017-benchmarks,5072.html
the rx460 2,2TF vs ps4 1,8TF looks to be the best match...
http://www.tomshardware.co.uk/answers/id-3135987/amd-radeon-460-tflops-ps4-tflops.html
on very high preset RX460@2,2TF runs between 28 and 40 fps vs 30 fps of ps4@1,84 on high preset... Somerhing like 20% more calc power... and you got something like a 30% better results... Lets say 50% better result considering the high vs very high preset... So I would say the CPU handicap here is a 30%
so my conclusion is that 8 jaguars@1,6 ghz behave like i5-6500 - 30%... Jaguars are not so bad into an unified memory system... In the article is stated this game is optimized specially for AMD... So ok, lets say i5-6500 - 50%...
so 16 Jaguars @ 1.6 ghz = 4 cored i5-6500@3,2 ghz.... (3,6 ghz turbo)