I mean according to "tests" - they are getting +10%. Maybe they just used same platform and just set pcie speed to 2.0 or used kind of close platforms (like 2600k vs 3770k). I thought same way, that actual performance drop is small (that's why I bought 1080 in old PC)...
But if you have really old platform like P67 - actual loose is more than 10%, and I guess not only because of pcie itself - but because of platform as whole thing.
>> Including the RAM & not just the Chipset, Massive CPU upgrade & Massive GPU upgrade
Well, another funny thing is that RAM settings doesn't affect gpu score in 3dmark at all, but affect cpu score: I tested memory speed from 2400 to 3600 and got like nothing in gpu test and about +5% in cpu test. So I don't know why I got such a huge difference between 1080 and 2080ti. It is not a RAM, it is not a CPU as CPU power (or 2600k @5ghz bottlenecking gpu)...
I did that to test what is bottlenecking Il-2 and found out that it was not memory, but cpu because of game engine. First I thought it must be both cpu and memory.
I did same for CPU - from 3200 to 5000, and again nothing: almost linear behavior for cpu test (well, obvious...) and zero changes in gpu test.
post edited by DeadlyMercury - 2019/09/27 13:47:21