![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
There are reasons to suspect that Win 7 being faster than Win 10 is a function of nVidia cards being used as GPUs. Swap that with AMD stuff and suddenly Win10 is way better than Win7. At least that's what looncraz claims.
Their SMT approaches are hardly different (it is difficult to be creative with SMT-2).
Because it is the nature of the threads and the data (such as size of chunks that can be processed,associated dependencies-sharing,etc) ; use is very different between RealBench and those game engines that are impacted, in fact some other game engines may be as well because it depends upon what options the reviewers/gamers set.
Have you seen any AMD GPUs used yet. Now that you mentioned it, I cant say I have but that would be interesting.
All of that is interesting, unverifiable, and completely irrelevant. The fact is no one should wait on a hypothetical. We all know new tech is released constantly, so you you make the best purchase you can when you need it and don't look back.
I call that using your head and not purchasing sight unseen a very expensive product that has not been independently verified.
I think the main reason for lack of Radeon tests is due to them not having a high end part to test with.
I mean, you could always just run games at 720p or below and even Fury would make CPU the bottleneck.
It's about the guarantee or otherwise so many reviewers would have used a 1070 or 1080 instead of a titan.
I mean, computerbase.de used to use 980 Ti for their CPU testing until this February and the only game that was GPU bottlenecked at 720p was TW3. It is still basically GPU bottlenecked at 1080p with Titan XP with sufficiently fast CPU though.
Sorry but seeing all the testing would leave me in a WAIT and SEE approach. The testing does raise questions between 7 and 10 so regardless of some statement even from AMD I am hesitant to accept it as something just doesn't add up. Besides aren't most of you always saying never trust AMD statements, all of a sudden you want to take them at their word. Given the plethora of tests, although some look a little off, I feel there is need for further in depth testing. Even PCPers tests were scheduled scripting that don't actually mimic games random behavior, as quite a few believe. So far we have no conclusive info either way, at least not a lot.A pile of BS. Now he pretends that the commercial chips behave worse than he said because they are in reality engineering samples
That thread started fine with some very interesting tech info posted by The Stilt. It has now degenerated on a conspiracy theories fest. People in that thread pretends that the scheduler issue is real and that AMD is officially denying it because "Wintel" (sic) is obligating them to do so. That thread title would be changed to "strictly BS".
It is not bad, it is *worse*.
But, and this is the baffling part, Win 7 shows in most benches with pics that it doesn't populate the SMT cores almost at all. Now in windows 10 it shows populating all cores + SMT equally. If that is accurate with no shenanigans from the posters/reviewer then there is some work to be done. Unfortunately most of the delving has been half ass, not because they aren't trying but I think they aren't entirely sure what they are looking for or what it means. Kind of like how most all 9590 reviews showed terrible performance simply because no one knew how to OC it correctly ( my 8350@ 4.6 with slower ram and worse GPU - a lot worse GPU, posted better scores).
Can you link the examples of both that you mean as it does not necessarily match what PCPer showed.
The easy one is a thread here
Cant say wish I had a 1700 to test. But that is how my FX does because of the modules. In both 7 and 10. It loads 0-2-4-6 then 1-3-5-7, 0-1 being 2 cores one module. I can say that win 10 loading so equally seems wrong, to what degree I cant say. SMT would make sense if the load were just a touch higher on the physical core.
Before your argument comes valid you must consider core count. You mention price and performance but not getting 4 more cores for the same price or slightly higher.
