I think Zen will be good, I've said nothing contrary to that.
But if Cinebench is so "bad" and "Intel biased", then why is it a 4670K is faster than an FX83 in Blender (Baring in mind Blender's being touted as some brilliant benchmark) but the opposite is shown in Cinebench (As you'd expect). This is something that would *never* be accepted because it's an outlier. An FX83 going 100% versus an i5 4670K is faster than the i5, not slower.
I understand people being hyped and positive for Zen (As I am) but it suddenly seems like because a certain benchmark supports an agenda, it's fair game.
I'm not saying the results are faked or anything like that, I just have absolutely no faith in a benchmark which posts outliers as per the FX83 versus i5.
I'm not shocked AMD released the wrong benchmark files originally either, but I've got no conspiracy for that. It's not an AMD product unless there's a mess up along the way.
You don understand right that in Cinebench the single threaded test shows a 4690k as scoring 154 to the FX8370's 102.... so you know it shows the 4960k as 50% faster IPC?
The 4960k has 4 cores and no hyperthreading, it's multithreaded score is 599, close to perfect scaling. The FX8370 gets a 643 score, a no where near perfect scaling score. It's still an 8 core cpu capable of running 8 threads where the 4690k is a 4 core cpu capable of running 4 threads.
There have been a half dozen pages now of you saying it's faster in Blender than the FX8370 but slower in Cinebench... but that is simply wrong. The 4690k is beating it in IPC by 50% and beating it in multicore scaling massively.
http://www.anandtech.com/bench/product/1261?vs=1709
Also, it's not an outlier there are several places the FX8370 spanks, outright spanks the 4690k, Doplhin benchmark, Agisoft build texture stage, 7-zip.
Also who is touting Blender as a brilliant benchmark? Can you find anyone saying it's an brilliant benchmark at all or is that just to add some hyperbole to your argument again? Blender is not an Intel biased benchmark, it's not depending on Intel compiling it for optimisations and it's actually a real world application.
You don't do work in Cinebench, you can do work in a SEPARATE application which hasn't got the same code or versions. YOu buy Maxon software and work on that, not on Cinebench. You get Blender, you can benchmark in it or work in it, it's that simple. It's an actual real work application that has real world meaning to people, Cinebench has no relevance.
That doesn't make Blender a brilliant benchmark, it makes it a better choice for AMD to showcase. Same with handbrake, it's not a benchmark, it's an actual application people actually use in the real world.
Number of people who do work, personal or professional, in Blender who would buy a new CPU just for more performance in a program that performance is important is magnitudes higher than those who would buy a new CPU because it was faster in Cinebench... yet you still seem to lack the ability to understand why AMD would benchmark something actually relevant to sales.