I believe they're an inept company... plain and simple. Don't confuse me belief of their ineptitude to actually provide tools to OC well with a thought that I actually believe it doesn't OC well.
Simply put, the company does stupid things. I don't need to explain that statement any further really you can just use your memory.
So no, I expect the card to OC at least decently. OC dream? Who knows, but I expect it to be ok. I just at this point my expectations of AMD are actually this low that AMD launching a card, claiming it to be an OC dream, the card OCing decently, and not launching with OC capabilities is something AMD would do. It's literally not something that would surprise me.
That's just where I'm at and that's really not a point you want your consumer to be at where a used sub R9 290 is still something your consumer has to think about because they just don't really trust you.
You might be right, I've been thinking something similar.
This whole business of 'forward looking architecture' doesn't make any business sense, because gamers are going to buy what's good NOW.
unless, of course, they're planning to stick with GCN for a while longer. I guess that could explain that decision.
Anyways, my point being they need to release an architecture that's good now because most people are n99b idiots who just buy whatever benches best and don't honestly evaluate GPU architecture or have the credentials to forecast what will be a better architecture in 3 years time.
It probably was them planning to stick with GCN for 3 more years, in which case I can forgive them. However, hiring two engineers to manage memory issues NOW is way too late-- they needed to do that 2 years ago when they decided Fury X was going to be 4GB only, and they needed to do that by promoting or shifting resources within, not hiring new guys and asking them to do it, having had no exposure to the driver architecture or implementation.
But for overclocking, yeah, I can see that they produced the hardware for it but didn't have the communication to make it happen on the software side.
Lately my greatest theories on how to efficiently run a company would be having a few highly paid full time low-levels in each section of the company who observe everything and report their opinion to the CEO, who then is basically required to implement their direction.
Before I get to that, let me touch on synergies, which probably would have solved this OCing problem: if you were friends with someone in the Catalyst software development side of things, you eat lunch with him on Fridays-- eventually over the course of 2 year dev cycle it would become apparent 'no, I haven't been given the specs to implement the voltage control on the Fury X, in fact no one has even mentioned it, we just assumed it would be same as the 290x' 'no, it's not' 'ok, lets schedule a meeting with Bob and get this worked out'.
ok, returning to CEO idea-- for example, the low level embedded engineer who's been there for 8 years needs to be friends with the driver architecture engineer who needs to be friends with a game dev, and between them they would probably be able to cook up the future AMD's GPUs need to go better than the CEO could. Actually, scratch this idea, this is probably what happens to some extent, just abstracted through management. I think what I'm getting at is this: I have sometimes had visionary ideas, which if implemented correctly, could make someone a lot of money. Most don't implement correctly, which is why when someone does, like Apple, they get so far.
It kinda makes sense though. If you take any ten devices that all draw 1000Watts, they will all output the same amount of heat. The only exception would be whatever amount of that power is transformed into radio waves. So a microwave or a radio transmitter would heat a room less than say a toaster or a hairdryer. I guess it gets more complicated if you take a 1000W freezer and run it with the door open, but I'm not even gonna go there.
Facebook
Twitter