No matter how you look at it, they were hammering Intel... and then the Core 2 Duo's came and it was done. A strong "reply" is long overdue.

Intel sat on Netburst for an eternity and it got dumpstered by the Athlon 64... then AMD did the same Intel did, sat on their asses for an eternity spewing out the same technology. Core 2s came around and Intel has been dominating since 2006. AMD won't even have a reply by 2016, or 2020 for that matter. At this rate they'll never have a reply and it really does seem like problems at the top of the company.

Intel sat on Netburst for an eternity and it got dumpstered by the Athlon 64... then AMD did the same Intel did, sat on their asses for an eternity spewing out the same technology. Core 2s came around and Intel has been dominating since 2006. AMD won't even have a reply by 2016, or 2020 for that matter. At this rate they'll never have a reply and it really does seem like problems at the top of the company.

I tend to believe that they've learned their lessons. If AMD would release something more powerful than current i7s, I wouldn't be surprised if Intel would blow them out of the water again with a new launch just a few weeks after. Just like nvidia's driver optimizations shortly after mantle's release. I mean let's be honest, they can't pop up such a polished driver with such performance increases across the board in such a short amount of time. Nvidia's driver team is good, but it's not that good. From a business point of view, it makes sense to hold technology to yourself until a release is needed.

A bit? https://www.cpubenchmark.net/high_end_cpus.html As you can see, intel's top-end, which is the i7-5960x is 60% (!) faster than AMD's top-end, the FX-9590. Intel DOES have high desktop/server prices for that exact damn reason, that i7-5960 is over 1k USD. How much do you think it would cost if AMD wasn't so damn behind? And the FX-9590 is a poor excuse of a cpu. They just took good FX-8350 chips and overclocked them. Nothing more than a money grab, those cpus were 1k dollars when they launched. Look at the price now. And you know what the deal is? You can't even blame them, a lot of people bought them because 4.9GHz must mean they're faster than intel's 3GHz right? No. And do you really think intel makes no money from mobile chips? I really hope you're joking. If there was no money in that market, nobody would have bothered making them. Mobile chips are a direct result of desktop chips R&D, they do not spend a lot more in making mobile chips after they have their architecture up and running. Take my i7-2670QM for example. It's just a gimped version of my i7-2600k. Reduce the clock, cut the cache size, gimp the igp and pass the least amount of voltage possible through it. Result? A gimped 2600k. http://ark.intel.com/compare/53469,52214

Firstly, I wasn't talking about gaming. Since I said I own 2 i7s, that should be enough of a hint that I need that raw processing power. Rendering, encoding and sheer cpu power for audio work. Take the following scenario: Cubase with 20 tracks, all of them vst plugins with big and complex libraries. Do you know how much cpu power that needs just to play my compositions? You would be surprised. And it's pretty obvious you haven't worked in the field of modeling and rendering (I'm not a professional, just a mere hobbyist), there are many limitations with using opencl or cuda when rendering. Some examples will include: you sometimes have to triangulate all your meshes (although this will not be the case soon, many gpu renderers already support quads), you can only use certain raytracing and antialiasing algorithms. You can only use certain materials, certain lights. Textures are also generally limited, you have to use time-consuming workarounds in order to get the desired effect. Don't you tell me about rendering boy, it's not all sunshine and bunnies as you seem to believe. And about getting more cpu power than you need, it's justified getting an i7 for gaming. The immediate factor is that you WILL see a bit of a framerate increase, the long-term factor is that you will not need to change that cpu for a long long time. It's actually more cost-efficient getting an i7 (obviously not the 1k dollars extreme ones) over getting an apu and changing it every 2 years. Look at my rig. This cpu is in it for 3 and a half years already and I'm willing to bet it will still be here for 3 more years at least. I don't need to change it. If you had a 3-year-old apu, you would probably need to change it now if you would buy a 970 for example. And then change it again in 3 years. And if need be, I can push it to 4.8 GHz like it sais in my specs since it's currently hovering at 4 GHz. I can squeeze a lot more time out of it. Your vision on high-end cpus is very flawed my friend. They do have a purpose, even for normal-ass gamers.

You don't understand. Vst instruments simulate the way real instruments actually work. As an example, to get a guitar sound it's not possible to trigger a signal and filter the sound so that it resembles a guitar sound. I mean it IS possible, but it sounds bad and artificial. You need the pick sound, the picking direction (up, down), automating the direction in which it strikes the strings (alternating between up and down when needed), variations in volume, variation in velocity, calculating the fact that you cannot possible strum 2 strings at the exact same time (there is a slight delay), making sure you don't hit 2 notes which conflict with each other (eg 2 notes on the same string cannot be played at the same time). And this is just 1 instrument. I compose symphonic metal. I need 2 guitars, a bass, a drumkit and a lot of orchestral instruments. That requires a good amount of processing power. This is also why good organs/keyboards are so expensive. Soundcards have nothing to do with this process, they just take the data and output it in a manner you can hear. The better the soundcard, the more accurate the sound. You are right about video encoding, opencl/cuda helps a lot with that. But I stand by my statement about 3D renderers. They are more limited in what they can do compared to classic CPU renderers. The end result might look approximately the same but the process is different. You don't have the same amount of 'freedom'. And trying to bottleneck to under 60 fps highly depends on the game used. Depends on how cpu-intensive the game is mostly. Crysis, metro, arma 3 come to mind.