Actually, it's not too far.  At worst you've have to drop the AA, soft shadows or volumetic fog in games but after that, you'd not really see much difference.

So why are you only comparing the performance of 1 of the GPU's when the system I posted has 2?

And yes, I used GFLOPS as it the closest means of measuring comparable performance across architecturally different products.

$550 without OS, if you take all the cheapest components, but by the time PS4 launches it will probably be under $500...which will be great, I fully expect DF to build one and make face-offs with it against PS4 and XOne.

a) Not sure what are you asking me - your config has GPU inside APU plus 6670, working in Dual graphics mode...together they perform (at best) at the level of 6770/7750 (as for Dual graphics with 6770, official specs do not support that)

b) closest means of comparing performance across different architectures are...benchmarks

Anyway, you can get PS4 comparable PC (without OS) for as low as $550 (if you don't mind some really cheapish components)...and to be on the level of PS4, due to Win overhead and due to PS4's architectural enhancements, it does need GPU that performs better than HD7850...so either 660GTX or HD7870 (I would most certainly go with later).

That would still not match the ps4's performance. You simply cant build anything with 8gb DDR5 dedicated to graphics. The reason the ps4 could not be made is because it's an SOC that reduces cost. Building from the equivalent part would cost more. One more thing about gpu, there is that weird notion that you guys believe that single point performance is the sole meausure of power. Even though the 5870 matches the ps4 in single point performance, that does not make them equal. And don't forget about the controller and other accessories.

single point performance of gflops is the worst measure of performance across architecturally different products. You have it backwards. The Gflops difference between two gpu of the same architecture is the best way to go. However across different architecture it's a dead end. The reason being that one architecture might be suited for optimal performance per cycle, whereas the other one might perform few calculations per cycle but a lot more cycles. Now over the course of the same time they will perform the same amount of calculation know as flops. However if the task at hand deals with more calculation per cyle than they one best suited in that area will will win by a wide margin even though they can compute the same amount.

So flops comparison is no good unless we are dealing with the same architecture. Need further proof think of what a cpu does at around 200-400 gflops against a gpu computing physics at 2tflops. In fact gpu only started to compute physics only three years ago.

You could build a PC with a comparible price with comparible specs to a ps4. It's not hard, you also fail to realize that bluetooth is necessary on a PS3 because it only has to usb ports, while a cheap motherboard will come with 4-6 along with 2 more on the case. Wifi doesnt cost much either. That's only $10. You also need to realize, that computers are better price per performance once you get into the $600+ range. So a $600 pc will have better price per performance the a PS3. It makes sense, that's what happens near the end of a console generation.

You're under the assumption the PC needs 8gb of GDDR5 on the graphics chip.
Remember, it's only Ram, it doesn't do *any* form of processing itself.
In the PC world it's essentially another level of cache that can hide the bandwidth and latency hit of using System memory (Which PC GPU's can also use.)

By your logic, because I don't have 8gb of GDDR5 of system memory, my PC by default is slower than the PS4, which is laughable.

Besides, the PS4 doesn't even have all 8Gb dedicated to graphics tasks anyway, 1Gb is reserved for the OS, that drops it down to 7Gb, the rest of the 7Gb is certainly not going to be dedicated to the GPU, a massive chunk... Sure.

I actually own one for 3D blu ray movies.....i just have my computer set up in my living room on my 60 inch tv...

I also use it to burn cd's as my car is a 97' WRX STi and has no adapters for an mp3 player, but that could be on a typical DVD as well.

It doesn't need 8gb of gddr5 dedicated memory and it is not just cache. I used 8gb because I don't know how much the os will occupy it could be one gig or more who knows. When it comes to graphics memory the speed is what matters thus the gddr5, so when you have 8gb of super fast memory it's a clear advandtage over 8gb of ddr3 in graphics for the following reasons. As you play a game you are constantly streaming assets from the cpu and gpu. How fast you ram can take the assets, therefore it speed, the faster you can show them on screen. in order words if you have slow ram with powerful gpu and cpu the performance will be hindered by the speed of the memory. In order words, even thought the assets would be more lifelike the frame rate would be choppy.

As far as I know there are no graphics card that support 7gig of ddr5 memory. So yeah it's not just cache. Hard drive are cache as well but the difference between ram and your hdd is speed. SSd is derived from ram architecture. Any kinda of memory or anything that holds on to your ressources are cache. But they behave differently based on their speed. For that reason, you just can't build a ps4, not saying you can't match its visuals, but you can't meet its visuals on your own.

I dont ever remember saying that anybody's computer would be slower, but there would be areas where you would have shortcomings. You can't fit 7gb worth of data in a 6gb machine. Your machine would slow down, and the effect would be based on how fast your memory can quickly swap data, THEREFORE THE IMPORTANCE OF SPEED IN RAM. This why intel's next generation of cards will support their crazy fast way of swapping memory. A computer with 20gb of 100 hertz memory would be outperformed by a wide margin by a ps2 with bareley 56mb at 200 or so mhz. Not sure what the real values of the ps2 are but the analogy refers to how fast the ram displays on screen, which directly affects frame rate

Based on rumours, the os reserves 1gb of memory leaving 7gb to games. There is no way on earth the cpu will use 1 gb of memory, unless you're not optimizing or you're sharing graphic tasks with the gpu. Ultimately you still can't match that performance with equivalent parts, because you can't build as of today a computer of similar parts. You can exceed its performance, absolutely can you recreate not a chance as of now.

Learn more about memory, it is a fun topic.

So you are agreeing, yet disagreeing with me.

The GDDR5 Ram on a PC is just a cache! A cache dedicated to the GPU.

For example, take AMD's Cayman architecture, the L1 and L2 caches in the GPU were unchanged over Cypress, but AMD still threw twice as many Tessellators in the GPU. - That's allot of geometry information which can't fit in those tiny caches, so where is the Geometry supposed to go? Back to a larger but slower cache, enter: GDDR5 memory.

AMD and nVidia call the memory next to a GPU an "Off-Chip Buffer" for a reason, it's a cache.
https://en.wikipedia.org/wiki/Cache_%28computing%29

And then you contradict your entire post with this simple line.

That's the point of a cache hierachy's on the PC, so you don't have to.

Why do you think all memories in a PC go from Slowest+Largest to Fastest+Smallest? It's to hide latency and bandwidth deficits in all stages of memory retrieval whilst keeping costs down so you don't need 16Gb of L1 cache each on the GPU and CPU.
Remember, GDDR5 is pitifully slow compared to L1 and L2 caches.

The need for 8Gb of GDDR5 on a GPU isn't needed now or in the immediate future on the PC, main graphics tasks will be stored in the current GPU memory  with excess data in system memory which will get swapped into the GDDR5 on demand, in-fact, that's how it's always worked with PC graphics, even before the AGP bus or DDR was ever invented.

And by the time that GPU's will need 8Gb of GDDR5, JEDEC should ratify the GDDR6 standard anyway, making the PC a generation ahead in memory technology. (And is due out next year.)