AMD's a-comin!!!... and Intel's been a-dunnin!!

[Bubbamusic]

Here it is! The ultimate DAW.......and it is affordable:

http://www.cray.com/products/computing/xc-series

[fluffy_little_something]

Here it is! The ultimate DAW.......and it is affordable:

http://www.cray.com/products/computing/xc-series

Is there a mobile version?

Some get incredible performance out of ordinary chips, seems cooling is the key obstacle to higher performance:

https://hwbot.org/benchmark/cpu_frequency/halloffame

Crazy, standard AMD chips that cost less than 200 dollars, clocked at twice the turbo clock speed

[Kaine]

And still under-performing an quad core i7 at stock

[fluffy_little_something]

How do you know? I have not seen any benchmark results for those extremely overclocked processors...

[Kaine]

Because overall system performance isn't just about clock speed. Even with the benefit of the doubt the point is moot because you can't hold a system stable enough to work on whilst running under N20, as you say, you've not seen benchmarks because all those guys do is run Pifast or whatever the flavor of the month is, pure maths and never for more than a few mins.

The reason they are behind is down to the overall architecture and the amount of time it takes to pull the data through the various memory and cache levels, none of this improves with overclocking the CPU, well it does if done right, but not to the level we'd need to see to set the world on fire (maybe the desk if you run out of coolent however). This is the reason they haven't really advanced much over the last 5 or 6 years, they've bottle necked the chipset design. It's the same problem that kicked Intel in the ass with Prescotts and at that point it was AMD moving in a new direction that gave them the edge whilst Intel back pedaled. When Intel moved the memory and other various controllers onto the CPU die, that is when we saw the mega advancement in performance and is exactly what AMD is doing with the new platform... which is the reason why we are excited.

[fluffy_little_something]

Sure, that kind of cooling is not sustainable. But for a couple of minutes the performance might actually reach or surpass that of the fastest 4-core i7.

I don't know how clock speed scales in terms of performance, but the performance gap between the FX-8370 (which seems very popular with overclockers, but is only AMD's #3 so far) and the fastest 4-core i7 (7700K) is about 23,400 : 38,800 according to this benchmark site: http://www.anandtech.com/bench/CPU/39

According to this site the gap between those two processors is even smaller, namely about 8950 : 12,300.
https://www.cpubenchmark.net/high_end_cpus.html

So, if the performance increase due to overclocking is nearly linear...

[Burillo]

Some get incredible performance out of ordinary chips, seems cooling is the key obstacle to higher performance

i'm surprised you just found out about that. yes, among other things, it's about cooling. the reason Pentium 4 was such crap is that it had an overly long pipeline that was designed to eventually run at 10GHz, but laws of thermodynamics thought differently, and you still can't run any chip past 9GHz, liquid nitrogen cooled or otherwise.

[fluffy_little_something]

It says you are from hell, cooling your CPU must be even more of an issue down there

[Burillo]

nah, we get enough thermal energy to power air conditioning

[Kaine]

So, if the performance increase due to overclocking is nearly linear...

Yeah, if your just asking it to carculate Pi (which seems to be one of the core tests for extreme overclocking) and all the information is already available to the CPU, then yeah, the raw performance is there. As soon as you have to call and reference data not already on the CPU the performance issues raise their head and it all comes crashing down... as I say the CPU core performance isn't really the issue here. It's like all the Bulldozer performance claims last time around, yeah it looked decent on paper, but real world didn't reflect it.

In fact the pipeline statement above is exactly what I was referencing, hopefully Zen nails it finally, the design briefs they've put out are extremely promising at least.

[DJ Warmonger]

I expect all this naural network stuff to shine in case of badly optimized or badly conditioned programs where memory access is the issue. For processing-heavy tasks, as Kaine said, it will make no difference. All that matters here is a number of instructions performed per second, which comes directly from core clock.

[Debutante]

WOOT WOOT !!!!!?

http://wccftech.com/amd-launching-ryzen-cpu-gdc-2017/

[DJ Warmonger]

WOOT WOOT !!!!!?

http://wccftech.com/amd-launching-ryzen-cpu-gdc-2017/

End of February is still a long time to go , especially since I already have the money.

But maybe something like GeForce 1030 shows up until then, I need cheap GPU for my new PC.

[Burillo]

For processing-heavy tasks, as Kaine said, it will make no difference. All that matters here is a number of instructions performed per second, which comes directly from core clock.

umm no not really. if clock speed was all that mattered, then old P4's would've been faster than modern day Atoms. there's a lot more complexity to it than just clock speed.

[DJ Warmonger]

For processing-heavy tasks, as Kaine said, it will make no difference. All that matters here is a number of instructions performed per second, which comes directly from core clock.

umm no not really. if clock speed was all that mattered, then old P4's would've been faster than modern day Atoms. there's a lot more complexity to it than just clock speed.

I'm talking about changing the core speed of one and the same processor