Shogun 2 CPU Benchmarking (formerly "3570k CPU Bottleneck?" joke title)

[Master&Puppet]

Ah I see what you mean now:

Oc'ing the thing to 4g+ I'd not dream of being the equiv or better than adding another 7950.

Yea that makes sense in terms of potential (but you'd still need to OC the CPU to get the use out of the second card).

 

[Rastalovich]

Think that's more relevant with AMD drivers also. Believe someone somewhere compared drivers and cpu use.

 

[Master&Puppet]

Ah well that's just another complication to add to the list!

 

[Rastalovich]

There's far too many variables in the computer & game stakes to hit the nail exactly and precisely on the head, covering every-single base that exists. Too many different components and variables to do it - even on a corporate level.

It just can't be done.

You're better off sticking with the idea you had in mind and follow it through like you're doing already. It'll always be the case that as you go along with something you'll think "dammit I should have done xyz..." but you're already halfway through your tests - better of keeping them in mind and do them the next time instead of starting again.

 

[Master&Puppet]

Yea - after all I only have a certain amount of hardware anyway. Would love to get hold of a CPU monitor which displays during game like afterburner does for GPU load. Do you know of one?

Update, found out how to do it - and log all system details too!

 

[Master&Puppet]

Sexy Update. Logger reported in graph

A note of formatting

There were a lot of coloured lines when excel made the graph so I simplified it so that all cores are the same light blue colour with the Total CPU use in black and obviously 2 separate colours for each GPU. Does that look ok? I can re-colour the lines of people want to see which actual core is working.

Unfortunately the logger doesn't have a toggle key to start so I had to run it from immediately before I started the 1080p Benchmark. It therefore logged the use during the loading time which I could have deleted but I thought it could be potentially interesting. The actual benchmark obviously starts when the GPUs flare up halfway through.

I can literally report any sensor now...temps, volts, clocks, RAM use. Is there anything people want?

This gem of a find will change the way I report these benchmarks...I'll overlay the FPS onto these logs if it doesn't get too cluttered.

Interesting to see that the CPU was largely working off one core even when both GPUs were working their asses off.

M&P