[quotemsg=17231525,0,440653]Why even purchase the older chipsets with prices like that?[/quotemsg]
If you bothered to read the rest of the SBM articles, you'd see that these parts were purchased back in November, not right now. Also, you'd see this quarter was a "if I would've known then what I know now" theme in how we would've built the machines from last quarter if we didn't have money restrictions. You'd also know how I restricted myself to purchasing parts that were only available three months ago since it was a "how I should have built it then" type thing going on. Please people, fully reading the articles answers 90% of the "why did you do this? questions.
Joe, the cooler fitting is more on Deepcool than Gigabyte. Nearly every ITX FM2 board has the CPU slot oriented to the RAM slots in the same way as here. No, my beef with the mboard is:
1) Poor RAM support. It took 1.65V to even run a standard 1600 CAS9 1.5V profile. This same RAM before easily made it to 2400, so 2133 is a little disappointing. I got seemingly random crashes on it. Example, I can run Prime95 in blend with 4096 MB of RAM used for four hours at 2133 without a problem. It can go through the rest of the bench suite, even the very RAM intensive 3DS sections, at that same speed with no problems. However, Arma couldn't run past 1866 ( why? ) and the Sandra memory bandwidth test could only run once ( run it again without rebooting and it crashes the computer ). It's inconsistent performance and that really bugs me.
2) No, I didn't like the UEFI layout much. It had a somewhat logical idea of grouping all the voltages into one area, all the frequencies into another, etc. Except you could find the same setting in a few different places. The RAM multiplier could be set in both the frequencies and RAM sections. The RAM voltage could be vied in the RAM section, but only changed in the voltage control section. That's a little confusing when you change it one place, encounter it in another, and it doesn't match what you just set it to.
3) It was limited in features. The adaptive VCore control was limited and confusing and you can't set a per-core multiplier ( load on one core goes up to 45, two cores is 44, three cores is 42, etc ).
Overall, I wouldn't recommend it. The ASRock board I used last time was okay in features, except RAM OCing was tedious in that every timing had to be manually set and the HD audio header is inexcusably located under the GPU. MSI had some promising boards, but they were all $100+ at time of purchase. If I had known this one would be so troublesome, I would've spent the extra to see what MSI offered.