Assembly And Overclocking
All three SBM builds housed within the Asgard II enclosure have gone just the way we like: smooth and without issue. This one was no exception. The assembly was easy and the finished product was impressively stable.
Snapping the front bezel off by pulling from the lower edge makes 5.25" drive installation easier. There are not enough tool-less drive clips to populate all drive bays, but the ones provided are adequate for retaining the system’s drives. If you really need them, screws work too.
Once again, the only real difficulty is threading the black standoffs into the painted motherboard tray. Not really a big deal, but it's worth noting that your fingers won't suffice. A 5 mm nut driver is the perfect tool for this job.
Though this should probably be pretty obvious, we weren't able to fit a P67- or Z68-based board into our budget. Both of those platforms would have allowed us to exploit the Core i5-2400's "limited overclocking" capability, which would have facilitated a certain number of 100 MHz bins over the chip's maximum Turbo Boost ratio. Intel’s multiplier cap of 38x would have been good for 3.8 GHz without the need for base clock tampering, which is already very limited.
MSI's firmware teased us with CPU multipliers up to 34x (the highest used by Turbo Boost). But on this H61 platform, the ratio is locked, and there was no way to set four-core frequencies in excess of 3.2 GHz. It's better to just let Turbo Boost do its thing, which by our observation means switching between 3.2 and 3.3 GHz, depending how many cores are active.
With no base clock adjustment, a locked ratio, and a maximum memory frequency of 1333 MT/s, our tweaking was limited to setting lower RAM timings. MSI only gave us four settings for memory voltage: 1.35 V, Auto (1.5 V), 1.65 V, and 1.8 V, but they all fell under those numbers somewhat. I’m not crazy about pushing above 1.6 V, but the 1.65 V setting yielded 1.632 V and allowed for stability at CAS 7.
This Sapphire Radeon HD 6870 had even more headroom than the last one. The GPU core was stable at AMD OverDrive’s maximum 1000 MHz setting, and the memory was stable though our maximum tested 1200 MHz (4800 MT/s). We backed these down a bit and ran our second set of data with the core at 990 MHz and 1180 MHz (4720 MT/s) memory. As you pore over the benchmark results, remember that this system benefits from a 30 MHz-faster GPU clock and newer Catalyst drivers, while the prior AMD Phenom II X4 enjoyed a 3.8 GHz overclock.