Test Rig, Performance and Overclocking
First up, a disclaimer: Overclocking can cause hardware death, instability, data loss, time loss, explosions, smoke, fire. It also voids the warranty on all parts involved and is generally a dubious idea.
The following hardware was used during this portion of testing:
| Motherboards: |
ASRock FM2A85X Extreme6 Gigabyte FM2A85X-UP4 |
| CPU: | AMD Trinity A10 5800k 3.8ghz APU , 4.2ghz boost w/Radeon 7760D IGP |
| Ram: |
Corsair Dominator Platnium 2666mhz 11-13-13-35 2t Ran at 2400mhz @ 11-13-13-35 2t |
| GPU: | Integrated Into APU – Radeon 7660D IGP |
| Storage: | Western Digitial Blue 320gb SATA2 2.5″ |
| PSU: | Enermax Revolution85+ 1050watt |
| Cooler: |
Thermaltake Water2.0 Performer w/ Nidec BetaV 220cfm fan |
Since we didn’t have anything else to compare this board to on the site since desktop APU’s are just really hitting the market. I Grabbed one of the boards currently on my bench slated for death by overclocking to use as a comparison. The board in question is the Gigabyte FM2A85X-UP4.
The same board AMD released with their press kits. So we are going to use this as our baseline comparison. All Stock benchmarks were run three times, then averaged together. At this point in the review I was running out of time to get this done so the OC results at this point are not an average but the result of a single stable run.
This is mostly due to the amount of hurdles I ran into in the bug department while attempting to get overclocks that worked correctly.
After consulting and several bug reports and later recived fixes from ASRock I was able to complete the 3D portion of the bench marks.
The stock clock benches where all ran with the fan set to Idle, approximately 5% positive duty. While overclocking it was opened up and let run at 70-100% duty based on core temps and load. Believe me airflow wasn’t a problem. I was also blessed with 5c ambient temperature while doing the 3D OC benching. Resulting it wonderfully low overall core temps.
CPU and other special settings used while testing:
| Multiplier | Voltage | Bclk | IGP Voltage | IGP Speed | Thermal Throttling | |
| CPU stock | 42x |
1.4v (Stock) |
100mhz (Stock) |
1.275v (stock) | 800mhz (stock) |
Enabled (stock) |
| CPU OC | 46x | 1.5v | 100mhz | 1.47v | 1267mhz | Disabled |
I used a 42x multiplier to remove flutter from turbo core up and down clocking. Since we where not running outside of the thermal envelope for stock testing thermal throtteling never kicked in during testing despite being enabled. Showing that we would of been running at a 42x multiplier anyways. This was done the same across both boards.
Why no cold benchmarks you might ask?
They will be featured in a second article along with a more in depth experience of overclocking with this board. Due in part to the original technical bugs and the additional work required to insure the board doesn’t instantly die due to condensation.
