| ASRock X48TurboTwins-WiFi Motherboard Review | ||
| by Eldonko | June 30, 2008 | ||
| Overclocking Results Overclocking Results Now for the section you have been waiting for, overclocking results!! The board is packed with features, uses the latest and greatest chipset, but all that means nothing for an enthusiast if the board does not overclock well. Well it turns out the X48TurboTwins-WiFi will live up to many user’s expectations other than a few minor BIOS issues. First the good news, we were able to achieve an overclock of 4203Mhz (about a 30% overclock) which was stable enough to run every test thrown at it. Sticking with good news, the maximum FSB benchable was 520, not an easy feat for many X48 boards. It is important to note however the max FSB that runs with a degree of stability is around 510, but still good nonetheless. Stability using high FSB is a great benefit as this will not bottleneck your max overclock. Testing memory independently, the board clocked quite well for both DDR2 and DDR3. DDR2 was tested briefly and the results for a 4GB kit were about on par with a DDR2 P35 board. However, this was just tested to ensure the hybrid memory feature was working as specified. What we will focus on in this review is DDR3 performance. Even with a budget kit ($120) of DDR3, 820Mhz (DDR1640) was achieved at timings of 7-7-7-20 with a command rate of 1T and 1.82v. Not bad at all, now it should be simple math to find which divider will give the best memory clocks at 467 FSB, which is what is needed for 4203Mhz on the CPU. Right? Well, not quite... First off, the board has issues posting on some dividers. The 1:2 divider worked well since the clock crossing procedure is far more effective in sync mode plus there is always a cycle available for transfer. The problem with using the 1:2 divider is the CPU speed is bottlenecked at about 3700mhz because of RAM limitation (max 820mhz). So POSTing is ok, but this is not an ideal divider to use with this kit of memory. It is always best to maximize CPU speed and then focus on memory. This divider would ne fine on memory that can run over 1000Mhz, but this is not cheap. Running the CPU @ 4203Mhz (467 FSB) on 5:6 divider is another option, however that only gives 560Mhz on RAM which is far from ideal. In addition, the board really has a tough time POSTing at this speed on 5:6 so it was necessary to clockgen up from about 435 FSB most of the time. This is likely because of one of two reasons. Either the BIOS is not engineered to allow proper posting of high FSB on this divider or secondary RAM timings are being auto adjusted too tight to allow the board to POST. Moving on, the 5:8 divider (533Mhz – DDR3 1066) would be best for a FSB of 467, it keeps memory within it’s stability range and would be the optimal configuration for the CPU overclock (467 x 1.6 = 747Mhz memory speed). The problem is the board will not POST past 380FSB on this divider. So unfortunately, this option is out. With FSB issues on 5:8 and the memory bottleneck on 1:2, we are just left with the 5:6 divider. This option is not THAT bad though, 4203Mhz and 560Mhz at 6-5-5-20 on RAM is a very respectable overclock. This configuration will be used throughout the benchmarking sections. | ||
| |
| Latest Reviews in Motherboards | |||||||||
|