Assembly And Overclocking
There was plenty of room once again within the Antec Three Hundred case in which to work, and cable management was simple with the NeoPower’s modular design. However, we did run into one problem worth mentioning with this month’s component selection: the excessive height of the heat shields installed on the Patriot Viper RAM interfered with the Freezer 7 Pro’s fan assembly, preventing the first DDR-2 socket from being used. Thus, in order to run dual-channel mode, channel 0 was not an option and we needed to use channel 1 instead. This would prevent four sticks of this same memory from being used, thus making this motherboard/cooler/memory combination less than an ideal match.
Last Month, our E5200 didn’t like a high-clock FSB but was still a blast to overclock and Prime95 was stable at over 4.0 GHz for air cooling. Armed with the same motherboard and cooler, we had high hopes that this month the E7300 would reach similar clock speeds with the added bonus of more L2 cache.
Luckily, we did not encounter the same FSB limitation, which would have drastically limited our overclocking with the much lower multiplier of the E7300 and we were able to push beyond a 1,600 MHz bus speed if need be. But apart from this, our overclocking experience paled in comparison to that of last month with the E5200.
At just 3.6 GHz (400*9), we already needed almost 1.4 V Vcore for stability. After bumping up to 3.8 GHz, Prime95 would fail after 14-17 minutes until we raised Vcore all the way to 1.488 V. This was already more voltage than we intended to run, but with ambient room temperatures hovering near a chilly 18 degrees Celsius, maximum load temperatures were plenty safe at 61 degrees Celsius, so we decided to use these settings for testing.
With this disappointment, we pressed on for fun to try and reach 4.0 GHz. Vcore was cranked to 1.552 V, yet Prime95 still instantly failed on one core and crashed with any attempt to game or benchmark at 4.0 GHz. This is the same Vcore that allowed us to game and bench just fine at 4.3 GHz on the E5200. As we have said before, successful overclocking is, up to a point, the luck of the draw, but our E7300 already leaves us feeling that we won’t revisit this CPU again in future SBM reviews. Another glitch occurred at both 3.6 GHz and 3.8 GHz when both Prime95 and memory test 86+ stable settings failed to cold boot one time each.
The Sapphire HD 4850 we received this month again had no problem running CCC’s maximum manual settings of 700 MHz core/1,200 MHz (2,400 MHz effective DDR). Rather than stop there, we decided this month to try other means of overclocking. AMD GPU Clock Tool didn’t work with our card and drivers, but with Riva Tuner 2.21 we were able to push the Core to 750 MHz and the memory to an amazing 1,220 MHZ. For testing, we settled for a little more conservative 740/1195.