Related Articles

Overclocking

Intel has recently put a serious crimp into overclocking by raising the FSB for its processors.

Given that even the best enthusiast boards have problems running at FSB's much greater than 450MHz (with perfect stability) raising the FSB means that enthusiasts buying a lower multiplier processor will no longer be able to overclock quite as highly as before - for example, a future E8200 with an 8x multiplier would be limited to 3.4GHz maximum on a board that was only usable up to 450MHz (1800MHz data rate).

The Core 2 Duo E8500 uses a 1333MHz (333MHz quad pumped) FSB with a 9.5x multiplier. Due to this choice, the E8500 is largely limited to newer boards that officially support a 1333MHz FSB - and the voltages required by the Penryn based core - and while some older boards with updated BIOSes may be able to run with the chip, realistically, you will probably want at least a 965P based chipset for this processor.

The 1333MHz FSB also means that overclocking is going to be limited by the highest stable FSB of the board - and most "enthusiast" boards don't really want to go beyond 450MHz.

The E8500 still performed like a champ on the X38 based board I tested it on; and I was able to reach a stable overclock of 4.275GHz while only air cooled.

I was able to get to the Windows desktop at 4.52GHz - however it was not stable, and I did not wish to crank Vcore high enough to stabilize it. The Intel data sheet shows 1.46V to be the maximum safe Vcore for the E8500, so I did not want to exceed that by much. 1.468V was enough to get the system stable for all our benchmarks at 4.275GHz.

In order to reach 4.275GHz, I had to do the following:

• used a Noctua NH-U 12 heatsink with two 12cm fans
• set Vcore to 1.468V
• set Vfsb to 1.45V
• set FSB to 450MHz
• set memory to 1800MHz
• set memory timing to 8-8-8--24, 2T
• set memory voltage to 1.9V (officially supported for these OCZ modules)

This chip is a great overclocker.

Power Consumption

When idle at 3.16GHz, the E8500 uses about four watts less than the E6750 at 2.66GHz, and when fully loaded, there is still about a four watt difference in favor of the E8500. Not bad given the 500MHz increase in speed!

It is quite interesting that at 3.8GHz the E8500 actually uses a smidgen more juice than the E6750 at 3.8GHz when they are both loaded - no doubt this is due to the increased throughput of the SSE2 units when executing rendering code.

Conclusion

Intel has a winner on its hands with the Core 2 Duo E8500.

Unfortunately the E8500 also signals the end of easy, crazy-high overclocking for Intel chips due to increased standard FSB speeds. Because even if you end up with a really great board that runs fine with a 500MHz FSB, the lower end E8200 with an 8x multiplier would max out at 4GHz - which as we can clearly see is far below the potential of the 45nm Wolfdale. When Intel switches to 1600MHz FSB for its mainstream parts, the maximum overclock achievable on a good board with a lower multiplier less expensive chip will be limited to about 12.5% which is not exactly earth shattering.

We should enjoy the 35%-50% overclocks that we can currently achieve with reasonably priced parts while we can - the E8500 may well be one of the last easily significantly overclockable mid range part from Intel for a while.

At the stock 3.16MHz it already provides excellent performance for office use, encoding and gaming; and given that it is also an excellent overclocker, it absolutely shines and screams when pushed to the limit.

Frankly, for a gamer, an E8500 is a better choice than a quad core QX9770 - and it will remain a better choice for a gamer until such time as the games are no longer GPU bound and are extremely multi-threaded - which will not happen for quite a while.

For lots more gaming benchmarks of the E8500, check out OverclockersClub's E8500 Review.