A run down of the performance benefits of overclocking an Asus P5E3 Deluxe, Intel Pentium Dual Core E2160 CPU and OCZ Gold DDR3 combination including benchmarks and graphs.

Well I’ve promised this data for a while now and I thought I’d better deliver, so here it is at last. First off is a screenshot of CPUz showing the overclock in all it’s glory, for those sceptics who might not believe it possible without proof.

Before I get into the benchmark results I’d better give a listing of my system components as these have an effect on performance and overclockability.

• Motherboard: Asus P5E3 Deluxe
• CPU: Intel Pentium Dual Core E2160
• Memory: OCZ Gold DDR3 OCZ3G10662GK
• Graphics Card: GeCube 512Mb X1950pro
• Hard Drive: Samsung HD250HJ 250GB SATA2
• Operating System: Microsoft Windows Vista Home Premium 64-bit
• Power Supply: Hiper Type-R 580W

First off I always start with 1M SuperPI because it’s quick and easy to run and gives a rough guide to raw cpu power.

As you can see from the graph the overclocked system manages the test in almost half the time (54%) which is a good increase as expected.

My next benchmarking test is the memory benchmark in Lavalys Everest which gives us an idea of raw memory speeds, I’ve only reported the first three results (read, write and copy). If anyone is particularly interested in the others post a comment and I’ll let you know.

This benchmark suprised me a little as I wasn’t expecting such a massive increase, after all the memory is running at loosened timings and the clock speed is only 38% faster yet the benchmarks come out 95%, 84% and 87% better. This clearly shows that the increased fsb speed and the incresed cpu speed affect the memory performance massively.

These benchmarks are quick to run and provide a rough estimate of computing power but they aren’t real life tasks by a long way. So I’ve also run the following benchmarks that mimick real life task much more closely. The first one is a benchmark suite provided by CustomPC magazine and is a media benchmark suite that tests image editing, video encoding and multitasking (by playing the previously endoded video and archiving a huge number of files).

Custom PC’s test rig based on an E6750 scores 1000 in every test, so score near to 1000 are good and anything over 1000 is excellent for a low end cpu like this. Prior to overclocking the results are those of a modern budget PC, fine but nothing special. Overclocked, the results are very respectable, sure they don’t come anywhere near those of a quad core extreme edition but then I wouldn’t expect that! What is interesting is that it performs remarkably well in the video editing test but much more poorly in the multitasking even though it’s a dual core processor, compared to the E6750. I believe this is due to the smaller L2 cache that the E2160 has compared to a fully fledged E6xxx series cpu, which of course cannot be improved in anyway by overclocking or other means!

As for percentage increases, they are quite similar to the previous tests, 75% for the image editing, 78% for the video encoding, 97% for the multitasking and 83% in the benchmark overall. So although the overclocked E2160 doesn’t quite match a stock E6750, it does beat it in one test and it once again proves to be well worth overclocking.

My final benchmark test is the famous 3DMark from Futuremark, I’ve chosen to run both the 3DMark 05 and 06 versions in their default settings as you would run in the free versions. These are predominantly graphics tests but there’s also a cpu test which will offer a good idea of speed increase. It’s also interesting to see how much the cpu aids the graphics performance.

As you can see from these the increase on the cpu tests are of a similar order to those seen in the other benchmarks, 87% for the 3dmark05 cpu test and 86% for the 3dmark06 cpu test, but what I think is interesting is that the graphics tests get a decent boost too, 24% in 3dmark05 and 14% in 3dmark06. This additional boost to the graphics could equate to the difference between a choppy 25fps and a playable 30fps in some games, obviously depending on which graphics card is fitted.

If I average the percentage increases from the tests, ignoring the custompc overall percentage and the graphics card limited 3dmark tests then I get an overall percentage increase of 83%. So my conclusion basically is that if you don’t have a huge budget to blow on a higher end cpu then you can certainly make an E2160 go a long way. Sure, the majority of these chips will run at stock speeds in budget PCs made by OEMs and sold through high street chains but there they’ll do their job at a stroll, after all the average PC user just wants internet and email and a few word documents. However some will fall into the hands of overclockers like myself who will give them a real run out and allow them to stretch their legs and they will thank us by performing admirably.

Popularity: 1% [?]