|
Written by Maxit
|
|
Page 2 of 5
To give the new technology a good run down, we ran a series of tests in the MaXit labs to compare the results with those obtained
on an similarly priced and specified AM2 based system. The Intel and AMD chips were also taken up from their factory speeds to give an idea of the performance possible by overclocking. Stock cooling was used throughout. The Intel chip had plenty of room for further overclocking and went on to reach a prime95 stable speed of 2.86GHz  . Not bad for an entry level processor costing only £135!
Testbed:
Core 2 Duo system
- Intel Core 2 Duo E6300 ES @ 1.86 GHz and 2.5 GHz
- ASUS P5W DH Deluxe
motherboard
AM2 system
- AMD Athlon64 3800+ X2 @ 2.5
GHz
- ASUS M2N32 SLi Deluxe
motherboard
Common hardware
- 2 x 1 GB Team Xtreem PC2 5300
DDR2 RAM
- 256 MB Connect 3D ATi Radeon
X1800XT (Catalyst 6.6 Drivers)
- 80 GB Maxtor DiamondMax 10
Hard Drive
- Antec True Power 2.0 550W EPS
12v PSU
- Windows XP Professional SP2
Testing methodology:
We chose several synthetic tests to give us a general idea
of the performance trends which could be expected with most CPU and GPU
intensive tasks. In addition to these tests, we also made up a few real-world
tests in which we simulated a few scenarios that you would likely to be faced
with in day to day computing such as gaming and MP3 encoding.
We decided to raise the stakes a bit and decided to compare the E6300 running at its default clockspeed of 1.86 GHz with an AMD 3800+ X2 overclocked to 2.5 GHz. We felt it would be interesting to see how it would fair against the opposition which had a clock speed advantage of almost 700 MHz.
However to be fair, we also ran the benchmarks on the E6300 at 2.5 GHz to give a clock to clock comparison.
A summary of the tests chosen follows below:
Synthetic Benchmarks:
3D Mark 06
Futuremark's 3D Mark 06 is an excellent way to measure 3D performance of
your rig. The game tests are based on the same algorithm used in programming
real games, so you can be sure that if your PC does well in this benchmark, it
shouldn't have any problem coping in real games.
The CPU test in 3D Mark 06 is very useful in providing information on how you
will fair in games that are significantly dependant on processor performance.
3D Mark 2001 SE
Although considered obsolete by some people as a measure of GPU performance,
it still makes an excellent tool for measuring processor performance. The fact
that graphics cards are faster than ever means that this test is
predominantly CPU limited.
Super Pi
This benchmark is useful in comparing the floating point performance of different
CPUs. It times how long it takes to calculate the value of pi to a preset
number of digits which can be selected by the user. For our test, we used the
8M calculation which calculates the value to 8 million digits.
SiSoft SANDRA
SANDRA's CPU arithmetic and multimedia benchmarks are a good way to compare
the general performance of a variety of processors.
Real World Tests:
MP3 Encoding
In todays digital age, practically everyone owns an MP3 player. MP3
encoding is a fairly CPU intensive task and is benefited hugely by a faster
processor. To do our test, we ripped Moby - Play from the original CD to
MP3 files using Windows Media Player 10. We timed how long it took on each
system using a stopwatch.
Archiving/File Compression
It's fairly common these days for us to compress files and folders to share
or use for backup purposes. Like MP3 encoding, this is a CPU intensive task.
After ripping Moby - Play to MP3 files, we archived the folder (which
happened to be 200 MB) into a .rar file using WinRAR 3.6. Again, the process
was timed on each rig using a stopwatch.
Doom 3 and F.E.A.R. v1.02
To compare how the different rigs would compare in games, we ran
the built in time demos in Doom 3 and F.E.A.R. and recorded the number of
frames per second in each case.
Both Doom 3 and F.E.A.R. were ran at medium quality at a resolution of 800 x
600, to ensure that the CPUs were the bottleneck.
|
