Test Setups
ABIT AN7 (UDP 4.27s), Athlon XP 2500 (provided by ), MSI GeForce FX5950 Ultra 256MB (Forceware Drivers 61.77), 80GB WD 8MB Cache HD, Windows XP SP2
ABIT IC7 Max3, Pentium 4 2.4C, MSI GeForce FX5950 Ultra 256MB, 80GB WD 8MB HD, Windows XP SP2
Overclocking
Overclocking was performed on two different systems, an AMD Athlon XP system and an Intel P4 2.4C.
Starting with the AMD system, the first thing I did was try to remove the bottleneck as much as possible, which entailed a search of ABIT's forums for a new BIOS. I know many enthusiasts will be aware of the TicTac BIOS, of which I am running a derivative by EnduraCell based on the v16 BIOS for the AN7. The particular BIOS is the CPC2T version. Anyway, I've previously never gotten higher than 219 FSB, so if you too have had this trouble, give one of the TicTac BIOS images a spin (though be aware you do so at your own risk).
The other thing I did was to lower the Multiplier down to the lowest bootable that I could attain which was 9.5. These efforts, plus the water-cooling from WACC, gave me high hopes and I wasn't disappointed. I was able to push the system to a stable 240 FSB, which whilst only at 2.28GHz is still quite high for a non Mobile Athlon.
Of course the ram here wouldn't have been the limiting factor since it is rated for 250MHz, so switching over to the ABIT IC7 P4 platform allowed us to find the limits of the ram which pushed all the way up to 274MHz. I'd like to be able to say that the ram might have had more in it, but I was able to boot into Windows and perform non intensive tasks at 276MHz, with 274MHz becoming stable. However, we shouldn't detract from the fact that 274MHz from ram that is rated at 250MHz is a very impressive overclock in its own right.
In both cases I've had to bump up the voltage from a default 2.7v to 2.9v to maintain stability and in both cases the ram handled the extra voltage increase without issue. I would like to have been able to have tested this ram at lower speeds and lower latencies, but the AN7 can be real picky with its memory, and wouldn't let me POST with the Ultra PC4000 at anything other than 3-4-4-8. This is in line with other PC4000 I've tried such as the Corsair TWINX kits.
Test software will be
Unreal Tournament 2004
Quake 3: Arena
All of the test software applications were run 3 times and the results averaged between the 3. Unfortunately, the 2.4C wasn't mine, and due to time constraints I wasn't able to finish the benchmarking on the Intel platform. I've presented what results I was able to get, along with comparisons with the Corsair PC4000, and I hope to be able to come back and finish the results at a later date.
SiSoft Sandra
Although a synthetic benchmark, it's a popular one, freely available if you wish to make comparison benchmarks. We will be testing the memory speeds, using the 32-bit 2004 version.
|
|
Ultra PC4000
|
Corsair TwinX 1024 PC4000
|
|
|
Float
|
Int
|
Float
|
Int
|
|
AN7 @ 9.5x240
|
2996
|
3280
|
3085
|
3320
|
|
IC7 Max3 @ 250
|
5610
|
5690
|
5634
|
5630
|
MB/s, Higher is better
SiSoft Sandra's results show the Ultra to be on par with the Corsair PC4000 TwinX in the bandwidth test. This isn't to much of a surprise as with both being PC4000 and both being tested at equal speeds the results should be similar. If they were different, that is when we would have something to talk about.
PiFast
A good indicator of CPU/Motherboard performance is PiFast version 4.2, by Xavier Gourdon. We used a computation of 10000000 digits of Pi, Chudnovsky method, 1024 K FFT, and no disk memory. Note that lower scores are better, and times are in seconds.
|
|
Ultra PC4000
|
Corsair TwinX 1024 PC4000
|
|
AN7 @ 9.5x240
|
53.31
|
53.32
|
|
IC7 Max3 @ 250
|
45.12
|
45.03
|
Time is Seconds, Lower is better
Again the Ultra shows that it is more than capable of keeping up with the big boys with very similar times to the Corsair.
CDEX
I took a CD (Bad Company - 10 from 9), and ripped it to one long 414MB .wav file. I then turned around and encoded that wave file to a 320Kb/s sample rate MP3. I am sure there is a negligible HD performance influence in this test but with the amount of memory we are testing, it shouldn't be more than a fraction of a second.
|
|
Ultra PC4000
|
Corsair TwinX 1024 PC4000
|
|
AN7 @ 9.5x240
|
2:01
|
2:01
|
|
IC7 Max3 @ 250
|
1:26
|
1:26
|
Time in Minutes, Lower is better
This kind of test is very CPU dependent, however the memory speed and quantity do serve to speed up the process quite a bit. Once again their is little difference between the Ultra and Corsair.
TMPGEnc MPEG Encoding
Video editing is a taxing chore, and we'll be testing the TWINX using TMPGEnc 2.512 to encode an Animatrix short of 150MB to a DVD quality mpg of 5000Kbps. Note that lower scores are better.
|
|
Ultra PC4000
|
Corsair TwinX 1024 PC4000
|
|
AN7 @ 9.5x240
|
14:31
|
14:32
|
Time in Minutes, Lower is better
In this test the Corsair loses out to the Ultra although the difference is negligible and well within the accepted margin for error.
NEXT
