Hilfe
Neues Thema
Antworten
geschrieben 
Dabei lief das Dualkit durch folgende Tests;
SiSoft Sandra Memory Bandwidth
Everest Home Edition
3DMark2001SE
Far Cry
Unreal Tournament 2004
PCMark2002
Super PI
PiFast
Zitat
Test System:
* AMD Athlon 64 3000+ Venice
* DFI LANParty UT nF4 Ultra-D
* 2 x 512MB OCZ Gold PC3500 GX
* Sapphire Radeon X800
* Arctic Freezer 64
* Thermaltake PurePower 680W
Overclocking:
Keep in mind that during overclocking, the latencies are as follows: CAS-TRCD-TRP-TRAS
Throughout all tests, the multiplier has been kept at 9x. BIOS is flashed to the 510-2 and yellow DIMM slots were used.
One thing to keep in mind with the PC3500 GX is that it is not designed to be an awesome overclocker. These are speed binned to run at 217MHZ 2-2-2-5 with just 2.8v. Not all Winbond UTT can run that speed and timings with just 2.8v. I'm sure that all you hardware enthusiasts are very familiar with the process of speed binning. There are two different kinds of speed binning. The first and most well known is the process of finding the fastest memory frequency possible, i.e., the OCZ Platinum PC5000. The other kind is looking for the lowest latencies at the lowest voltage requirement. The Gold PC3500 GX is the result of the second type of speed binning process.
Since Winbond UTT IC's are used, the memory latencies during the overclocking process will be kept at 2-2-2-5. I began the initial test at 200HTT and set out to find the lowest voltage needed. I was pleasantly surprised that the memory only needed 2.5v for standard PC3200 operations. The next stop was to find the voltage requirement for the rated speeds and timings. 2.8v turned out to be the required voltage. I kept the voltage at 2.8v and proceeded to find the maximum overclock that the memory would allow. It only pushed a measly 3MHZ more to 220HTT.
Because we know that the PC3500 GX uses Winbond UTT IC's and that the IC's love voltage, it's time to really push things with increased voltages. With 3.0v, I managed to obtain an extra 15MHZ out of the memory, stopping at 235HTT. I upped the voltage a bit more to 3.2v and only gained 5MHZ to reach 240HTT. I was quite surprised that pumping 0.2v more would only yield a mere 5MHZ. Going even further to 3.4v allowed the memory to shine again. It reached 255HTT without a hiccup and allowed smooth operation of all benchmarks and memory stress testing. Since the DFI nF4 Ultra-D allows up to 4.0v, I pushed even more to 3.5v and received another 5MHZ, reaching the ceiling at 260HTT. Increasing voltages did not provide any additional help.
Overclocking Results:
* 200HTT 2.0-2-2-5 1T 2.5v
* 217HTT 2.0-2-2-5 1T 2.8v
* 220HTT 2.0-2-2-5 1T 2.8v
* 235HTT 2.0-2-2-5 1T 3.0v
* 240HTT 2.0-2-2-5 1T 3.2v
* 255HTT 2.0-2-2-5 1T 3.4v
* 260HTT 2.0-2-2-5 1T 3.5v
* AMD Athlon 64 3000+ Venice
* DFI LANParty UT nF4 Ultra-D
* 2 x 512MB OCZ Gold PC3500 GX
* Sapphire Radeon X800
* Arctic Freezer 64
* Thermaltake PurePower 680W
Overclocking:
Keep in mind that during overclocking, the latencies are as follows: CAS-TRCD-TRP-TRAS
Throughout all tests, the multiplier has been kept at 9x. BIOS is flashed to the 510-2 and yellow DIMM slots were used.
One thing to keep in mind with the PC3500 GX is that it is not designed to be an awesome overclocker. These are speed binned to run at 217MHZ 2-2-2-5 with just 2.8v. Not all Winbond UTT can run that speed and timings with just 2.8v. I'm sure that all you hardware enthusiasts are very familiar with the process of speed binning. There are two different kinds of speed binning. The first and most well known is the process of finding the fastest memory frequency possible, i.e., the OCZ Platinum PC5000. The other kind is looking for the lowest latencies at the lowest voltage requirement. The Gold PC3500 GX is the result of the second type of speed binning process.
Since Winbond UTT IC's are used, the memory latencies during the overclocking process will be kept at 2-2-2-5. I began the initial test at 200HTT and set out to find the lowest voltage needed. I was pleasantly surprised that the memory only needed 2.5v for standard PC3200 operations. The next stop was to find the voltage requirement for the rated speeds and timings. 2.8v turned out to be the required voltage. I kept the voltage at 2.8v and proceeded to find the maximum overclock that the memory would allow. It only pushed a measly 3MHZ more to 220HTT.
Because we know that the PC3500 GX uses Winbond UTT IC's and that the IC's love voltage, it's time to really push things with increased voltages. With 3.0v, I managed to obtain an extra 15MHZ out of the memory, stopping at 235HTT. I upped the voltage a bit more to 3.2v and only gained 5MHZ to reach 240HTT. I was quite surprised that pumping 0.2v more would only yield a mere 5MHZ. Going even further to 3.4v allowed the memory to shine again. It reached 255HTT without a hiccup and allowed smooth operation of all benchmarks and memory stress testing. Since the DFI nF4 Ultra-D allows up to 4.0v, I pushed even more to 3.5v and received another 5MHZ, reaching the ceiling at 260HTT. Increasing voltages did not provide any additional help.
Overclocking Results:
* 200HTT 2.0-2-2-5 1T 2.5v
* 217HTT 2.0-2-2-5 1T 2.8v
* 220HTT 2.0-2-2-5 1T 2.8v
* 235HTT 2.0-2-2-5 1T 3.0v
* 240HTT 2.0-2-2-5 1T 3.2v
* 255HTT 2.0-2-2-5 1T 3.4v
* 260HTT 2.0-2-2-5 1T 3.5v
EDIT: bleibt nur noch zu sagen das ich Sie mir auf jedenfall nächsten Monat kaufen werde
Dieser Beitrag wurde von ´quark oO'2 bearbeitet: 23. Juli 2005 - 15:23
Nach oben
