Gigabyte GA-P35T-DQ6: DDR3 comes a knocking, again
by Gary Key on May 30, 2007 2:00 AM EST- Posted in
- Motherboards
E6600 Dual Core Overclocking
We were easily able to reach a final benchmark stable setting of 9x420 FSB resulting in a clock speed of 3780MHz. We were able to run our Corsair DDR3 CM3X1024-1333C9DHX at timings of 8-7-7-20 with a minimal increase in voltage to 1.55V with 4GB or 2GB of memory. We did encounter one problem and that was trying to run our memory past 1400MHz while the FSB was above 400 with the E6600. The board would boot and enter Vista up to 1600MHz but we could not complete our entire test suite unless we kept the memory below 1400MHz. We attribute this to an early BIOS design that favors tighter memory timings and latencies over high memory overclocks. Vdroop was very acceptable on this board during overclocking with an average drop of .02 ~.03V during load testing with our E6300, E6600, and X6800 CPUs.
We dropped the multiplier on our E6600 to six and were able to reach 550 FSB without an issue on the F2N BIOS. We were not able to enter Vista at 6x551 up to 6x600 although the board would POST. We managed to eke out a 6x592 result on the DDR2 version of this board so we attribute the seemingly hard lock at 550FSB to the BIOS at this time. The issue at this time with the high FSB settings is that latencies and access times suffer at the expense of memory bandwidth.
We are still working on finding the best memory multipliers and timings to utilize along with exact chipset strap changes. They tend to float between BIOS releases but we will provide a full range of results and settings in a future article. Gigabyte still has some fine tuning left to complete but overall the BIOS is in really good shape from both a performance and stability standpoint. Once the board is dialed in, stability is absolutely superb.
| Gigabyte P35T-DQ6 Dual Core Overclocking Testbed |
|
| Processor: | Intel Core 2 Duo E6600 Dual Core, 2.4GHz, 4MB Unified Cache 1066FSB, 9x Multiplier |
| CPU Voltage: | 1.4875V / 1.4500 (default 1.3250V) |
| Cooling: | Tuniq 120 Air Cooling |
| Power Supply: | OCZ ProXStream 1000W |
| Memory: | Corsair DDR3 CM3X1024-1333C9DHX (2x1GB, 4x1GB) 8-7-7-20 |
| Video Cards: | 1 x MSI HD 2900XT |
| Hard Drive: | Western Digital 150GB 10,000RPM SATA 16MB Buffer |
| Case: | Cooler Master CM Stacker 830 |
| Maximum CPU OC: | 420x9 (8-7-7-20, 1344MHz, 1.55V), CPU 1.4875V 3780MHz (+58%) |
| Maximum FSB OC: | 550x6 (8-7-7-20, 1320MHz, 1.55V), CPU 1.4500V 3300MHz (+107% FSB) |
| . | |
We were easily able to reach a final benchmark stable setting of 9x420 FSB resulting in a clock speed of 3780MHz. We were able to run our Corsair DDR3 CM3X1024-1333C9DHX at timings of 8-7-7-20 with a minimal increase in voltage to 1.55V with 4GB or 2GB of memory. We did encounter one problem and that was trying to run our memory past 1400MHz while the FSB was above 400 with the E6600. The board would boot and enter Vista up to 1600MHz but we could not complete our entire test suite unless we kept the memory below 1400MHz. We attribute this to an early BIOS design that favors tighter memory timings and latencies over high memory overclocks. Vdroop was very acceptable on this board during overclocking with an average drop of .02 ~.03V during load testing with our E6300, E6600, and X6800 CPUs.
We dropped the multiplier on our E6600 to six and were able to reach 550 FSB without an issue on the F2N BIOS. We were not able to enter Vista at 6x551 up to 6x600 although the board would POST. We managed to eke out a 6x592 result on the DDR2 version of this board so we attribute the seemingly hard lock at 550FSB to the BIOS at this time. The issue at this time with the high FSB settings is that latencies and access times suffer at the expense of memory bandwidth.
We are still working on finding the best memory multipliers and timings to utilize along with exact chipset strap changes. They tend to float between BIOS releases but we will provide a full range of results and settings in a future article. Gigabyte still has some fine tuning left to complete but overall the BIOS is in really good shape from both a performance and stability standpoint. Once the board is dialed in, stability is absolutely superb.
Facebook
Twitter
RSS
22 Comments
View All Comments
yzkbug - Wednesday, May 30, 2007 - link
Will we ever see boards supporting both DDR2 and DDR3 memory? It would be nice to be able to run DDR2 for now, and switch to DDR3 in the future without buying a new mobo.Stele - Friday, June 1, 2007 - link
There's already at least one in existence - the Asus P5KC. Check it out http://www.asus.com/products.aspx?l1=3&l2=11&a...">here.It would be interesting if Anandtech could get hold of this board and see if having support for both memory types sacrifices fine tuning and hence performance/overclocking capability by a measurable degree.
slayerized - Wednesday, May 30, 2007 - link
I know it is a bit premature, but do you have n estimate on the targeted price points for these boards and ddr3 memory modules?gigahertz20 - Wednesday, May 30, 2007 - link
You can already buy the Asus P5K Deluxe for $225 from here.http://www.xpcgear.com/p5kdeluxe.html">http://www.xpcgear.com/p5kdeluxe.html
My guess is once Newegg and some other places get them in hopefully around $200 or below but maybe not. They will be expensive at first.
xsilver - Thursday, May 31, 2007 - link
are p35 boards recommended for midrange overclocking systems just yet?a gigabyte ds3 + e6320 vs. a asus p5k + e4400 combo; which system is likly to have better performance after OC?
Sunrise089 - Wednesday, May 30, 2007 - link
The above comments reminded me of something many reviews have said recently - that "additional airflow needed to OC" line. What exactly does that mean? Does it simply mean airflow inside the case, as in you first tested with so case fans at all, and had to add some? Or does it mean you added some sort of motherboard specific additional cooling? If the latter, a motherboard that does not require such an added part would be much more appealing.Googer - Wednesday, May 30, 2007 - link
Good Article. But where are the disk and I/O benchmarks?Treripica - Wednesday, May 30, 2007 - link
What the hell is a niggle?johnsonx - Thursday, May 31, 2007 - link
a 'niggle' is a minor complaint, or perhaps a complaint about a small detail. It's completely unrelated to another word like that with an 'r' at the end.TallBill - Wednesday, May 30, 2007 - link
All applications were run with administer privileges.