We performed all our overclocking experiments in the following testbed:
- Intel DX48BT2 mainboard, BIOS 1513, 1521 and 1554;
- Intel Core 2 Duo E8400 CPU (3.0GHz, 333MHz FSB, 6MB, Wolfdale, rev. C0);
- 2 x 1024MB OCZ PC3-14400 Platinum Series (OCZ3P18002GK);
- NVIDIA GeForce 8800 GTS 320MB graphics card;
- Seagate Barracuda 7200.10 HDD (ST3320620AS, 7200RPM, 16MB, SATA 320GB);
- Zalman CNPS9700 LED CPU cooler;
- Antec NeoPower HE 550 PSU (550W).
We faced some difficulty during system assembly: the voltage regulator capacitor contacts hanging off the reverse side of the PCB didn’t let us install the backplate for Zalman CNPS9700 LED CPU cooler. Luckily, it was a plastic backplate, so we simply shaved off one of its sides and it fit snuggly into the spot. For better chipset cooling we attached a 50-mm fan to the chipset North Bridge heatsink.
We received Intel DX48BT2 mainboard with the BIOS version 1513. Our Intel Core 2 Duo E8400 processor with its default x9 multiplier can overclock to 450-455MHz bus frequency. However, the board failed to boot at 450MHz FSB. To find out the maximum FSB frequency when the mainboard remains operational and stable, we lowered the processor clock frequency multiplier to the minimal value of x6 and tried to boot at 450MHz and then at 425MHz. Both these preliminary tests failed, so we reflashed the BIOS to the latest version 1521 and continued out vain attempts.
Intel DX48BT2 mainboard supports WatchDog Timer technology that monitors POST, and although it was enabled, it didn’t work well. The mainboard responded in several different ways to over-overclocking. For example, it could simply refuse to boot, which would be the best outcome, no matter how strange it may sound. The noise from the fans indicated that the mainboard was powered on, but didn’t start, so we had to stop wasting time and go ahead and correct the BISO settings.
It is much worse when the board powers on, quickly shuts down, then powers on again and continues booting. The thing is that Intel DX48BT2 mainboard, like Gigabyte boards, doesn’t report when the processor settings have been reset to defaults and continues to boot the OS. So, until you launch some diagnostic tools, you will not really know if your overclocking attempt succeeded or not. You simply waste the time on booting Windows and then rebooting the system. There is a variation of the latter situation, when instead of powering on and shutting down once the mainboard needs to do it at least five times to come to senses, but the worst thin when it loops at the on and off stage.
Once the starting FSB frequency was dropped to 400MHz, the board could finally boot, but still failed to load Windows. It succeeded only at 380MHz FSB, however, this poor overclocking result will hardly be of any value, so we didn’t perform any detailed performance tests at this speed. We haven’t come across FSB Hole for a long time already (when the board doesn’t work in a certain frequency interval, but functions normally at lower as well as higher speeds). Therefore, we tried booting at 450MHz as well as 475-480MHz FSB as well.
We increased processor Vcore, NB voltage, FSB voltage and Vmem; changed the bus frequency, processor clock multiplier and memory dividers; disabled power-saving technologies. However, if the operating system booted, it always meant that the mainboard reset all the parameters to defaults again. Just like with the previous Intel DX38BT mainboard, we didn’t manage to achieve any more or less acceptable overclocking results on Intel DX48BT2.
I am sure you may already imagine the conclusions we might want to make in the end of this article. However, while working on this review I checked out Intel web-site a few times and was very surprised to find out that there was a new BIOS version – 1554 - already available. Strangely enough it was dated May 1, although I never saw it there on May 12, when I downloaded the latest available BIOS version at that time – 1521.
Once I read the BIOS Update Release Notes, I realized that I absolutely had to check it out. They listed all those problems that I have just described among the corrected bugs: hanging on boot-up, continued booting in safe mode, boot-up with maximum clock multiplier and disabled EIST despite the setting in the BIOS, looping system restart. Actually, the list of made corrections is much longer: I was just lucky not to have experienced all of the issues. Besides, the new BIOS version allows disabling C1E, supports fractional processor clock frequency multipliers, and offers higher maximum supported FSB frequency of 720MHz instead of 500MHz.
As a result, we had to put together our Intel DX48BT2 based testbed one more time, update the BIOS with the newest version 1554 and attempt to boot at a modest but yet unattainable FSB speed of 425MHz. I have to say that the board changed dramatically. It started acting much more predictably. Now it responded adequately to over-overclocking: rebooted in safe mode, stopped at some point and offered the user to continue booting with the last operational settings or enter the BIOS Setup to make a few changes. I was offered this choice quite a few times, because I still failed to successfully overclock the CPU.