Overclocking and CPU Degradation

To check out the above discussed features and potential of MSI X48 Platinum mainboard we used an open testbed built with the following components:

• Mainboard: MSI X48 Platinum (MS-7353), ver. 1.0, P.2B4 BIOS;
• Processors:
  • Intel Core 2 Duo E6300 (1.86GHz, 266MHz FSB, 2MB, Conroe-2M, rev. B2);
  • Intel Core 2 Extreme QX6700 (2.66GHz, 266MHz FSB, 8MB, Kentsfield rev. B3);
  • Intel Core 2 Duo E8400 (3.0GHz, 333MHz FSB, 6MB, Wolfdale, rev. C0);
• Memory: 2x1024MB OCZ PC3-14400 Platinum Series (OCZ3P18002GK);
• Graphics card: NVIDIA GeForce 8800 GTS 320MB;
• HDD: Seagate Barracuda 7200.10 (ST3320620AS), 7200 RPM, 16MB, SATA 320GB;
• CPU cooler: Zalman CNPS9700 LED;
• PSU: Antec NeoPower HE 550 (550W).

For our preliminary tests of MSI X48 Platinum mainboard we used an old Intel Core 2 Duo E6300 processor. By raising its Vcore to 1.45V it can overclock from the nominal FSB frequency of 266MHz to 490MHz FSB, which is 3.43GHz frequency. This tests the board passed flawlessly, it only set very strange timings for some reason, which we had to correct manually afterwards and the Performance Level was different for each channel, but this is something MSI X48 Platinum mainboard cannot really fix.

In our previous articles we have already mentioned that for our tests we also use Intel Core Extreme QX6700 processor. It overclocks to 3.4GHz with Vcore increased to 1.45V and thanks to its “hot temper” serves to load the voltage regulator circuitry and cooling systems during our test sessions. Unfortunately, the Circu-Pipe cooling of the MSI X48 Platinum mainboard failed to cope with this task successfully. The CPU overclocked just fine, but the heatsinks warmed up excessively. We had to install an additional 50x50x10mm fan, because the standard 25mm fan would not fit between the chipset North Bridge heatsink and the installed graphics card. However, this measure didn’t really help that much: the heatsinks remained scorching hot.

It could be insufficient chipset cooling that in the end affected the performance of MSI X48 Platinum mainboard. The mainboard BIOS formally allows increasing the chipset NB Voltage to 1.83V that tells a lot on the platform stability during CPU and memory overclocking. However, in reality we had to stick to maximum 1.5V. Variable adjustment increment for this voltage setting allows setting it to 1.46V, then to 1.5V and then to 1.58V. While 1.5V was not enough to ensure desired stability and the platform reported errors during the test session, the next available setting of 1.58V produced a completely opposite result: instead of improved stability the errors started to pop up even sooner, probably because of extreme overheating.

Intel Core 2 Duo E6300 and Intel Core 2 Extreme QX6700 processors were used only for the preliminary tests of the MSI X48 Platinum mainboard. The main work had to be done by Intel Core 2 Duo E8400. As far as we know, this processor can overclock beyond 4.2GHz on abit IP35 Pro and ASUS Maximus Extreme mainboards. This result can be achieved by pushing the FSB frequency to 470MHz with the default x9 clock frequency multiplier, or to 530MHz with a lower x8 clock frequency multiplier. The processor Vcore should be raised to 1.55-1.6V.

The first attempt to work at 470MHz FSB with 1.55V Vcore setting failed: the board would boot but couldn’t load the Windows OS and displayed the BSOD. When we increased Vcore to 1.6V we could see the desktop, but BSOD would inevitably follow after the first seconds of Prime95 test. Lowering the bus frequency to 460MHz didn’t improve the situation that much: BSOD would now appear after a few minutes into the test. We tried playing with the memory timings and increasing a few other voltages. It did help and the system worked stably for some time at 460MHz FSB, but we had to interrupt the test when the processor temperature reached 76-78ºC. It is way too much even for an open testbed. Maybe the CPU was additionally warmed up by the hot heatsink on top of the chipset North Bridge located nearby.

We decided to lock the CPU Vcore at 1.55V and check out the results in this case. At 455MHz FSB the system ran for about 5 minutes, however at 450MHz FSB we ran Prime95 test for half an hour and the processor temperature didn’t exceed 72ºC. Some of you might be pretty frustrated with these results and could probably give up at this point blaming MSI X48 Platinum for poor overclocking potential, but not us. We decided to check if our Intel Core 2 Duo E8400 processor still overclock as good as it used to on other mainboards, before making any conclusions about our today’s hero.

On ASUS Maximus Extreme we set the same parameters as before when it overclocked our Intel Core 2 Duo E8400 CPU to 470MHz FSB. It was especially easy to do since all our BIOS settings were saved. So, we start the system at 470MHz bus frequency, boot the Windows, launch Prime95 and in a few seconds get the familiar BSOD. We drop the FSB to 465MHz – the result is the same. Then we check out 460MHz, increase memory timings, but cannot reach any stability whatsoever. It was a shocking discovery, as nothing changed and just recently the system worked flawlessly with the CPU overclocked to 4.23GHz! We had to admit that the overclocking potential of our Intel Core 2 Duo E8400 processor worsened, degraded, and it happened not in a year and even not in a few months, but after just a few weeks! So what happened?

I am sure you remember our recent review of MSI P35 Platinum Combo mainboard. As we reported, there was a software error that made the system send over 1.8V to Intel Core 2 Duo E8400 processor although the setting in the BIOS was of 1.55V or 1.6V only. Extremely high processor temperature indicated that there was a problem and the MSI Dual Core Center sidebar that was responsible for it was immediately removed from the system, even before we launched any tests. However, looks like this particular incident has gravely affected the overclocking potential of our Intel Core 2 Duo E8400 processor.

Frankly speaking, we are not 100% sure that this is how it all happened, maybe long-term tests at 1.55-1.6V voltage eventually let to this effect, as it is still relatively high core voltage for a CPU with the nominal setting of only 1.225V. However we first tested Intel Core 2 Duo E8400 processor on ASUS Maximus Extreme mainboard where it worked at 470MHz FSB with 1.55V Vcore. Then we continued the tests on abit IP35 Pro mainboard and the same result could only be achieved at 1.6V Vcore. After that we returned to ASUS Maximus Extreme testing and didn’t observe any changes in the Intel Core 2 Duo E8400 overclocking potential: it was still running fine at 470MHz and 1.55V Vcore. And now, right after we completed our MSI P35 Platinum Combo tests, we notice significant changes. We tend to believe that it is the core voltage beyond 1.8V that caused it.

However, this is when another question pops up: how far can we now push our Intel Core 2 Duo E8400 processor? Instead of proceeding with our MSI X48 Platinum mainboard tests we had to go back to ASUS Maximus Extreme to investigate the new overclocking potential of our Intel Core 2 Duo E8400. I will not describe all our experiments in detail this time. All I have to say is that we managed to achieve acceptable stability only at 455MHz FSB. Despite lower frequency, we couldn’t reduce the Vcore setting, it remained at 1.55V. Of course, lower FSB speed required lower memory frequency, so we could also reduce the timings and Performance Level parameter. However, it could hardly make up for the significant worsening of our Intel Core 2 Duo E8400 overclocking potential.