ECS PF5 Extreme Review 3

ECS PF5 Extreme Review

Value & Conclusion »

Overclocking

Overclocking is a pain with this board. Every time the board doesn't complete the POST the CMOS checksum will be invalid, which means you have to redo all settings in the BIOS. This also happens sometimes when the board crashes in Windows, for example during a stability test.

The next issue is that manual memory timings are not possible, because of a BIOS bug, which is fixed now. Let's hope ECS gets the update out fast.

Another issue I had was with the BIOS flashing. Every time I updated the BIOS using the flashing utility in DOS, I would get "CMOS checksum failure" on every reboot after the update, not only the first one. This basically left me with a non-booting motherboard: CMOS checksum failure, hit F1 to enter BIOS. F1. Save. Reboot. CMOS checksum failure, hit F1 to enter BIOS, and so on ...
The only way to get a working system was using the ECS Top Hat Flash and use the BIOS flashed on it to boot. Then remove the Top Hat Flash and flash the original BIOS back.
After getting the first board replaced by ECS, and seeing the same issue again, I found out that you MUST use the Windows flashing tool to update your BIOS. The DOS tool does not seem to write to all areas.

Next issue: the TRCD and TRP settings are not correctly applied. If you set 4-3-3-8, it becomes 4-4-4-8. 5-3-3-9 becomes 5-4-4-9. This can be changed with SysTool in Windows though. I think the BIOS always uses TRCD and TRP from the memory's SPD info.

Last but not least, the i945 overclocking lock is in effect. The chipset compares the PCI-Express clock and the FSB clock. If the difference is too high, the system will not start (more details in our ABIT AL8 review). In order to get around this, the "PCIE CLK Synchronize" setting in the BIOS exists. When it is enabled, any increase to the FSB will also increase the PCI-E clock a bit. A distinct setting where you can freely pick the frequency would be better (the clock generator does support it). The problem with the linked setting of the PF5 is that you can't just indefinitely increase the PCI-E frequency. At some point the SATA drives are no longer detected, and a bit further the video card does not initialize right.

With all these issues and gotchas I could still get the board to run at 241 FSB = 3618 MHz. One MHz more and the SATA drives were gone.




With the BIOS set to 241 FSB booting into Windows works fine and the system is stable all day. Next I used SysTool to reprogram the motherboard's clock generator, which let me use frequencies up to 293 FSB = 4395 MHz. Any higher setting would not apply to the clock generator and the clock would not change.




The benchmarks were run at 3618 MHz and at 4395 MHz as well. Some benchmark results do not seem to improve because changing the clock within Windows affects of how some benchmarking program measure time.
Next Page »Value & Conclusion
View as single page
Nov 19th, 2026 18:26 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts