|
|
|
|
Before you read on, I would like to thank Mr. Nikos Papanikolaou [email protected] in contributing this article on overclocking his ASUS P2B at 133Mhz FSB. Here is a list of my components and how they are configured: Intel PII-300 (non-boxed code SL2HA)
The system can run at 2.5x133=333MHz without crashes in both Windows 98 and Windows NT 4.0 Workstation. The only note that I have to make is that Monster 3D gives "tearing". One should say that this is caused due to the 44.3 PCI bus. However, this is not the case. I will explain why not: The ASUS P2B manual (rev.102) says that at 112
the bus is 112/3=37.3, at 133 is "Dear Customer, The PCI frequence is always kept in 33Mhz (for 440BX M/B) if the CPU external bus is over than 100Mhz, and it's the half of external CPU bus speed if less than 100Mhz. For example, if setting as 83Mhz, PCI is running as 83/2 = 41.5Mhz. Best regards, Woody Huang Technical Support Division" So as you can see ASUS did a great job keeping the PCI specs at default since very few PCI cards would be able to run at 44.3(133/3).
Problems The first problem was to verify that the PCI clock remained at 33 when the bus speed is over the 100MHz . So: I removed all me cards and leave my SCSI controller and the VIPER (AGP) only and run at: 4x75: When the system enters windows 98 or NT 4 the SCSI CDROM makes weird seeking noises and the system hangs a little bit later with either a blue screen or a complete freeze. 3.5x83: At the boot screen where the SCSI initialises and lists the devices attached it will give me the message "CHIP FAILED" (i.e. the SCSI controller chip). The above clearly demonstrate that my SCSI controller does not work with neither 37.5 nor 41.5 bus speeds. The second problem was to verify the reason for the Monster giving tearing. So I removed the SCSI and insert the Monster 3D. I can run in the two above configurations (i.e. 37.5 and 41.5 PCI clock) without any problems. This certifies that the Monster 3D handles even the 41.5 PCI clock speed. The only thing that I can think about the tearing phenomenon in the Monster 3D is that it can't handle the increased speed of the memory..... the VIPER on the other hand does not have any problems whatsoever at 2.5x133… Since at all other configurations (i.e. 4.5x66, 3x100, 3x112, 2.5x133) the SCSI works flawesly, this proves that at 112 and 133 bus speeds the PCI clock in the ASUS P2B is kept lower than 37.5 or 41.5, and of course this confirms the reply from ASUS Taiwan that at 133MHz bus the PCI clock remains within specs at 33 but I believe that at 112MHz bus the PCI clock is at 37.3MHz. Since I do not believe that you can use a divider of 3.36 in the BX chipset... to get the 33 PCI clock at 112 bus!.
Recommendations In my opinion, the fact that I am now running at the 133 bus speed, is due to two specific parts that make this happen. I list them in order of importance:
Concluding I would firstly recommend everyone who wants to run at this bus speed to search before he buys memory, since it is one of the most crucial factors. The way to do this is to ask your memory supplier about the alphanumeric code that is printed on the chip, and then go directly to the manufacturers page and see the specs for yourself. Secondly I would suggest to buy a motherboard that either keeps the PCI bus speed within specs when above 100MHz bus or has a way to control it. Thank you for sharing my experience and Celeron Overclocker’s Page for hosting my article. Nikos Papanikolaou Email: [email protected]
13 September 1998 11:34 AM |
