View Full Version : Epox owners please help out
killem1x1
12-25-01, 09:10 PM
I posted this thread here (http://forums.overclockers.ws/vb/showthread.php?threadid=54074) but didn't get a response. If any Seti brothers can help, it'd be appreciated;)
Eliminator
12-25-01, 09:15 PM
I think the Bios has a divider that kicks in around 166 to bring your specs down.
LandShark
12-25-01, 10:35 PM
but i've heard that /5 or /6 divider is "only" from a beta BIOS which u have to get it from somewhere else other than Eprox's web site.
and according to my 8KHA+, it didn't let u manually set the divider. it seems like use whatever divider the BIOS's "pre-set" is from the factory depending on ur FSB speed. doesn't like the old KT133a which WILL show u not only the FSB, but also the PCI bus speed too.
btw, i'm soooooooo impressed that u "need" the /5 divider...!! my Corsair is only running at 153 right now, and my other Crucial is running at 150. however, i didn't try to push it tho.
does anyone think it would be safe to up another notch?? like 155? 160?? :rolleyes:
Eliminator
12-26-01, 09:20 AM
Originally posted by LandShark
does anyone think it would be safe to up another notch?? like 155? 160?? :rolleyes:
I don't see why not. The only harm that may come from raising your fsb is that it might not boot and you may have to clear your CMOS.
The divider on the 8K7A kicks in at 166. It is not a beta bios. The first bios to implement it was beta, but that was months ago. There's no manual way to enable or disable the divider. It is done by the PLL-IC clock generator. What you may be thinking of is a beta bios for the 8KHA+ that lets you enable or disable a divider for the ram. These higher dividers are seldom talked about because they're frowned upon by AMD. Epox will hardly acknowledge they exist. You might enjoy reading the posts in the Epox forum at AOA. They have a tech from Epox who answers questions. He posted the special bios for the 8KHA+ which kicks major butt.
LandShark
12-26-01, 12:50 PM
AOA?? what's stand for?? (sorry, not too familiar w/ it......) :rolleyes:
AOA forums = alliance for the advancement of the overclocking arts. That's the place Daniel started. I don't like to mention it here because of what happened with Daniel and other guys, but I guess it can't hurt. The Epox forum is interesting.
killem1x1
12-26-01, 07:40 PM
I was curious because it seems that I can "cross a point" where the system runs more stable at around 142. So I didn't know if there was a "divider" there? Also Sandra tells me my AGP bus is to high, as well as my System bus? Thanks for all the comments thogh! ;)
Originally posted by killem1x1
I was curious because it seems that I can "cross a point" where the system runs more stable at around 142. So I didn't know if there was a "divider" there? Also Sandra tells me my AGP bus is to high, as well as my System bus? Thanks for all the comments thogh! ;) The dividers are usually located at 33MHz intervals - hence 66,100,133,166,200 etc. Sandra hardly ever gets the dividers and bus speeds correct. It probably just looks at the fsb and thinks there's not a divider that would make that a safe speed. I doubt it's actually measuring directly from the pci bus.
Originally posted by LandShark
does anyone think it would be safe to up another notch?? like 155? 160?? :rolleyes:
There's a lot of controversy surrounding this issue from both sides. Not being an engineer myself, I can't get too deep in the discussion of the existence / non-existence of the 1/5 divisor. However, from my own experience, I can say that the Epox 8KHA+ will run very high FSB speeds.
I've read the 1/5 divisor kicks in at about 160MHz. My current 166MHz FSB along with the reliability of my LincSys NIC, would lead me to believe that there is another divisor.
I would suggest that you inch it up until you run into stability problems, try starting at 161MHz and go from there. Increasing the voltage to your RAM will help.
As stated above, you can also expect even higher with the newest BETA BIOS out now, some have reported as much as 200+MHz FSB.
Good Luck!
The divider kicks in at exactly 166. This was confirmed by an engineer at AOA who put a scope on his board and measured the drop in speed when changing from 165 to 166. The bios that allows 200+ for the 8KHA+ features a ram divider that helps reach those high speeds. On the 8K7A the divider doesn't slow the ram down, only the pci and agp busses, hence the reason it can't hit much beyond 180 unless you have really good ram.
Originally posted by TC
The divider kicks in at exactly 166. This was confirmed by an engineer at AOA who put a scope on his board and measured the drop in speed when changing from 165 to 166. The bios that allows 200+ for the 8KHA+ features a ram divider that helps reach those high speeds. On the 8K7A the divider doesn't slow the ram down, only the pci and agp busses, hence the reason it can't hit much beyond 180 unless you have really good ram.
I've read those threads as well. There are also others that have different thoughts on this issue with good arguments on their side to support these opinions.
As I stated in my earlier post, I'm no engineer. But I don't understand how my peripherals, particularly the NIC, are able to function properly at FSB's in the 160 to 165 range if the 1/5 doesn't come in until 166.
TC, don't take this the wrong way, I'm not trying to give you a hard time on this. I have a lot of respect for you and what you have accomplished. I consider you to be an "ICON" in these forums. However, I haven't heard any other plausible explanation for this. I'm curious as to your thoughts.
Thanks.
I didn't take your comments as argumentative. I was just repeating what I've read elsewhere. It makes sense to me that the divider would follow the same pattern as they have previously. The spec sheet for the specific PLL-IC used shows the divider taking over at 166 and 199. I have no explanation for why your stuff might be working at such high speeds if the divider does not come into play until 166. However I've seen some nics and other hardware that worked just fine at 180 on modified 440BX boards, as well as i820, and those did not have a divider above 133. I suppose I could get a friend of mine to spare some time with his scope and see what we can find.
Mictlan
12-27-01, 08:42 AM
I will be addding this hard fact. When using my current mobo (an Iwill BD100+) I wanted to know if I had other settings besides the ones shown in the BIOS. Normally when I was overclocking my mbo I had to use the settings put there by the BIOS programmer, like 100, 103, 105, 110, 115, 120, 124, 133, 140 and 150. I used a little program called SoftFSB. With it you need to feed the PCL data (it has several ways to do it, choosing the PCL manufacturer and model, the mobo or feeding the actual data with another program). And I came that at 120 MHz FSB you could use 2 settings: one using 1/3 divider and the other using 1/4 divider. The same thing happened at 124. If I selected any of this settings with the BIOS the one that gave me a lower PCI speed was always selected.
I would think that if the BIOS as made around the same logic , the PCI divider must be selected so it will give you a speed between 28 to 40 Mhz in the PCI bus. So at 161 with a 1/5 divider you'll have 32.2 MHz insted of 40.25. That woud be very esay to program into a chip.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.