-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Ncch-dl
- Thread starter diehrd
- Start date
- Joined
- May 6, 2003
- Location
- Maine
It has a PCI lock at all FSBs other than the 166-199 range, from what I've read here and at 2CPU.
Ken
Ken
- Thread Starter
- #3
I have never read anything that says for sure it is locked.
I can also say that at any higher then 200 no matter the multiplier I use I get serious instability..I can run 18x200 but not 14x210 ? or 16x200 but not 12x a any clock over 200 ? ( I have HyperX pc3500 in rig and have Corsair Pc3200)
Seams top me it is not locked that is why i am asking if any one knows for sure and or has a spec sheet saying it is ?
I can also say that at any higher then 200 no matter the multiplier I use I get serious instability..I can run 18x200 but not 14x210 ? or 16x200 but not 12x a any clock over 200 ? ( I have HyperX pc3500 in rig and have Corsair Pc3200)
Seams top me it is not locked that is why i am asking if any one knows for sure and or has a spec sheet saying it is ?
I dont' have a spec sheet or anything, but I have also read the same things penquiss has and would have to agree with him.
Also using clockgen you can adjust the fsb/memory separate from the AGP/Pci speeds so that further leads me to believe it is locked.
It does however seem odd, on the PC-DL I can boot and run non-cpu intensive tasks much higher than where the setup stops becoming stable. On the NCCH-dl it definately seems to hit a 'wall' much faster.
Also using clockgen you can adjust the fsb/memory separate from the AGP/Pci speeds so that further leads me to believe it is locked.
It does however seem odd, on the PC-DL I can boot and run non-cpu intensive tasks much higher than where the setup stops becoming stable. On the NCCH-dl it definately seems to hit a 'wall' much faster.
- Thread Starter
- #5
I used another program that had the NCCH-DL clock frequiencies listed and it clearly showed the PCI changing with the FSB mater of fact I at that time wrote them down showing what fsb kept the pci clock closest to stock as ya oced the board.
So according to that clearly the PCI is not locked..And also any one here running a NCCH-DL at 220 fsb ? How about 233 ? I am betting the answer is no..
Also if ya use clock gen yes the PCI appears locked BUT also if you use clock gen Windows or Sandra appear to not see the clock increse.So it is hard for me to trust Clock gen totally.
The more i look the more I think it has an AGP lock but no PCI..Wish someone had a definate answer..
So according to that clearly the PCI is not locked..And also any one here running a NCCH-DL at 220 fsb ? How about 233 ? I am betting the answer is no..
Also if ya use clock gen yes the PCI appears locked BUT also if you use clock gen Windows or Sandra appear to not see the clock increse.So it is hard for me to trust Clock gen totally.
The more i look the more I think it has an AGP lock but no PCI..Wish someone had a definate answer..
Last edited:
Seems to be a wall around 215-220 or so (although only based on 3 users):http://forums.2cpu.com/showthread.php?s=&threadid=63911
About the software: I think its a known bug with Sandra that it only reads boot up clocks or something (there were a few posts somewhere at 2cpu about this I think)
Also, I know clockgen IS actually changing the clocks because every benchmark that I've run other than Sandra shows improvement with its settings. Following the reasoning that clockgen actually changes the clock, if the pci/agp are tied to the fsb, why could I boot at 133 and be stable at settings between 133 and 166, and be stable at settings between 166 and 200?
For instance I boot at 133 and run 180. so if not locked, at 133 divider would be 4 (133/4 = 33mhz) and so at 180/4=45 mhz which would most likely not be stable?
There definately does seem to be conflicting suggestions. maybe the divider of 4 automatically switches to 5 after 166fsb, and to 6 at 200? then I could maybe see the wall at ~215, although by that logic, at 180/5 = 36mhz so 215/6=35.8 mhz should also be stable?
On the other hand, how many people have prestonia chips running at higher then say 210-215 fsb on any motherboard?
About the software: I think its a known bug with Sandra that it only reads boot up clocks or something (there were a few posts somewhere at 2cpu about this I think)
Also, I know clockgen IS actually changing the clocks because every benchmark that I've run other than Sandra shows improvement with its settings. Following the reasoning that clockgen actually changes the clock, if the pci/agp are tied to the fsb, why could I boot at 133 and be stable at settings between 133 and 166, and be stable at settings between 166 and 200?
For instance I boot at 133 and run 180. so if not locked, at 133 divider would be 4 (133/4 = 33mhz) and so at 180/4=45 mhz which would most likely not be stable?
There definately does seem to be conflicting suggestions. maybe the divider of 4 automatically switches to 5 after 166fsb, and to 6 at 200? then I could maybe see the wall at ~215, although by that logic, at 180/5 = 36mhz so 215/6=35.8 mhz should also be stable?
On the other hand, how many people have prestonia chips running at higher then say 210-215 fsb on any motherboard?
I'm sure you've seen the responses here since you posted the same question back in february, but just for anyone else viewing:
http://forums.2cpu.com/showthread.php?s=&postid=512060&highlight=pci+lock#post512060
Seems to be based on clockgen again though. But sheeno claimed 230 stable...
diehrd - what program were you using showing the PCI fequencies changing? I'me sure you've tried running memory slower than fsb to rule that out?
EDIT: Also may want to read through this page of the huge thread
http://forums.2cpu.com/showthread.p...how+what+the+PCI+and/or+AGP+bus+is#post469247
Still no conclusion but it seems that some are able to run up to 260 (Wow) fsb with a single cpu, but are limited with two cpus. At this point I think I would point to something with the implementation of 2 cpus that is limiting the fsb and not a pci/agp lock. I may have to pull one of my cpus tonight and see how high it can go...
http://forums.2cpu.com/showthread.php?s=&postid=512060&highlight=pci+lock#post512060
Seems to be based on clockgen again though. But sheeno claimed 230 stable...
diehrd - what program were you using showing the PCI fequencies changing? I'me sure you've tried running memory slower than fsb to rule that out?
EDIT: Also may want to read through this page of the huge thread
http://forums.2cpu.com/showthread.p...how+what+the+PCI+and/or+AGP+bus+is#post469247
Still no conclusion but it seems that some are able to run up to 260 (Wow) fsb with a single cpu, but are limited with two cpus. At this point I think I would point to something with the implementation of 2 cpus that is limiting the fsb and not a pci/agp lock. I may have to pull one of my cpus tonight and see how high it can go...
Last edited:
- Joined
- Dec 20, 2001
All I can say for sure is that the 875P can lock the PCI/AGP bus, so if/why ASUS didn't allow for a lock is beyond me.
I think its somehting else, because I could not even run 205x15 but I could run 16x200. This is with hardware that has all been tested to run way out of spec.
I think its somehting else, because I could not even run 205x15 but I could run 16x200. This is with hardware that has all been tested to run way out of spec.
- Thread Starter
- #9
{PMS}fishy said:
You see my point exactly..If I can run at 17/200 24/7 why not at 14x220 ?
I think the PCI lock gets twisted at the 200 Mark and loses its lock.I think if ya can run at 215 try 209 or some weird number that would normally set the PCI to a weird clock and tell us if you are stable..I doubt ya are..And if u are not then the PCI lock dies at 200 as I suspect it does..I think it was CPU FSB that showed the PCI change with the fsb if i remember now.
Last edited:
Well, I can definately manipulate the Agp/pci frequency independently of the fsb using clockgen. I kept my fsb low and worked the agp/pci speed up to about agp ~73 a few times and my rig would become unstable and lockup usually.
However I am booting at 133 and using bios revision 1003. It seems at least this chip doesn't want to post at 200 fsb (bios setting) maybe because of the higher multiplier, so I wasn't able to boot with the BIOS set at 200 (may try the other chip though).
I can't imagine why the pci lock would go out the window when booting at fsb at/over 200, as I thought many P4 boards with the same chipset could run very high bus speeds.
I also took out one cpu and can run my fsb up to ~221 mhz stable. At 222 - 225 it starts to fail prime/OCCT tests.
I still think it has something to do with both cpu's being installed (this seems to be an issue on most boards if I am not mistaken).
diehrd - I bet you could also pull on cpu and run higher than 215. Are your chips M0's?
However I am booting at 133 and using bios revision 1003. It seems at least this chip doesn't want to post at 200 fsb (bios setting) maybe because of the higher multiplier, so I wasn't able to boot with the BIOS set at 200 (may try the other chip though).
I can't imagine why the pci lock would go out the window when booting at fsb at/over 200, as I thought many P4 boards with the same chipset could run very high bus speeds.
I also took out one cpu and can run my fsb up to ~221 mhz stable. At 222 - 225 it starts to fail prime/OCCT tests.
I still think it has something to do with both cpu's being installed (this seems to be an issue on most boards if I am not mistaken).
diehrd - I bet you could also pull on cpu and run higher than 215. Are your chips M0's?
- Joined
- Dec 20, 2001
diehrd said:
Even if the PCI lock didn't work at over 200, it does not explain why the MB refuses to run at 205. I think the MB just is tapped out at 200 fsb. Which would suck for anyone with Nocona CPUs.
My advise it to get a PC-DL and mod it, if you really must run 200+ Its the only MB that I have seen that runs 250+ with only a little modification.
- Joined
- Apr 22, 2004
- Location
- New Hampshire
does anyone know if this instability happens with the newer 800fsb chips?
- Joined
- Dec 20, 2001
veryhumid said:
I can only imagine that it does.
The MB only goes to 233 FSB (you can go higher with clockgen) but I dont see anyone making it to even 233.
{PMS}fishy said:
I'm not sure "the motherboard is simply tapped out at 200fsb." I mean, it has been shown to run higher with only 1 cpu installed, so there is something (apparently fundamental to the operation of the board) that happens when 2 cpus are installed causing instability. What that is I have no clue.
{PMS}fishy said:
I know you had problems with your PC-DL, but what was the highest fsb you obtained and which mods besides vdimm did you do?
Yeah, nothing seems 'wrong' with the NCCH in single cpu mode:
267 fsb (memory downclocked) Matlab bench:
Only prime stable for 15 minutes but I was impatient. Not stable at 271. I think running daily at 265 wouldn't be a problem though. Its a shame this potential has yet to be tapped in dual-cpu mode...
267 fsb (memory downclocked) Matlab bench:
Only prime stable for 15 minutes but I was impatient. Not stable at 271. I think running daily at 265 wouldn't be a problem though. Its a shame this potential has yet to be tapped in dual-cpu mode...
- Joined
- Dec 20, 2001
Albigger said:
vDIMM and vCPU I ran 230 FSB for abit with no problems. I only had the MB for a week so I didn't have a chance to play with it much. With a vAGP mod the board sure flys.
Similar threads
