Printable View
could it be your pci cards? Upping the fsb means upping the frequency to your pci cards too. Even with the pci dividers, your pci cards will still be running out of spec.
180mhz / 5 = 36mhz.
pci cards are meant to run at 33mhz.
Is there an option to make your pci divider divide by 6? That setting will bring your pci cards back into spec.
If not, then do you have any crappy pci cards installed on your system? Some cheap cards can't handle much. Remove them all and then try upping your fsb again.
FYI Patlo, the 8RDA+ locks the PCI bus at 33 and the AGP bus at 66. ;)Quote:
Originally posted by patlo
...Upping the fsb means upping the frequency to your pci cards too...
Guess i'm a little out of touch with oc'ing nowadays. But thanks for the info, i honestly didn't know that.Quote:
Originally posted by BlueOval
FYI Patlo, the 8RDA+ locks the PCI bus at 33 and the AGP bus at 66. ;)
Have you been changing your settings in small or large increments? It made a big difference for me. When I made big changes I tended to lock up the system and would fail to post. When I made small changes I was able to reach any setting I tried but would reboot instead of lock up when I went too far.
How small flap, like 1mhz increases? I tried 5mhz increases, maybe I'll try just 1 mhz at a time.Quote:
Originally posted by FlapJacks
Have you been changing your settings in small or large increments? It made a big difference for me. When I made big changes I tended to lock up the system and would fail to post. When I made small changes I was able to reach any setting I tried but would reboot instead of lock up when I went too far.
I've narrowed it down to the board or ram because I can get the proc up to 14 multi.
Do you guys think the NB has anything to do with it? I put a Vantec Iceberq on there with some Artic Silver Ceramique, so I don't think that's the issue. I'll start trying the 1mhz increases.
P.S. Here's another thing I thought of, could I be supplying TOO much voltage? Would that cause instability? I always upped the Vcore and Vdimm just so I was sure that too little voltage was not the issue. Could too much voltage be causing a problem? There's heatspreaders on the ram + I have my case open so I don't think the ram is getting too hot.
Ha, you'll love this sh1t.
Tried again to go just 1mhz to 181mhz. Multi set at 13, Vcore 1.950, Vdimm at 2.77.
Windows BSOD's halfway through booting. So I reboot and do the INSERT + DEL thingy to get back into bios, and just for the sake of arguement, try again, only this time I bump the Vcore WAY up to 2.1 and Vdimm to 2.9 (max).
Reboot, XP BSOD's again. This time, the INSERT + DEL thingy won't even work to get me back into the bios. I tried just regular delete and that got me in twice, except the bios kept 'freezing' halfway through making all of my changes. Third time I got into the bios, it turned into just a blue bar and cursor across the top of the screen, INSERT + DEL still wouldn't work, so I had to power down and clear CMOS with the jumper.
This just keeps getting wierder and wierder. But now it has me wondering and maybe it isn't the ram. I've got three possibilities I believe at this point:
1. Bad batch of ram (although bothsticks of ram used one at a time produces the exact same effect so then both sticks would have to be bad)
2. Bad mobo (I'm starting to lean toward this now because it seems totally absurd to me, no matter what setting I throw at it, I can't get 1 f'in mhz more out of the bus andthe fact that even oc'ing just 1mhz seems to screw up the bios when i try to get back in and make changes)
3. Mobo/Ram incompatibility (This seems unlikely as it seems like alot of other people are having success with the HyperX. I haven't seen too many people with the LL 2x512mb PC3500 sticks though, so this still could be the case).
Sadly, the only other ram I have to try is old CL2.5 PC2100 Crucial and I know that won't go over 180mhz, so it's a mute point to even try.
I don't know. What do you guys think about the three choices above? I can't RMA the board cause I've replaced the heatsinks (obviously), so I voided my warranty. I've RMA'd the ram and Kingston is sending me two fresh sticks (same type, size) and when I get those, I'm sending this ram back to them). But if those sticks produce the same result, it still won't prove it's the board cause I don't have another board to try this ram in and I don't have different brand ram to try in this board. This is so frustrating.
What do you guys think of those three?
Actually, I don't know if anything is bad. You reached an overclocked level, perhaps not one in which you felt satisfied, but it was overclocked nonetheless. I don't believe it to be #3 since many others are using the same mobo and RAM without issues. #1 and #2 are answered by my initial statement.
I think, perhaps, that it may have been due to your trying to jump by 5 MHz increments rather than by 1. See my reply in your new thread. Also, did you check your RAM ala memtest at all prior to RMA'ing it?
Have you visited www.8rda.com? They have tips on how to reach a high FSB. Things like don't use certain PCI slots, don't hook up fans to MB headers, things like that.
Also, check to make sure your Iceberg is making full contact with the NB, mine wasn't. That's worse than stock HS.
If this is more hyperx A1 stepping combo...its incompatbility issue. However some boards will work ...as with philip and his pc 3500 are at 231 stable. It maybe a vdimm voltage issue with the new stepping maybe? ITS CERTAINLY not anything to do with cooling or anything like that. As i have the same ram in an a3 stepping epox 8rda+ and it would do 200 before and now in the A1 stepping is only capable of 180
kirkyg
I checked your other reply. Yeah, using memtest86 @ 181mhz, I get 100,000+ errors (didn't even bother letting it go any further).Quote:
Originally posted by BlueOval
Actually, I don't know if anything is bad. You reached an overclocked level, perhaps not one in which you felt satisfied, but it was overclocked nonetheless. I don't believe it to be #3 since many others are using the same mobo and RAM without issues. #1 and #2 are answered by my initial statement.
I think, perhaps, that it may have been due to your trying to jump by 5 MHz increments rather than by 1. See my reply in your new thread. Also, did you check your RAM ala memtest at all prior to RMA'ing it?
2 things though, yes, the FSB is overclocked, BUT, the RAM is not overclocked. The ram is rated to 216mhz, so it should be able to run that. I tested this by setting the ram to 66% of the FSB. I could ramp the FSB all the way up to 230+!!!!! So the board can definetly attain that high a FSB when asynch with the ram. That proves that it's the ram holding it back.
I'm VERY satisfied with the overclock of my processor/FSB. I'm VERY DISSATISFIED with having to UNDERCLOCK my ram (180mhz < 216mhz). Granted, as stated by Epox, the board is only rated to support PC3200 (200mhz). I would be very happy to get 200mhz, but I can't even get to that with ram rated for even HIGHER than that!!! It's madness I tell you.
Plus, if you look at the "Can't get to 168mhz thread", check my reply on page 3 about the new 200mhz proc's that are supposed to be fully supported with this board.
::blanches:: Wow! No doubt it's the RAM then. :(Quote:
Originally posted by janko10
...Yeah, using memtest86 @ 181mhz, I get 100,000+ errors (didn't even bother letting it go any further)...
Ok let me try to get this straight..I've got this same exact problem but it goes a little further.My origional 8rda+Rev 1.1(i don't know wich stepping,N.B. had a metal center plug)had this exact problem of not being stable over 180fsb 100% memory but would run at 217fsb,66% memory with 2 sticks of corsair 512mb PC3200LL memory.Thought it was the memory so i RMA'd it.Herd great things about the Kingston Hyper-x pc3500 so i bought that.BAM!..same problems!I then did the V-Mod on my board.BAM!..same no go over 180.Then i read about these new A1 stepping boards so i whip out the o'll AMEX and BAM!...still at 180.I'd bet all the cash i spent on these boards and memory that if i still had that Corsair PC3200LL memory it wouldn't be running over 180 on the A1 stepping board either.Thats 2 different N.B. steppings with the same problem and the same problem with 2 different memories!Kingston's fault?Epox's fault?Who knows.Love this board in some respects..but this board makes working with VIA chipsets look like a walk in the park.Epox and Kingston are both pointing the finger at each other,which means to me this ain't getting fixed!Hope i'm wrong.