865 Motherboards, High FSB's 1:1 SHARDING!!!...

[bugsy]

Has anyone else experienced this?, where when using dual channel ram on high fsbs 250mhz (1:1) what can only be described as 'sharding' appears in high intensive games and mainly on 3dmark cpu tests

but taking the ram to sigle channel mode completly removes the problem.... any1 have any fixes for it? (abit is7, corsair ddr500 pro)

[STEvil]

AGP is losing its lock.

You need to try values other than 66mhz to fix this.

[bugsy]

What values should I try, Ive never heard of agp losing its lock before on p4... ( its a p4 northwood system btw )

[STEvil]

Dont remember what the values were, was between 71 and 81mhz it seems, though...

Just up it 5mhz at a time starting at 66 going to about 85, over 81 may have issues..

[bugsy]

Fidled around with it... 66 is default 67, 68, 69, 70, 75, 80, 85 all had issues not detecting my sata drives and it did not fix the problem....

[STEvil]

darn, that sucks.. more chipset voltage might get you somewhere, or bumping agp voltage a notch or two... but if you're losing your sata drives by bumping it 1mhz dunno what your gonna be able to get.

[bugsy]

Ive been told its a 865 chhipset limitation from a compatibility issue with difefrent brands of memory... :(

[mtb856]

I had that problem when I tried to go above a 300FSB on my board..... I had to raise the AGP to 72mhz to get it stable. That trick worked with both my 865 and 875 chipset BTW.

I've heard of the problem with SATA drives when you do that, but I don't think I've heard of a way of getting around that.....other than sticking a IDE drive in for when you want to do a max overclock.

Like STEvil said, you might try raising your AGP voltage to 1.6 or 1.65V and see if that helps..... it can give you a little more memory and chipset stability depending on the board.

[meke]

I am not aware of "SHARDING" but, is it something like resizing and refreshing and some color changing but everything comes back on a sec.?

if so,
I am facing the same stuff with p5p800 865 lga board with a dual vitesta ddr600.

yes, it is not avaliable on daily fsb's, but very often avaliable on 275 fsb + especially on benchmarking.

I will try to uplift the agp clock !

[bugsy]

I had that problem when I tried to go above a 300FSB on my board..... I had to raise the AGP to 72mhz to get it stable. That trick worked with both my 865 and 875 chipset BTW.

I've heard of the problem with SATA drives when you do that, but I don't think I've heard of a way of getting around that.....other than sticking a IDE drive in for when you want to do a max overclock.

Like STEvil said, you might try raising your AGP voltage to 1.6 or 1.65V and see if that helps..... it can give you a little more memory and chipset stability depending on the board.


Nope didnt change it one bit... I said before when I run the memory in 5:4 (ie 200mhz) the problem is gone, and when the memory is run in single channel the problem completly dissapears even at high fsbs.... :confused: :confused: :confused:

[mtb856]

What PSU do you have right now? It might not be supplying enough power to handle an overclocked processor and ram.... although I'm not sure how much power ram uses. Just throwing out ideas. I'll try to dig up some more info....I've never had sharding, so I don't know of a solution off the top of my head.

[bugsy]

Superflower 500w, I have tried other psu's....

[bugsy]

any ideas???

[STEvil]

chipset/agp voltage.. ram timings.. sideband addressing.. drop to agp4x.. etc?

[bugsy]

chipset/agp voltage.. - tried

ram timings.. - 2.5-3-3-7 to 4-4-4-10 nope...

sideband addressing.. nope...

drop to agp4x.. nope...

[bugsy]

I Have The Solution!!!!!!!!!!!!!!!!!!! Use 865 Tweaker And Select Dram Latency Paremeter To 01!!!!!!!!!!!!!! It Fixes It!!!!

[STEvil]

free bump, maybe someone with a high FSB setup that wont do 1:1 2-2-2-x should try this and see if it fixes high fsb with tight timings?

[bugsy]

im was running 1:1 @ 250 mhz

[STEvil]

high fsb is 260+

[bugsy]

Picky picky :D I know this is it cause as soon as I chnage it back to its default of "00" it goes CRAZY!!!!!!

[enzoR]

northbridge unable to handle it..

[poch]

anyone seeing this on an 875 board? i seem to be encountering it every now and then and it's pretty annoying! is there a solution?

[bugsy]

I Have The Solution!!!!!!!!!!!!!!!!!!! Use 865 Tweaker And Select Dram Latency Paremeter To 01!!!!!!!!!!!!!! It Fixes It!!!!

dude read up abit...

[GazC]

In addition, is it worth checking that the chipset cooling is effective? I've seen some Nforce2 mobos do this when the NB was over heating.

[poch]

I Have The Solution!!!!!!!!!!!!!!!!!!! Use 865 Tweaker And Select Dram Latency Paremeter To 01!!!!!!!!!!!!!! It Fixes It!!!!


just wonderin where i can set this on an IC7 :confused:

[bugsy]

In addition, is it worth checking that the chipset cooling is effective? I've seen some Nforce2 mobos do this when the NB was over heating.

I have the solution!!! ^^


as far as I know this only happens on 865p motherboards, you use the tool called "865 tweaker"

[WeldZilla]

I run 3.0C at 269 fsb 1:1 on my 865 and have never seen this. I run my 2.4C daily at 303fsb 0n the 5:4 divider and have not seen this. I will tell you once I hit 303fsb I had to turn my agp up to 68 mhz just so I could keep my vid card clocks high. Both my processors run on the Epox 4PDA2+Ver.2. I am running Gskill mem with my 3.0C at 269 1:1 and OCZ 4000Gold Rev.2 with my 2.4C on the 5:4 divider.

WZ

[GazC]

I have the solution!!! ^^


as far as I know this only happens on 865p motherboards, you use the tool called "865 tweaker"

I know you solved, it. That is why I said in addition.

[cpulloverclock]

My best run is 278 1:1 DC with my BH-5 full timings and a friend runs at 300 with 2048MB and his P4C800-E-DX (A-data PC4800)