Weird wall at 240Mhz bus. Need a bit of help from the forum gurus

[TheStigma]

Perhaps someone can help me out with this...

Im running a 3000+ winchester on a MSI K8N Neo2 plat, together with 2 gigs of OCZ 3200 Platinum rev2 (mmmm love that RAM).

Anyway, im currently doing 239FBS 4xHTT and running ram at 2-2-2-10 (166). The PC seems rock stable. I havent done extended testing, but I did half an hour in prime without a single error, so I should be pretty far away from pushing the CPU over the edge. At this point im still using default 2.4 Vcore.

However... (and this is the weird part), whe second i set the FBS to 240, I get frozen a few seconds after loading windows. Tried lowering HTT multiplier, tried eliminating RAM from the equation, tried upping Vcore up as far as 1.55, and stil nothing.

I cant understand that the chip would simply go from perfectly stable, to "no-way-no-how" over that single Mhz. There must be something im doing wrong.

Unfortuantely I dont have BIOS version or core data availiable, but I can dig up the bios version atleast if that becomes relevant to giving me tips to try.

of course, let me know if theres any other data i can provide to help make an analysis.

-Stigma

 

[DerwenArtos12]

please say you meant 1.4vcore not 2.4, if you actually mean 2.4 turn it down before that thing lights on fire. here is my suggestion for your settings:
1.5vcore,
2.7vram
3xhtt
2.5-3-3-10 timings
synch up the memory
that thing will be a beast and a half a 239 x 9 =2.15ghz

if you insist on going higher then you can try but i really have come to the point where i don't expect top of the line results from MSI. this is just an opinion though. I haven't worked with a neo 2 but after the coupel i build of the origonal new plat I just stopped buying MSI.

 

[TheStigma]

yea, i meant 1.4 obviously, or else it would be one hell of a stock voltage hehe.

Il try those settings out, and see what I end up with. Couple of questions tho:

Why run HTT at 3x? Allready at 4x its below the default 1000. (4x239=956). I don't understand the reasoning behind lowering it further.

Also, its not that im not happy with a 19,5% speed increase, but it seems to me from all the posts that its become allmost routine to get these chips up near 2.4 with good air cooling, or even stock air. I firmly believe that the chip has more potential, since it ran prime95 stable for 3 hours yesterday at 239, but refuses to work more than 30 seconds in windows at 240.

Lastly, I read just now somewhere that the SATA1&2 on the MSI K8N NEO2 plat. are NOT locked, infering thaty the 3&4 are locked i guess. I think this may be a possible cause for my problem, as i have a raptor running on SATA1. I will try later to move it to 3 to see if that corrects the windows freeze. If its really true 1&2 arent locked, then im amazed that it even worked at 239 actually, perhaps 240 was too much, but for the drive, not the CPU. Can you (or anyone else) confirm that 1&2 are not locked, but 3&4 are?

Seems ike you don't have much love for the MSI Neo2. I just bought it because it had a lot of nice reviews of it, especially in regards to overclocking. I very much doubt that the motherboard chockes on 240, but just for the sake of interrest, wich board would you buy to overclock a winnie? Im gonna buy a second board for a winnie anyway, since im upgrading my server as well as my gaming rig, so I suppose I could just as well get a different brand to check out.

Thanks for the help
-Stigma

 

[TheStigma]

Wooohoo! I believe i just smashed the wall infront of me into a gazillion pieces hehe. This winnie is smokin'! (thankfully not in the litteral sense hehe). Looks like it was the Raptor that couldnt handle higher than 239 (wich is impressive tho, I was thinking it was locked). Set it to SATA3 and now everything works liek a charm! I've hit 250 now at stock Vcore (1.4) and still climbing. No indication that im close to any roof yet whatsoever. Prime stable for about 15 mins at 250, so il keep climbing a bit more I think =)

I'l report back later with updates. I think this baby might not be so bad after all hehehe.
-Stigma

 

[TheStigma]

Stable in short prime test at an amazing 275Mhz bus now. Im sure I can get it even higher, but all this testing takes a ton of time, and inching forward from this point will take even longer. I'l drop an progress update in here though when i get the time, for those that might be interrested.

-Stigma

 

[peonyu]

To get your true HTT speed you multiply like this :
HTT divisor speed [4 or 3 etc] times your FSB, then multiply the number you get by two.
so if you 200 fsb at 4xhtt x 2 you have a 1600 htt.
If you go above 1800 mhz htt then you may have problems so thats why most peps use 3x htt cause it gives much much more headroom for oc.

 

[DerwenArtos12]

wow, well good job, your almost at 2.5ghz. peonyu thanks for explaining the HTT for me. I completely forgot that the the msi nf-integrated sata aren't locked. You must have a winchester 90nm to be doing 275 at stock voltage, they are great chips.

 

[TheStigma]

Well im using HTTx3 now of course, but now im way way further up than before =)

Update: 275 rock stable in prime for an hour now, using 1.5 Vcore.
I have no doubt it can go higher, but I think I will pause it for now, until I get my proper cooling system in place. I'l have to recalibrate everything then anyway ^_^ Aiming to run CPU at about ambient temps using waste water cooling. Combined with my the new 480W watercooled PSU thats comming, the OC results could be pretty interresting =)

Thanks for the help guys.
-Stigma