E6600 woes

The DS3 is actually an intel 965 chipset
Click to expand...

blah, I've been seeing the numbers six, eight, and zero for like a month and half straight, so my brain automatically just assumed thats what the DS3 was lol, my bad. I don't know if your last post was hinting that you might try a 680i, but if you are I'd point you to EVGA's A1.Better stability reports and EVGA's support sounds good from what I've read. My friend has an A1 and loves it. Admittedly I know very little about the 965 except that it out performs the 680i's RAID capabilities, so I don't have 2 cents to give you there =/
 
J_Devil said:
blah, I've been seeing the numbers six, eight, and zero for like a month and half straight, so my brain automatically just assumed thats what the DS3 was lol, my bad. I don't know if your last post was hinting that you might try a 680i, but if you are I'd point you to EVGA's A1.Better stability reports and EVGA's support sounds good from what I've read. My friend has an A1 and loves it. Admittedly I know very little about the 965 except that it out performs the 680i's RAID capabilities, so I don't have 2 cents to give you there =/
Click to expand...

That was the board I had been looking at... however I'm really in the game for big OC's.. and the 680i has yet to prove it's solid capable to do that, it's really a different kind of board from what I'm really looking for.
 
cerebrex, it's almost a foregone conclusion that your cpu is sketchy as far as the overclocking. I bet if you ran the cpu at stick speeds or say, 2.8ghz, the board would not have a single issue. Instead, you go on a rant and trash the board, as if getting a huge overclock is GUARANTEED, and now you are going to send back a probably good board, replace it with another board, and probably have the same problems. Thanks for raising prices for us overclockers. Then you go post in other threads, trashing the DS3 board, when in fact it still is probably NOT the board's fault. Stop being a jackass and try a different cpu.

You apparently seem a bit confused about the entire troubleshooting process.
 
TheRapture said:
cerebrex, it's almost a foregone conclusion that your cpu is sketchy as far as the overclocking. I bet if you ran the cpu at stick speeds or say, 2.8ghz, the board would not have a single issue. Instead, you go on a rant and trash the board, as if getting a huge overclock is GUARANTEED, and now you are going to send back a probably good board, replace it with another board, and probably have the same problems. Thanks for raising prices for us overclockers. Then you go post in other threads, trashing the DS3 board, when in fact it still is probably NOT the board's fault. Stop being a jackass and try a different cpu.

You apparently seem a bit confused about the entire troubleshooting process.
Click to expand...

You know what? re-read every post I made, your reply is completely misinformed, you obviously only read the first few posts I made and decided you understood the situation. I don't think I've ever witness someone jumping to radical conclusions and progressing to a completely militant stance so quickly in my life.
 
TheRapture said:
cerebrex, it's almost a foregone conclusion that your cpu is sketchy as far as the overclocking. I bet if you ran the cpu at stick speeds or say, 2.8ghz, the board would not have a single issue. Instead, you go on a rant and trash the board, as if getting a huge overclock is GUARANTEED, and now you are going to send back a probably good board, replace it with another board, and probably have the same problems. Thanks for raising prices for us overclockers. Then you go post in other threads, trashing the DS3 board, when in fact it still is probably NOT the board's fault. Stop being a jackass and try a different cpu.

You apparently seem a bit confused about the entire troubleshooting process.
Click to expand...

Unless your a DS3 3.3 user having these problems it’s pretty hard to relate to the frustration we’re having. I’ve been dealing with the same issues. Stable settings posting back at default or not posting at all requiring a CMOS reset. I finally decided to just leave it at 335*9 a few days ago and last night it posted at 200*9. This is definitely the most frustrating out of probably 20 MB’s I’ve used in the last few years, especially after doing a lot of research on what board to get. If you take a look here you can see this isn’t isolated to a couple of users.

http://www.xtremesystems.org/forums/showthread.php?t=138348
 
cerebrex, I did read all your posts, and it seems you have decided the cpu is NOT the problem, when in fact, it MIGHT be. But yet you keep trashing the board, in this thread and in another I have read. That's why I jumped on you. But it's ok, I hope you get your issue fixed, but I know you are going to hate it when you get a new board and have the same thing happen. Maybe it won't.


Have you even TRIED running the board for a couple days at a mild overclock or even stock speeds?


ALSO, if you have ram rated at 2.1v, the board defaults to 1.8v at bootup, and this may not be enough to get it going. I researched into this board carefully before I bought one, and bought ram that would run at 1.8v at stock speeds, and have not had an issue yet with it. Once booted I can change the ram volts to whatever I want and then overclock from there. The thread you reffered me to over at extremesystems also notes this issue.
 
TheRapture said:
cerebrex, I did read all your posts, and it seems you have decided the cpu is NOT the problem, when in fact, it MIGHT be. But yet you keep trashing the board, in this thread and in another I have read. That's why I jumped on you. But it's ok, I hope you get your issue fixed, but I know you are going to hate it when you get a new board and have the same thing happen. Maybe it won't.
Click to expand...

Yes, that would definately suck, and if it does, you will know it, just follow the thread, you will have all the time you want to trash me. I do however, feel quite confident the processor is not the issue... a processor would not cause bios settings to randomly erase themselves, or cause the computer to fail to post at a 3.0GHz clock when it ran 3.5GHz stable through 24hrs of stress testing... and then a few days later.. it runs stable at 3.5GHz again...?!


TheRapture said:
Have you even TRIED running the board for a couple days at a mild overclock or even stock speeds?
Click to expand...

Yes, 3.0GHz is solid. - 9x @ 334MHz - 9x @ 335MHz won't post (seriously, wtf)


TheRapture said:
ALSO, if you have ram rated at 2.1v, the board defaults to 1.8v at bootup, and this may not be enough to get it going. I researched into this board carefully before I bought one, and bought ram that would run at 1.8v at stock speeds, and have not had an issue yet with it. Once booted I can change the ram volts to whatever I want and then overclock from there. The thread you reffered me to over at extremesystems also notes this issue.
Click to expand...

My ram is actually 2.2v rated (Crucial Ballistix PC2-8000) and I'm running +.04v for the mem, it has never failed a mem test, so I've completely eliminated this is an issue.. primarily because the board might not even be physically able to reach a speed this memory could handle on stock settings.
 
cerebrex....are you sure the board is booting the ram at 2.2v? If you reset the cmos or crash during a cold boot, the ram is maybe getting set back to 1.8v and causing issues....just trying to help. Sorry I got so rough to start off. I tend to do that even though I meant no real harm. I apologize, shit, thanks for getting me straight. I reread my post and yeah, it was a bit out of line. :rolleyes:
 
TheRapture said:
cerebrex....are you sure the board is booting the ram at 2.2v? If you reset the cmos or crash during a cold boot, the ram is maybe getting set back to 1.8v and causing issues....just trying to help. Sorry I got so rough to start off. I tend to do that even though I meant no real harm. I apologize, shit, thanks for getting me straight. I reread my post and yeah, it was a bit out of line. :rolleyes:
Click to expand...

hehe, no sweat.. I do the same thing.. as long as everyone gets treated with respect the world would be a great place. The memory will boot at 1.8v.. it's never been a problem for me.. very versatile memory (I'd definately recommend it, outside of the fact it was more expensive than my 8800GTS). The memory always cooperates.. I was able to overclock it with a 550MHz FSB to 1100MHz and did it on stock voltage.. didn't even have to bump it to 2.3v.. The board does have some advanced timings that I don't exactly understand so I live them to auto.. but it doesn't act like RAM is the issue.. like I said I can save a 3.5GHz overclock which will run stable through orthos.. never lock freeze, hell it doesn't even run hot... and keep screwing around, somewhere I'll fail to boot, and if I attempt to revert to the 3.5GHz settings that worked perfectly... it won't even post.

My whole objective here is 3.6GHz minimum... It's not for performance, or games specifically, its because I went the extra mile to throw down on pricey memory and processor.. and damnit, if I wanted 3.0GHz I would have bought a 4300 with some cheap ass 533 ddr2
 
The instability and random problems all "seem" (I say that with great discretion) to originate from power issues revolving around the motherboard itself. I noticed the big problems started to happen when I was running raised voltages across the MCH, PCI-E, FSB, 0.1v across all were okay, when I brought the MCH to 0.2v it stablized my 500fsb 3.2GHz clock.. when I then increased voltage on the FSB it went nuts... but it did it randomly, and randomly changed bios settings around too.. so it's too difficult to be for sure if this did cause the issues, or if it was something entirely unrelated.

on a seperate note - my conversation with Gigabyte support - After enduring about 10 minutes of engrish accusations that my processor was bad, or my ram, etc.. etc.. my fault.. etc.. I continued to push the fact that the board is doing as it wishes despite my efforts.. and it's completely random. I kept trying to drive the rep into a corner and admit that if the system is rock solid at a clock.. and then restart it and it won't even post.. at all... something doesn't make sense.. clearly it's proven itself stable... he didn't speak good enough english to understand he lost the argument.. eventually he just started spewing "sen back do bo" which is when I thanked him for his time and disconnected the call... didn't really get anywhere with that. I didn't appreciate the inquisition I experied from their support, he was ANGRY at ME for having the problems.. good god.
 
Read this thread but there are several others scattered around so you might have mentioned it, I dont know.

How are you cooling the MCH ? For the kind of OC you are looking for just putting a fan on it will not do. The factory application of crappy thermal compound done in a poor manner will trap the heat in the chip making the fan of limiited usefullness once you really start to push it.

Default MCH voltage is 1.25 V
VCC 1.25 V Core Supply Voltage with respect to VSS max - 1.375 V

Avoid the 350-400 FSB range, lower mulitplier and jump over it.


Pull the board out, take off the heatpipe, trash those crappy push pins, clean everything, AS-5 or ceramic, replace the pushpins with nylon 6-32x3/4 screws nuts washers, I think it was 6-32, its cheap so get a set of 8-32x3/4 hardware too, use 8-32 if you can, even if you CAREFULLY have to enlarge/ream out the holes just a TINY bit with a knife with a thin sharp point. I cant remember which size fit when I did mine, if the 8-32 is not close to fitting, dont use it and the 6-32 will go in easily. Get extra nuts so you can double nut the screws so they will not loosen up. Snug the hardware down and then give 1/2 turn , use your good judgement, there is some give in the nylon but we dont want to crack the chip. Firmly snug with no movement when wiggleing the heatsink. Fan back on, pointing down. ICH/southbridge too.

IF watercooling a northbridge block would be the way to go but I would do the above and see if that causes a lot of your issues to go away and if so then consider the bigs bucks for a NB block.
 
BillParrish said:
Avoid the 350-400 FSB range, lower mulitplier and jump over it.
Click to expand...

I can post at around 365 FSB, but no luck with anything over. It generally defaults back or doesn't post a second time anyway. even if I jump to 401 it won't post anything over 365. Is there any better bus speed over 400 that would be more likely to post? My RAM's memtest stable at least to 840 on this MB. Sorry if I'm hijacking the thread.
 
yeah, this whole strap business is kinda hard to figure, to me at least :)

I got another e6600, L630B016. This one starts at very low voltage, like 1.2 or something (1.18 in cpu-z), but runs kinda hotter than my other e6600 L642A297. I'm still using Intel HSF (although a different one, but how different can they be?). I played around with both, and it seems the L630 is much more overclockable than L642, but it needs a beefier cooler... Of course, I had L630 for only a couple of hours in the system, so AS5 hasn't cured yet but still, it runs much hotter than L642
 
BillParrish said:
, replace the pushpins with nylon 6-32x3/4 screws nuts washers, I think it was 6-32, its cheap so get a set of 8-32x3/4 hardware too, use 8-32 if you can, even if you CAREFULLY have to enlarge/ream out the holes just a TINY bit with a knife with a thin sharp point. I cant remember which size fit when I did mine, if the 8-32 is not close to fitting, dont use it and the 6-32 will go in easily. Get extra nuts so you can double nut the screws so they will not loosen up. Snug the hardware down and then give 1/2 turn , use your good judgement, there is some give in the nylon but we dont want to crack the chip. Firmly snug with no movement when wiggleing the heatsink. Fan back on, pointing down. ICH/southbridge too.

.
Click to expand...

I thought it was you that I got the idea of using screws instead of push pins.... thank:) I used 4-40 x 3/4 on my NB and it fits without any mods....

If you think its the mobo thats giving you a headache, send me some $$$$ and Ill ship you my commando lol.... Guaranteed 550fsb;)
 
Thanks for the info, I'll pick up some hardware on the way home from work tonight to do that. The MCH never heated up to an even warm temperature with the fans on it... but if what you say is accurate about it's poor heat transfer that could explain it, I will give it a whirl... I just hope it doesn't piss newegg off when they end up getting it back if this doesn't work out.
 
BillParrish said:
Read this thread but there are several others scattered around so you might have mentioned it, I dont know.

How are you cooling the MCH ? For the kind of OC you are looking for just putting a fan on it will not do. The factory application of crappy thermal compound done in a poor manner will trap the heat in the chip making the fan of limiited usefullness once you really start to push it.

Default MCH voltage is 1.25 V
VCC 1.25 V Core Supply Voltage with respect to VSS max - 1.375 V

Avoid the 350-400 FSB range, lower mulitplier and jump over it.


Pull the board out, take off the heatpipe, trash those crappy push pins, clean everything, AS-5 or ceramic, replace the pushpins with nylon 6-32x3/4 screws nuts washers, I think it was 6-32, its cheap so get a set of 8-32x3/4 hardware too, use 8-32 if you can, even if you CAREFULLY have to enlarge/ream out the holes just a TINY bit with a knife with a thin sharp point. I cant remember which size fit when I did mine, if the 8-32 is not close to fitting, dont use it and the 6-32 will go in easily. Get extra nuts so you can double nut the screws so they will not loosen up. Snug the hardware down and then give 1/2 turn , use your good judgement, there is some give in the nylon but we dont want to crack the chip. Firmly snug with no movement when wiggleing the heatsink. Fan back on, pointing down. ICH/southbridge too.

IF watercooling a northbridge block would be the way to go but I would do the above and see if that causes a lot of your issues to go away and if so then consider the bigs bucks for a NB block.
Click to expand...

how high should I be increasing my voltages across the FSB, MCH, and PCI-E? I've been running your default recommendations of +0.1v - however I did end up running +0.2v FSB & MCH to post at 500MHz FSB... the board "seems" to respond better with lower voltages on itself, and the processor... which I see as a little odd.. if my cooling is sufficient, it should increase the stability, not decrease... so I will take your advice and really dig into it tonight and try to get it worked out.
 
cerebrex said:
Thanks for the info, I'll pick up some hardware on the way home from work tonight to do that. The MCH never heated up to an even warm temperature with the fans on it... but if what you say is accurate about it's poor heat transfer that could explain it, I will give it a whirl... I just hope it doesn't piss newegg off when they end up getting it back if this doesn't work out.
Click to expand...

My stock nb never got hot either but now that I switched to HR05, it gets hot when I dont have a fan blowing at it..... As I expected, the HR05 didnt help my overclock one bit, I admit though that I bought it more for "Bling" than anything else.....
 
Alright Bill.. I did everything you suggested.. oh and btw.. I had to drill out my motherboard (TERRIFYINGLY CLOSE to circuits) just to fit the 8/32" - Your DQ6 must have larger mounting holes for your heat pipes than the DS3 has for it's heatsink.

I also lapped both heatsinks and polished them to chrome like finish, and layed a paper thin coat of AS5... I mounted a 60mm fan on my NB and a small fan on my SB which I also lapped.. no changes as of yet, still wont post.. I'll really get it a whirl tomorrow and let you know.
 
cerebrex...have you tried running the board outside of the case? I had an issue with mine because a high solder point on the back of the board would contact the heatsink backplate when I tightened it down firmly. Sometimes it would post, sometimes it would not, and it about drove me crazy. Loosening the bolts would let it post but temps were too high due to not enough pressure. A quick filing session cured that problem.
 
TheRapture said:
cerebrex...have you tried running the board outside of the case? I had an issue with mine because a high solder point on the back of the board would contact the heatsink backplate when I tightened it down firmly. Sometimes it would post, sometimes it would not, and it about drove me crazy. Loosening the bolts would let it post but temps were too high due to not enough pressure. A quick filing session cured that problem.
Click to expand...

I can give it a shot.. but that's getting rather desperate don't you think?

It again won't even post above 3.06GHz... it did once post 3.15GHz but I headed back into the bios and it never did again.

I never thought it was my northbridge but I'd give it a shot just to make sure the board was junk.. it doesn't seem like a heat issue since the board isn't posting and becoming unstable as it heats up, it just never posts....

I'll say it again so we don't waste time on it anymore on this - I've already spent a combined ~15 hours of labor - heat is not my issue - The system temp is 26c and NB has never risen above 32c, ram doesn't even get warm, and with the new much thinner coat of AS5 on my lapped psu and waterblock, It hits a max load temp of 39c now... an impressive 6c drop.. I knew that coat of AS5 earlier was way way too thick.. coupled with what I believe was some air in my loop I didn't notice. I will post up some pictures this evening of my set up so you can see why I appear so confident about my cooling.

The important (and troubling issue) to remember is after the new PSU was installed yesterday, it posted AND benched 3.15, 3.20, 3.50?! what the hell is this things issue? My next strategy is pulling the battery to completely wipe the cmos (just to be sure) and if no success I will flash to F10, and then from there F7 (the most proven bios to date second to F9 which I already run) and if I still do not have success I will pre-order the DFI 965 board.

What I would really like to see is some successful overclocks, and all of their MIT settings... most importantly what voltages they are running for the RAM, MCH, FSB, PCI-E and their memory clock settings... I don't want total voltage, that doesn't help me since the bios doesn't even show that, it only shows +0.1v +0.2v etc... A few times, rarely, when I reset to a new bios setting, the system will begin quick beeping.. as if the memory is not present, or it could be what gigabyte refers to as "power error" - whatever the hell that's suppose to mean. If someone experienced with this board knows the circuit contact points, I'd like to fire up my fluke multimeter and manually verify voltages across the various circuits. If we do in fact run into some voltage droop issues, I can trace down some stronger circuits and manually re-route power with some soldered jumpers.

My goal here is to hit 4.0GHz (I'll take 3.6GHz), or show that I clearly without a doubt have a defective product... the last thing I hate seeing is baseless accusations.. If we finally come to the conclusion that the DS3 Rev. 3.3 is junk, I want solid evidence (this thread). Okay, let's hear some more suggestions on what could cause the board to post sometimes, and then not others... acting like a complete different piece of hardware with no noticable changes - all happening at random it would seem. Like I've said before, the weirdest thing is when it IS working and I'm making decent progress in my clocks.. it doesn't reset itself after a change. When the board starts being retarded and failing to do anything, it starts shutting down after a bios change, and then tries restarting - this is a clear transition I've noticed. Again, I can't think everyone enough here that's [H]ard enough to help me fight through this beast.
 
Apparently some lunch break research has revealed that w/ an 8800 vid card PCI-E frequencies of 110MHz+ have stabilized OC's.. I'll give that a try...

Also, I saw Anandtech made tonsof references to apparent incompatibilities and problems with Micron D9 based memory on the DS3 - I'm running the Ballistix DDR2-8000 sticks - Micron D9GMH IC's .. was this problem completely resolved yet? I remember they mentioned F9 was suppose to 'address some' of the problems.
 
So what now, try different ram, or a different board? Decisions, decisions...
 
TheRapture said:
So what now, try different ram, or a different board? Decisions, decisions...
Click to expand...

I don't have any other ddr2 unfortunately.. I'd like to see if anyone knows if it has been fixed in the more recent bios updates - anandtech review was pretty dated.
 
Hey I got your pm. I skimmed over the 4 pages of this thread, and from the looks of it, you've already done some extensive fiddling with unsatisfactory results. Again, I only skimmed through it all, so I apologize if something I suggest has already been tried.

To get into the low 400's stably, you shouldn't need more than 1.3vfsb for dual core c2's. So setting +0.1vfsb should be fine. For the purpose of testing, higher is probably doable, and a safer bet, but I have also seen situations where too high a voltage doesn't help. It's usually not the case though. And again, I don't need to be telling you this since it looks like you've done extensive testing already. And you've been able to get 500fsb, and you say even 550fsb?? 550 is quite stellar!!

vmch, a lot of people have said in the past that 1.55vmch reduces cold boot issues on the ds3. This is +0.3v. However, for running 400fsb, especially at 1:1, it shouldn't requite much more vmch if any.

Setting voltages to auto will cause them to automatically scale up, which I'm not a fan of, especially in the ds3 since there's no +0.0v option to maintain a baseline.

Since your temps are so good, we should be able to set vcore up at a higher value to boot; let's try 1.55v for starters. With droop and everything accounted for, you should be in safe zones.

Can you try f10 bios? Also, keep ram divider at 2.00x so it runs 1:1, since it's easier on the chipset, and also for the most part rules out ram as a concern. I know your sticks are rated 2.2v, but even still, I've had d9's flake out and require cooldown periods before working again when running 2.2-2.5v without direct fan cooling.

Let's try running 1:1 for 800ddr2, and have vdimm at +0.3v for 2.1v. You can use speedfan to look at the second vcore reading, which is actually vdimm. Set timings to 5-5-5-15 for the time being to try to rule out ram. Subtimings, I think you can probably set them to auto. IIRC, the ds3 sets subs pretty loosely. But even still, if you could post a screenshot of memset once you get into windows, it'd help, in case we see that sub's are being set too tightly still.

Also, I think there's some fast memory mode or something to that effect that might only show up after hitting ctrl+f1. Set it to normal or disable or whatever that setting is.

pcie, you can try 110 like you mentioned and report back. It shouldn't require any extra vpcie, but if you want to be safe, you could +0.1v it.

I wish I could help more, but I haven't touched my ds3 and am starting to forget it.

That said, let's do one more round of testing if you don't mind to verify that the motherboard is in fact decent.

If you don't mind, could you run these cookie cutter settings:

7x400=2800mhz
2.00x 800mhz 5-5-5-15 auto's, ram not at fast
1.425vcore, +0.1vfsb, +0.2vmch, +0.1vpcie, +0.3vdimm (some of these are higher than needed already, but since there's no +0.0v, I figure we'd settle for these)
110pcie

If you wouldn't mind, could you use the computer like that for a day or so, with orthos or wahtever to see if there's stability. If so, you know you can rule out motherboard. Then you can try raising to 8x, and the only thing you should need to possibly change is vcore.
 
ziddey said:
To get into the low 400's stably, you shouldn't need more than 1.3vfsb for dual core c2's. So setting +0.1vfsb should be fine. For the purpose of testing, higher is probably doable, and a safer bet, but I have also seen situations where too high a voltage doesn't help. It's usually not the case though. And again, I don't need to be telling you this since it looks like you've done extensive testing already. And you've been able to get 500fsb, and you say even 550fsb?? 550 is quite stellar!!
Click to expand...
Well, it was stellar for about an hour, up until I shut off orthos, and rebooted since I have verified it was stable under load, and tried to go higher.. then it all went to shit and since I've never even made 3.2... on the exact same bios save.


ziddey said:
vmch, a lot of people have said in the past that 1.55vmch reduces cold boot issues on the ds3. This is +0.3v. However, for running 400fsb, especially at 1:1, it shouldn't requite much more vmch if any.

Setting voltages to auto will cause them to automatically scale up, which I'm not a fan of, especially in the ds3 since there's no +0.0v option to maintain a baseline.

Since your temps are so good, we should be able to set vcore up at a higher value to boot; let's try 1.55v for starters. With droop and everything accounted for, you should be in safe zones.

Can you try f10 bios? Also, keep ram divider at 2.00x so it runs 1:1, since it's easier on the chipset, and also for the most part rules out ram as a concern. I know your sticks are rated 2.2v, but even still, I've had d9's flake out and require cooldown periods before working again when running 2.2-2.5v without direct fan cooling.
Click to expand...
Tonight I will flash over to F10 if I can't resolve my issues with your suggestions.

ziddey said:
Let's try running 1:1 for 800ddr2, and have vdimm at +0.3v for 2.1v. You can use speedfan to look at the second vcore reading, which is actually vdimm. Set timings to 5-5-5-15 for the time being to try to rule out ram. Subtimings, I think you can probably set them to auto. IIRC, the ds3 sets subs pretty loosely. But even still, if you could post a screenshot of memset once you get into windows, it'd help, in case we see that sub's are being set too tightly still.

Also, I think there's some fast memory mode or something to that effect that might only show up after hitting ctrl+f1. Set it to normal or disable or whatever that setting is.

pcie, you can try 110 like you mentioned and report back. It shouldn't require any extra vpcie, but if you want to be safe, you could +0.1v it.

I wish I could help more, but I haven't touched my ds3 and am starting to forget it.

That said, let's do one more round of testing if you don't mind to verify that the motherboard is in fact decent.
Click to expand...
I'm just about to reboot and give the PCI-E 110 a shot just to see if I can get lucky and pin down the exact problem early.

ziddey said:
If you don't mind, could you run these cookie cutter settings:

7x400=2800mhz
2.00x 800mhz 5-5-5-15 auto's, ram not at fast
1.425vcore, +0.1vfsb, +0.2vmch, +0.1vpcie, +0.3vdimm (some of these are higher than needed already, but since there's no +0.0v, I figure we'd settle for these)
110pcie

If you wouldn't mind, could you use the computer like that for a day or so, with orthos or wahtever to see if there's stability. If so, you know you can rule out motherboard. Then you can try raising to 8x, and the only thing you should need to possibly change is vcore.
Click to expand...
Actually, I've been running at 9x334 for the paste two weeks, it is rock solid stable, I'm rather confident the cpu will cooperate if given what it needs, so thankfully I won't have to get into that, we've already passed it.

I really appreciate all your insight.. I will report back momentarily (and if I don't, call gigabyte tech support and tell them to call me, haha) as to what I've uncovered... we're idling at 19c right now, so it should be ripe for the pickin'

thanks again!
 
PCI-E was the culprit..

Here we go:
--------------------------------------------------------------------------------------------------------
FSB: 400MHz
PCI-E: 110MHz
1:1
vcore: 1.425
vmch: +0.3v
vpci-e: +0.1v
vdimm: +0.4v
vfsb: +0.2v
--------------------------
Posted, windows loading screen, resolution change - kernal start..
CRASH
--------------------------------------------------------------------------------------------------------

reboot & changes:
--------------------------------------------------------------------------------------------------------
FSB: 400MHz
PCI-E: 110MHz
1:1
vcore: 1.450
vmch: +0.3v
vpci-e: +0.1v
vdimm: +0.4v
vfsb: +0.2v
--------------------------
Posted, windows loading screen, resolution change - kernal start..
SUCCESS



http://jadefalcon.shackspace.com/Pics/Computer/32_boot_001.JPG
http://jadefalcon.shackspace.com/Pics/Computer/32_boot_002.JPG

sorry for the res.. fresh format, only mspaint.

--------------------------------------------------------------------------------------------------------
now, stress testing...FAILED @ 17 seconds in orthos.
developing...
 
temps are probably 10c higher than reported in speedfan, my core 1 and 0 are at 22-23ish right now, and I know they arent that cold.

Anyways, I had a similar problem, my 3ghz on my e4300 would boot fine on the gigabyte ds3... anything higher, than it would either not reboot, or reset to stock settings..

Anyways, I left it at 3ghz to boot windows, downloaded gigabyte easytune 5, had it boot load 3.2ghz on start up and I am golden :)
 
aZn_plyR said:
temps are probably 10c higher than reported in speedfan, my core 1 and 0 are at 22-23ish right now, and I know they arent that cold.

Anyways, I had a similar problem, my 3ghz on my e4300 would boot fine on the gigabyte ds3... anything higher, than it would either not reboot, or reset to stock settings..

Anyways, I left it at 3ghz to boot windows, downloaded gigabyte easytune 5, had it boot load 3.2ghz on start up and I am golden :)
Click to expand...

you can see coretemp is running as well, and my thermistor agrees with coretemp within 2c normally, I'm water cooled btw.

Should I even bother with easytune for overclocking?
 
cerebrex said:
Should I even bother with easytune for overclocking?
Click to expand...



Eyewww....under XP it would probably work OK, but under Vista, it's a no go. You appear to be running XP though...try it out...



Have you gone into the ram timings and manually set them?
 
TheRapture said:
Eyewww....under XP it would probably work OK, but under Vista, it's a no go. You appear to be running XP though...try it out...



Have you gone into the ram timings and manually set them?
Click to expand...

yes the basic 4 settings, the sub timings are what I haven't played with.

I was running Vista but the overclocking was giving me issues.. and I couldn't decide if it was vista, or my hardware, so I decided to go back to XP to verify.. then I plan on going back to vista.

im going to run a few more tests and then up the vcore.
 
With my previous X2 and now this rig, I found that Vista vs. XP was no difference as far as overclocking went.

I wish you had some different ram to try out to eliminate that possibility. Have you ran the timings way loose?
 
FSB: 400MHz
PCI-E: 110MHz
1:1
vcore: 1.475
vmch: +0.3v
vpci-e: +0.1v
vdimm: +0.4v
vfsb: +0.2v

Orthos 12 minutes FAIL

Load temp peak 46c - increasing vcore..
 
Forget easytune.... For Vista or XP I like setfsb.... I find it more stable than crapgen...

Glad you got your problem solved... Ive been running at cpi-e at 110 and Ive had no problems, I was running 120 at first but I find it corrupts my HD but now its working good....
 
FSB: 400MHz
PCI-E: 110MHz
1:1
vcore: 1.4875
vmch: +0.3v
vpci-e: +0.1v
vdimm: +0.4v
vfsb: +0.2v

Orthos running
 
drop your vdimm. There's a good chance you're burning up your ram, causing instability. It'll also cause degradation if you do it enough. Look at vcore2 in speedfan. 2.34v. Change your vdimm to +0.3v, or even +0.1v or +0.2v for now. It looks like it's setting 0.15v higher than what you select.
 
I did'nt read the whole post, but isn't PCI-E @ 100 safer for a good boot; my DS3 has all kinds of problems if it's above 100. 110 pci-e is a pretty big overclock on this buss.
 
ziddey said:
drop your vdimm. There's a good chance you're burning up your ram, causing instability. It'll also cause degradation if you do it enough. Look at vcore2 in speedfan. 2.34v. Change your vdimm to +0.3v, or even +0.1v or +0.2v for now. It looks like it's setting 0.15v higher than what you select.
Click to expand...

The ram is actively cooled, case has a 25cm fan.. using my temperature probe (finger) the memory is case temp, remember Ballistix is 2.2v memory.. but it is running high, I will drop it down, thank you for the information, you've been an unbelievable help!

Also, I can't find ANYTHING on the net to configure speedfan for my board, is there a key anywhere to what everything is? temp1, temp2, temp3, etc... and the voltages?

btw Orthos STABLE 22min - load temp 49c ---eeek getting hot now.
 
You must log in or register to reply here.
Back
Top