Gigabyte P35 DS3L not holding overclock overnight

[Billbo]

No one seems to know what causes it. I think the DS3R's (and maybe other P35's) can suffer the related endless reboot cycle, where it just keeps rebooting as it tries to find memory settings that will work for the overclock you've set. A guy posted a reply from GB in another forum. They said it seems the newer bioses may not like certain memory, and the only thing to try is to flash an older bios. This is probably similar to what SerpentRoyal said about trying to find a bios that's friendlier with your setup.

For me though, that's all hogwash. GB advertise these boards as "overclocking enhanced". That's what we should be getting for our money, no? Most of the people having these problems aren't using "no name" parts. So I just don't understand it.

 

[nevbie]

I wonder what the differences are between rev 1.0/2.0/2.1 boards..
Is anyone using memory from the recommended memory list and still encountering problems..? (which in many cases means kingston, hah)
I'm going to build a system one day.. and these P35 DS3* boards seem to be the most interesting choises otherwise.

 

[mingsoup]

Well, I've tried F7,F6, and F5. None of them will hold. I've tried clearing CMOS as well as trying very very modest oc'es. Nothing will hold, yet everything I've tried is perfectly stable. I emailed Gigabyte about the issue.

 

[Billbo]

I was just about to post a huge "this is exactly what I did" but chickened out. I really think the solution is to stop fiddling with the bios. At first that's what I was doing, and it just made things worse.

I think your best bet is to leave it run a few days at stock and be certain it is stable in all situations - booting, resuming from standby etc. When you decide to try OC'ing, don't just increase the FSB to some stupidly high value, even if it has ran at 3.6gig are whatever previously. Do it the old-school way with small 5-20MHz increments. This allows you to spot when a bump in vcore may be needed. If you test the memory first while lowering the cpu multi, you know the ram will run at 400MHz or whatever you're striving for so you'll never have to touch the memory settings again. So it only leaves FSB speed and the voltages to play with. I actually increased the MCH, PCIE and FSB volts by 0.1V as it is supposed to help with stability. But I don't see why they'd have to be adjusted from that unless you're trying for a huge OC.

Does anyone think they could actually give this a go, so we can see if it truly does bring these boards back from hell, or it was just a fluke for me?

 

[toadeater]

I think the instability, or incompatibility, is triggering some kind of safety feature in the BIOS. The BIOS thinks the board failed to boot last time, so it goes back to defaults. I don't think it's BIOS corruption, because then you should lose more than your OC settings. Hmm, maybe Windows isn't shutting down properly? Try a Linux live CD, see if the OC sticks after a few cold boots with that.

Maybe something is timing out, or not responding properly? Maybe try turning off as much of the "auto" features on this board as possible and make sure you set everything manually to correct settings. It sounds like some kind of glitch to me.

 

[Billbo]

mingsoup, what did you say to Gigabyte in the email?

Just got a new reply from them saying once again they've never come across the problem and have never had complaints from anyone else. That is BS.

I urge you all to send a support request to them as they don't seem to be taking it seriously. I'd just like to find out what the problem is. Whether it's a compatibility problem, a bios problem or whatever, it'd be good to know.

See another person started a thread about the DS3P doing this as well.

 

[mingsoup]

Well they replied.....

"We do not guarantee any results from overclocking"
They also asked how high I had set my settings and whether I had any other issues besides oc'ing. They suggested lowering the settings. But that doesn't help. The board won't keep any oc I've tried. Currently I'm trying a 90Mhz oc, with auto voltage. If that doesn't stick...nothing will.

I told them my OC settings would not hold overnight or on a cold boot. I explained that these settings were perfectly stable in windows. I also said that quite a few other people have this same problem. Thats the main gist of it.

 

[Billbo]

Ok. Pretty much what I said. The point is the only thing being overclocked on my board is the cpu. The memory is still running within it's specs. The DS3L is designed to run at much higher settings than I'd ever be willing to go. So I don't understand why it happens. And why it will work for a while, then not????

 

[TomConsole]

Hi all,

Just found this thread since I am having the same problem with my new P35-DS3L board. For the couple of days it was fine, but after that if the pc is left off for more than an hour or so my bios settings are lost. I am relativly new to overclocking so thought I had damaged one of the components at first, but after reading this thread obviously not. As with many of you all my OCs are as stable as can be in windows and I am able to restart or even switch off for a period of time without losing my settings....but once it has been off for a while I lose my settings.

I assume the majority of these boards are ok, I think my plan is to order a new one then once received do a straight swap of all the parts then return the old one....unless the company is will to do a swap through the delivery company...I dont fancy having all my other components sitting around waiting for a new board!

Tom

 

[nevbie]

How about FSB underclocking? Will the settings still reset? If so, then you could complain about that to Gigabyte.

 

[mingsoup]

Well that is frustrating. I can't make it work. Gigabyte has basically told me to F off. I just LOVE being f'ed again and again by consumer electronics giants.

FSB under clocking.....interesting.

 

[Billbo]

I have always been a Gigabyte fan and have received ok but slow support from them in the past. But yeah, the way they are treating me about this problem is absurd. Even after telling them I'd like to help in any way possible to find out what the issues are they don't care. And to continually say they've never heard about it from anyone else is ridiculous. I think there are too many people with these problems for it to just be a few bad boards.

nevbie, my DS3L played up a couple of times just by setting the FSB control to manual but leaving the FSB at default 200MHz. This should not happen since it isn't an overclock.

 

[TomConsole]

That's interesting that it fails even when just setting to manual. Its very strange that they won't accept it as a problem...especially how they even go to the efforts to call it an overclocking board on the box! I just hope the company i ordered from sees it as a fault and sends a replacement out for me...otherwise I guess it will just be back to the search for a good p35 board! Im not sure whether to get a refund to buy a different make or just to risk a replacement...I just don't want to encounter the same problem a few months down the line.

Just a thought but next time you email Gigabyte maybe you should link to this thread along with the many others on different forums...then at least they can't say they've never heard of anyone else having the problem!

 

[mingsoup]

I tried setting the multiplier low....and the FSB high to get a 1.8Ghz again. That held just fine....if I simulate long power off via switching the PSU on the back...it will hold with low multi, high FSB.

Anyways, I'm back at F7 now. I can't help but get the idea that its my memory that is doing this. The post always gets stuck at "Memory Test" "Tes" or some variation. Thats when it resets on cold boots.
Here's the memory I'm using. Once again, my oc's are perfectly stable in Windows. I did notice that there are approved memories for Gigabyte boards.....I don't think mine is one of them.
Anyways here it is: G.SKILL 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory

 

[Billbo]

TomConsole, I told them yesterday about this thread and a few others at other forums.

mingsoup, why do you say it could be the memory when you tested it at high FSB? Did you actually run memtest/orthos on it at those settings?

 

[Billbo]

Well I'm back at 3.0GHz today. No problems so far with standby or anything. Memory running at 416MHz or so. Getting 20.8s in 1M Super Pi test which is comparable to others results. I'll leave it at these settings for a few days and see what happens. But I really think running the system at stock for a few days got rid of the gremlins.

To the people still having problems, have you ever overclocked with settings on auto by chance? Because I don't know what Gigabyte programmed other things to do on automatic, but vcore goes way over the specs with auto settings. I know that trying to increase the ddr volts manually made my system less stable. Maybe it increases volts or uses too strict NB latencies which may initiate the instabilty.

 

[RapiDJunioR]

hi dudes

im new in this board

i have got same problems with yours.

i can't overclocking with my system.

e4500 m0 rev
p35 ds3l rev 2
kingston ddr2 800mhz
fsp bluestorm II 500W


what i changed in MIT and start computer. while booting cmputer didnt complate the memory testing. after restart and computer started with default settings. i tryed clear cmos, put out battery, put out all cables(power, ide , sata) and waited 8 hour. but anything changed.
i tryed that with f5 and f7 bios.

anybody help me about that problem.

 

[Billbo]

What did you change in bios? Is memory set as per spec? Are voltages on auto or manual?

 

[RapiDJunioR]

first i disabled cia2 eist tm2 c1e. after i try auto voltages and +1 +2 voltages(cpu 1,3250 def , 1,3750 i try). i try all. rams are in 400mhz spd2 4 4 4 12, auto 800mhz 5 5 5 15.

then i try clear cmos and only change fsb. 200 default. i try 210fsb but again back default after post screen.

 

[RapiDJunioR]

i tryed my oldu cpu(p4 506) with oc. but same problems are running.

 

[Regalk]

Over the past few days I have been trying to see if I could emulate these issues above - and my observations
1. Try not to get overly aggressive with RAM particularly OCZ rev B (I hate this memory BTW) - stay close to specs
2. make sure you are not setting voltages to auto (and sufficient for the CPU)
3. Use standrd for memory timings (not Turbo or extreme)
4. make sure PS is at least 500W

Using Bios F6 for a while now with an E2160 now E4500 running at 412 X 8 - no issues even after a cold boot (PS switched off overnight). My board is now very abused
Swapped several HDs - SATA/IDE, Video cards back and forth and still no issues.
Yep definitely holding OC until you start screwing with memory timings
The only issue I have is double boot
I did change my chipset fan with a Thermaltake Heatpipe one - don't think that really mattered
I have seen more issues with the ABIT IP35-E (running a Q6600) in that and in the near future I will be tossing (darn thing refuses to run over 400 FSB)

 

[RapiDJunioR]

thanks for reply
but i can't doing 1MHz overclock. i cant changed my fsb. default 200. when i try 199 and 201 . anything was changed. problems are running.

sorry for my bad english.

 

[Billbo]

Hey Regalk, when does the double boot occur? I'm also using F6 at the moment since it seems the most stable for me. Were you using the same 412x8 with the E2160? How much vcore did you need?

RapidJunior, did your board always have problems or did it overclock ok when you first got it?

 

[RapiDJunioR]

already i bought my board(with cpu and rams) before 1 weeks. i have this problem when i bought it.
but i can 1 time worked 10*250 ratio: 2 . and after i updated my bios to f7 and tryed overclocking but problems are back. after i back to f5 and waiting solution this problem.

 

[Regalk]

Originally posted by: Billbo
Hey Regalk, when does the double boot occur? I'm also using F6 at the moment since it seems the most stable for me. Were you using the same 412x8 with the E2160? How much vcore did you need?

RapidJunior, did your board always have problems or did it overclock ok when you first got it?

Sorry messing around with my Quad for the last 3 hours - cleaning up the PC for faster boot etc. - WinXP accumulates a lot of junk so time to houseclean
Yes the E2160 worked at 412 X 8
The board (and the IP35-E) refused to go beyond 417FSB. Like it is a wall there for both E2160 and E4500. I guess the low end boards aren't really up to snuff.
The double boot occurs randomly - on first start up (though not all the time) from a cold boot or sometimes on a hard reset. My DS3L is out of the case (on the box) for now - basically I have to scrap 1 of my 939 A64 systems. Otherwise it works really well. I have no real complaints about this board.
Sorry guys what can I say - works for me (as long as you know what your HW is capable of)