EP35-DS3R Won't Hold 400 FSB

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

You will have much better luck with Extreme Performance OFF

Does that setting do anything besides change the RAM timings? I have tried setting it to standard a few times already at various FSBs and it did not help booting. I'll try it again next time it fails to boot before doing anything else.

[Lsdmeasap]

Re: EP35-DS3R Won't Hold 400 FSB

I am not sure what all it does sorry, but I do know it may be using too tight of ram timings causing you to not be able to overclock very well. But I do know you want it on standard for overclocking which you are doing, and especially if you overclocking ventures are not holding or staying stable.

You could also be at the upper range of your CPU's limit. Have you tested your particular CPU in a different board/setup?

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

You could also be at the upper range of your CPU's limit. Have you tested your particular CPU in a different board/setup?

No, but I don't think it has reached it's limit because it was stable in Windows at 400x9.5 FSB and 1.296v (after vdrop). But then it wouldn't boot the next time. Once I get it to boot overclocked, there is no problem with stability whatsoever. I've ran Prime95+Orthos simultaneously, and tried a bunch of games and video. Not a single BSOD or crash, the board runs great other than the boot thing.

Right now I'm trying to get 385x9.5 to boot consistently at 1.31v (after vdrop). This isn't an unusual voltage for an E7200, what bugs me is that I've seen it running 100% stable all day at lower voltages and higher speeds. The boot sequence is the only thing forcing me to raise voltages higher. I'm going to try some other timings later, such as 450x8, etc.

Btw, I tried Standard and Turbo performance enhance again today when the board failed to boot once, it did not make a difference. So I raised the CPU voltage another step and it booted.

[Lsdmeasap]

Re: EP35-DS3R Won't Hold 400 FSB

Right now I'm trying to get 385x9.5 to boot consistently at 1.31v (after vdrop). This isn't an unusual voltage for an E7200, what bugs me is that I've seen it running 100% stable all day at lower voltages and higher speeds. The boot sequence is the only thing forcing me to raise voltages higher. I'm going to try some other timings later, such as 450x8, etc.

Ya, well that is someone else's chip and hardware you are comparing to, which you Cannot do. But yeah, along the lines you are thinking I think you just need more voltage. Try 1.325 after Vdroop/vdrop depending on which you meant. 1.325 is fine as well if you have adequate cooling.

I need to test out my 8 multi, thanks for reminding me!

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

Ya, well that is someone else's chip and hardware you are comparing to, which you Cannot do.

I'm not comparing it to any other single E7200, I'm comparing cold boot failures with stable overclocks in Windows that I got on this same board. Why is the BIOS turning off the OC if it was stable in Windows? I have gotten zero, absolutely zero errors, corruptions, or BSODs running at 400x9.5 1.3v, but it will not cold boot that way no matter what I try.

Heh, sorry to be pestering you, but Gigabyte tech support was of no help whatsoever, they just said they don't guarantee overclocks.

[Lsdmeasap]

Re: EP35-DS3R Won't Hold 400 FSB

Well then you likely just need more volts. 1.3 isnt much you can use more then that safely for sure. I have set several boots that are windows stable, but are not fully stable for 24 hours which is what I consider stable. What your board considers stable and will keep is all up to your board, sounds like you just need a bit more voltage. Why are you hesitant to give it? If it is because your CPU is getting hot, then you likely just need better cooling. Which you should get anyway if you want a stable overclock at lowest possible voltages.

As for it running it before and not not at a certain voltage, your processor will degrade over time and has a "break in: period. Maybe you just had to go thru the break in period, some need more after this, some less. If you have at anytime given it a large amount of voltage then you could have degraded it as well

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

Yeah, the reason I don't want to give it much more voltage is because I don't want to degrade it. It's definitely the CPU voltage that is causing the board not to cold boot. It just needs significantly more to boot than to run in Windows. This isn't a guess, I've tested it numerous times already. Is it a bug in the BIOS? It could very well be a defect in my board or some obscure hardware incompatiblity, etc. But I've read of other people having the same issue. Whatever it is, I don't need nearly as much CPU voltage to be stable in Windows as to boot.

One more example: yesterday I ran it at 3.8GHz (400x9.5) after a reset. Today it didn't boot--it actually went through five restarts before it booted with default settings. So I decided to try the lower multi this time (with performance at standard). 420x8.5 booted. I then set performance to Turbo, it booted again. Same 1.32v CPU voltage as before.

It looks like I need that voltage bump at ~3.5GHz to boot, and after a reset it can run at that same voltage at 3.8GHz, but will not cold boot next time. IMHO that's a pretty big FSB margin for this to be caused by the CPU and not the board or BIOS?

[GiggaByte]

Re: EP35-DS3R Won't Hold 400 FSB

I am having the infamous "OC lost after cold boot" problem with my board. During resets it will hold, but if I power off it will disable CPU clock control and revert to stock speed. It only does this at an FSB of ~375 or higher, no matter what voltages or RAM timings I use.

When I can get it to boot at 400x9.5, it is 100% stable in Windows. I can run Prime95, Orthos, and ATITool simultaneously while multitasking and I get zero errors, and low temps. I get no crashes in Crysis, CoD4, GRAW2, STALKER, and other intensive 3D games. If that isn't stable, what is?

This is after it finally boots:

CPU-Z Validator 2.1

I have done everything including cleared the CMOS, removed all USB devices, flashed to the latest BIOS I could find (which wouldn't let me change vCore afterwards, so I flashed back), disabled/enabled CIE and EIST, disabled/enabled legacy USB support, keyboard, mouse, etc., etc. Nothing has made a difference. It just won't maintain the OC past 375FSB if I power off. What is the difference between powering off and resets to this board?

If anyone has any ideas, let me know.

E7200 Core 2 Duo - reaches 3.8GHz @ 1.296v
EP35-DS3R - tried BIOS F2 and F3e
2x1GB G.Skill F2-6400CL5D RAM
MSI 8800 GT
550W Rosewill PSU

These are the settings I tried:

Robust Graphics Booster___________ [tried all three]
CPU Clock Ratio ____________ [8.5-9.5]
CPU Host Frequency (MHz)__________ [266-400]
PCI Express Frequency (Mhz)_______ [100-102]
System Memory Multiplier (SPD)____ [2.00-2.50]
DRAM Timing Selectable_______ SPD __ [Manual & Auto]
Performance Enhance = [Standard]

CAS Latency Time________________ 5
Dram RAS# to CAS# Delay_________ 5
Dram RAS# Precharge Delay_______5
Precharge Delay (tRAS)__________ 13-15
ACT to ACT Delay (tRRD)_________ 3-4
Rank Write to READ Delay________ 3-11
Write to Precharge Delay_________ 4-5
Refresh to ACT Delay______________ 3
Read to Precharge Delay__________ 4-5
Static tRead Value_______________ 12
Static tRead Phase Adjust________ Auto

System Voltage Control____ [Manual and Auto]
DDR2 OverVoltage Control__ [+0-0.2V]
PCI-E OverVoltage Control_ [+0-0.1V]
FSB OverVoltage Control___ [+0.0-0.2V]
(G)MCH OverVoltage Control [+0-0.2V]
CPU Voltage Control_______ [1.1-1.375V]

After I saw your post I tried lowering FSB to 375 and long behold my P35-DS4 passed POST when cold and ran normally. I also noticed once again a P35 board having this cold boot problem with the E7200. I keep trying to tell GIGABYTE there is a problem with P35 boards and the E7200.. sigh... I can put my 65nm E4500 back in and I can go to ANY FSB I want, even unstable ones and it always cold boots fine..

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

After I saw your post I tried lowering FSB to 375 and long behold my P35-DS4 passed POST when cold and ran normally. I also noticed once again a P35 board having this cold boot problem with the E7200. I keep trying to tell GIGABYTE there is a problem with P35 boards and the E7200.. sigh... I can put my 65nm E4500 back in and I can go to ANY FSB I want, even unstable ones and it always cold boots fine..

Thanks for the info. It could be? Come to think of it, my board has booted cold during a few tests yet was unstable in Windows. At the same time, it won't boot with some stable settings. That kind of rules out that this BIOS is detecting stability accurately, doesn't it?

If you want, also try the 8.5 multiplier, because it might let you raise the FSB higher and get slightly better performance. Right now I have 420x8.5 booting consistently, while 400x9 refuses to boot with the same voltage and RAM timings. I'd assume it was the CPU's fault if it wasn't for the fact that it is stable in Windows at 400x9.5 at the same voltage. Just make sure if you OC your RAM to adjust the voltage and timings.

I have already contacted Gigabyte as well, but they don't want to hear it. They said they don't guarantee overclocks and that was it. Hopefully their techs are aware of the problem.

Here are my current bootable settings, in case this helps:

Robust Graphics Booster___________ [auto]
CPU Clock Ratio ____________ [8.5]
CPU Host Frequency (MHz)__________ [420]
PCI Express Frequency (Mhz)_______ [102]
System Memory Multiplier (SPD)____ [2.00]
DRAM Timing Selectable_______ SPD __ [Manual]
Performance Enhance = [Extreme]

CAS Latency Time________________ 5
Dram RAS# to CAS# Delay_________ 5
Dram RAS# Precharge Delay_______5
Precharge Delay (tRAS)__________ 15
ACT to ACT Delay (tRRD)_________ Auto
Rank Write to READ Delay________ Auto
Write to Precharge Delay_________ Auto
Refresh to ACT Delay______________ 52
Read to Precharge Delay__________ Auto
Static tRead Value_______________ 8
Static tRead Phase Adjust________ Auto

System Voltage Control____ [Manual]
DDR2 OverVoltage Control__ [+0.2V]
PCI-E OverVoltage Control_ [Normal]
FSB OverVoltage Control___ [+0.1V]
(G)MCH OverVoltage Control [+0.1V]
CPU Voltage Control_______ [1.368V]

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

TEMPORARY FIX FOR COLD BOOT POST FAILURES:

Find your highest bootable FSB (mine was 365 FSB). Next, find your maximum stable FSB, RAM and voltage settings in Windows. Now reduce the FSB to at least 80 less than your maximum stable Windows FSB. Enable CIA in the BIOS to Full Thrust. Your computer will boot at the lower FSB, but when CPU usage increases in Windows, CIA will raise the FSB by up to 80, depending on load.

Example:

Bootable setting: 333x9.5, 1.375v (to compensate for CIA) = 3.1GHz
After CIA OC in Windows: 413x9.5, 1.328v (vdroop) = 3.92GHz!!!!

If your bootable FSB limit was 365 like mine, this means you can use CIA to overclock in Windows to 445 FSB. An added advantage is it keeps temps down when your CPU is idle, and it works with C1E enabled.

This is certainly a great hack, unless your CPU has a low multiplier, but I am still not happy with Gigabyte's faulty BIOS. Dynamic overclocking by CIA proves this board is stable at higher FSBs, yet it still won't boot at those FSBs. I hope Gigabyte fixes this soon.

http://www.ocforums.com/showthread.php?t=547533

http://forums.overclockers.com.au/showthread.php?t=668920

http://forums.anandtech.com/messageview.aspx?catid=29&threadid=2131249&FTVAR_STKEYWORDFRM=&STARTPAGE=1&FTVAR_FORUMVIEWTMP=Linear

http://www.ocforums.com/showthread.php?t=546768

http://www.xcpus.com/forums/intel/2395-gigabyte-ga-965p-ds3-e6600-cant-cold-boot-any-overclock.html

http://forums.whirlpool.net.au/forum-replies-archive.cfm/990450.html

http://forums.pcper.com/showthread.php?t=446828

http://forums.pcper.com/showthread.php?t=439190

http://www.tomshardware.com/forum/247416-30-strange-boot-recycle-problem

http://forums.ebuyer.com/showthread.php?p=261063

http://www.overclock.net/intel-cpus/257736-gigabyte-p35-d3sl-reflashing-itself-stock.html

http://forums.overclockers.co.uk/showthread.php?p=10405536

http://forums.tweaktown.com/f19/ga-p31-s3l-overclock-advice-26996/

http://www.houseofhelp.com/v2/showpost.php?p=356151&postcount=4

http://www.xtremesystems.org/forums/showthread.php?t=178265

http://www.hardforum.com/showthread.php?t=1239860

http://www.hardforum.com/showthread.php?t=1193054

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

I just installed the F3 final BIOS update, it seems to have fixed the cold boot issue at last. And all this time so many people thought their boards were broken. I knew it had to be the BIOS.

Its cold booted 3 times now with 4-4-4-12 timings and tRD 7, which is awesome for a low cost motherboard and $40 budget RAM! Good work, Gigabyte!

Oh, the DPC latency has improved as well, unless something like CPU-Z is used.

[Lsdmeasap]

Re: EP35-DS3R Won't Hold 400 FSB

Nice to see you are having some Better luck Toadeater!

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

Nice to see you are having some Better luck Toadeater!

Luck had nothing to do with it, it was the new BIOS that did it. It's been a couple of days now and the board has booted every time, so I think it's safe to say it is fixed.

The only other thing is that now Performance Level seems to actually do something. Before I could set it to Standard or Extreme and it didn't make a difference. Now if I set it to Extreme, it won't boot. So maybe with the previous BIOS some of the board timings were stuck at Extreme the whole time? Anyway, it works now. Thanks for your help.

[Lsdmeasap]

Re: EP35-DS3R Won't Hold 400 FSB

Well performance level always changed things, maybe you just never noticed. It mainly changes tRrd for Sure (Static tRead Value) and if you have subtimings manually set then for sure using the performance setting on anything other then standard may toss your timings all off balance

Anywho... Glad you are up and running smoothly

[Toadeater]

Re: EP35-DS3R Won't Hold 400 FSB

Well performance level always changed things, maybe you just never noticed. It mainly changes tRrd for Sure (Static tRead Value) and if you have subtimings manually set then for sure using the performance setting on anything other then standard may toss your timings all off balance

I've been watching the timings in Everest and Memset. With the F2 BIOS Extreme didn't change anything on my board that is monitored by those programs. With the F3 BIOS the DIMM Clock Fine Delay has different timings now. Standard has 12T, 11T; Turbo has 9T, 8T, and Extreme won't boot anymore. In BIOS F2, I always had 5T, 6T timings for my RAM, no matter what I set performance enhance to. Perhaps these timings are dependent on the SPD somehow?

I realize it might set some other timings as well, unfortunately there is no program that I know of to manually set Clock Fine Delay in Windows to test out different timings and see if this was the cause of the cold boot bug.