HELP: Overclock of 2400+ Mobile XP

[Spudd]

Just upgraded some parts of my system. Installed the Corsair PC3200 ram first and ran that for a while at 266Mhz without a problem. I then installed a Mobile 2400+ with a Speeze WhisperRock II HS/F. The system FSB is set at 200Mhz.

The computer will boot up, after a couple of restarts, into Windows Xp PRo. After I login, the computer crashes! I can't get to the desktop or even hit Ctrl-Alt-Delete to open task manager. I tried to boot into safe mode but that thing just hung there forever.

The mobo is an NForce 2 board (by Epox, model 8rda+). Right now the vcore is at 1.61 volts. Is that the likely problem? I have it set at 10x 200FSB = 2 Ghz. What would you recommend? Thanks in advance. :beer:

 

[Acanthus]

Increase the Vcore to 1.75v and try again.

Watch your temps and voltages... What PSU and whats in the rest of the rig? You may be overloading your power supply which will cause power sags and make the system unstable.

Also windows instability can be caused by incorrect memory timings or memory overclocking. Check your memory dividers and timings.

 

[SickBeast]

It definately sounds like a memory issue to me. Relax your timings and try again. If that doesn't work, bump up the vcore to a maximum to 1.8v unless you have stellar cooling.

 

[Spudd]

The rest of my rig is in my sig. I'll try bumping the vcore to 1.75 like you said, because the memory timings are set at what Corsair rated them at 2.5, 3,3,8. I haven't touched any vcores on anything. I have the AGP/PCI locked just in case. I'll report back.

 

[Acanthus]

Is the corsair PC3200 or faster? 200fsb at 1:1 is 400mhz for the ddr. If its 2700, you will have to try VERY relaxed timings and hope it works, or use a divider.

 

[Spudd]

Yes, it's Corsair PC3200 ram. Right now, this reply is being written from the rig in question. All seems to be fine thus far. It seems it may have just needed the bump from 1.61 vcore on the CPU, to 1.75. So far so good. Would you recommend I run a test like memtest86 or something, or just let it "burn" in and just surf the net and stuff? Thanks again folks!

 

[SickBeast]

Prime95. Overnight.

 

[Spudd]

Damn thing froze on me in Win Xp Pro as I was writing an update about 5 mins ago. Went back to bios, and set the Vcore 1.75 again, but changed the BIOS settings under Power Options to Expert, and FSB/CPU clock setting to Aggressive so it would use my settings instead of trying to default. Some strange things I've noticed. On those boots that go through, when I get to Win XP desktop, ZoneAlarm pops up saying it found a network, but as a NEW network. AND all my programs from the zonealarm list are gone and have to be manually allowed and set all over again--so I've disabled Zonealarm for now. It didn't seem to get along with Panda Titatnium Antivirus 2004 anyway. Might dump both. I'm going to go find Prime 95 and run some tests again. (Note: this is all being written from the overclcoked rig).

 

[Acanthus]

Definately get a motherboard monitor program and check your voltages and temps @ load.

 

[Falloutboy]

bump you ram volatge up to see if that solves the problem

 

[Spudd]

Well, it's been running stable now since 8 pm. MBM5 has been monitoring the temps. I've been running DScaler with most demanding filters on so the CPU has been at 100% usage now for over 30 minutes. Temps are 47 C for CPU, and 31 for case. I have the alarm on the motherboard set to do an emergency shutdown at 63 C on CPU.

 

[Acanthus]

not bad at all!

 

[jagec]

It's topping out at ~2 GHz? That sucks...did you check MBM to see what Vcore you're ACTUALLY running? Could be your board undervolts.

 

[Spudd]

No, 2Ghz was just the most conservative setting I've set it at. Once I've got the stability going, I may try to push it to around 2.4Ghz. Right now, it's 11:37 so the computer's been running stable for more than 3.5 hours, at an average CPU temp of 40 C, and case temp of 30 C, at 1.75 vcore, 200Mhz FSB, with a clock multiplier of 10. So far, so good. Only thing left for the system upgrade this year is a 9800 Pro. Hehe.

 

[Avalon]

Was the 8rda+ a 200mhz FSB board? I was thinking that perhaps some revs of it were only 166mhz, and that may be the reason you are having instability. If that's the case, see if there is a bios upgrade for the board to allow 200mhz FSB. If not, I'm not sure.

That's very weird. Your mobile should not need 1.75v to do 200x10. If I recall correctly, in a laptop setting with Powernow! enabled, these mobiles had a range of up to 2ghz if the system needed that much power. That means the chip should be able to run 2ghz at 1.45. I may be wrong about the range statement, but either way, you definitely shouldn't need 1.75v to reach 2ghz. I'm thinking 1.55v max, unless your CPU is a dud, or the scenario I presented above is true.

 

[Spudd]

No, I lucked out because my board's revision of the 8rda+ has the 200 Mhz FSB option in the BIOS already. Everything's running fine now. I'll post some more results tomorrow. Thanks for the help folks.

Edit: as for the vcore setting, everywhere that I've seen someone posting their overclocking a mobile athlon, their vcore's have been above 1.6, usually around 1.75.

 

[Falloutboy]

at 2ghz it should need that much voltage though. at 2.4+ yeah voltages of 1.7+ are needed, hell I'm running 2.05 to get 2.7 out of my chip. thank god I have a pretty good water cooling setup or I could fry bacon on the core

 

[Spudd]

Ok, so far this is as far as I've gotten it: 200 x 11.5 = 2.31 Ghz @ vcore of 1.75, but my board (8rda+) overvolts, so according to MBM5 it's actually running at b/w 1.78-1.81 vcore. Temps are 38 C for CPU at idle, and 30 C for case. FSB is running at 400Mhz too (200*2). I tried running at 200 x 12, but it wouldn't finish loading windows. Tried dropping below 1.75, and it would either go into a reboot cycle, or just hang and scramble the screen when trying to load windows. So, right now I'm just going to run Prime95 and then DScaler to make sure it's stable. I'll report back later.

Edit: Just for the record, the dashboard of MBM5 says Core 0 vcore = 1.57, and Core 1 vcore = 1.78. :beer:

 

[Zap]

The 8RDA+ uses the original Nforce2 chipset which does NOT officially support over 166MHz FSB. You may want to try bumping chipset voltage a bit, or if BIOS doesn't give you that option, then run at 166MHz FSB and up the multiplier a bunch.

 

[Spudd]

Doh.

 

[Spudd]

Well, the computer just started acting up on me again. I was trying to play a game of Warcraft 3 with my friend, and the game crashed twice giving me an error of "Referenced memory could not be read at bla bla." This was after the forced update from Blizzard (hadn't played the game in months).

Also, while watching DVDswith PowerDVD XP, prior to the game of Warcraft 3, the computer would reboot half way or so into the movie. One time, it didn't reboot but rather the movie turned to garbled junk-- there were multi colored big pixels jumping around and alot of static being displayed instead of the movie.

Anyway, previous to this I ran memtest86 and on test 5 it gave me 4 errors. After the WarCraft 3 crash today, I ran Microsoft's memory tester 3 times on all 11 tests. Zero errors. Weird. Ran memtest86 again, all tests, 2 passes. Zero errors. Eh??

Currently running prime95 on Stress Test-->Torture Test--> Blended overnight. Current temps under this test have average 28 C for case, 53 C for CPU. CPU is running at 2304Mhz w/ 400Mhz FSB, with a Vcore setting in BIOS of 1.8 volts, but showing up as 1.81-1.87 in Mother Board Manager 5. +5.00 voltage averages +4.81, +12.00 voltage is averaging +11.92, and +3.3 voltage is averaging +2.54. The PSU is a Fortron 350W. :beer:

 

[VIAN]

If you have the original Nforce2, it is unstable at 200MHz. You would need Nforce2 400 chipset.

I would try and get your memory straight first. Try running your system non-overclocked and run prime 95 for however length of time you think is good. I choose to run it for 8 hours because I may never have the system run 8 hours under stress, max of about 6 and 2 hours for headroom is nice.

If the system is stable then that is good, if it isn't then that is bad, maybe your RAm is messing up.

If it is stable, find a good speed of RAM/FSB 1:1 ratio and test it with prime95 benchmark with each 5MHz incriment, don't forget to adjust memory timings as needed. After making sure this is stable by running prime95 torture again, then move onto the CPU.

This way of testing may make it easier to target what could be wrong with your system.

 

[Spudd]

Originally posted by: VIAN
If you have the original Nforce2, it is unstable at 200MHz. You would need Nforce2 400 chipset.

I would try and get your memory straight first. Try running your system non-overclocked and run prime 95 for however length of time you think is good. I choose to run it for 8 hours because I may never have the system run 8 hours under stress, max of about 6 and 2 hours for headroom is nice.

If the system is stable then that is good, if it isn't then that is bad, maybe your RAm is messing up.

If it is stable, find a good speed of RAM/FSB 1:1 ratio and test it with prime95 benchmark with each 5MHz incriment, don't forget to adjust memory timings as needed. After making sure this is stable by running prime95 torture again, then move onto the CPU.

This way of testing may make it easier to target what could be wrong with your system.

Does it matter that in CPU-Z it shows my chipset as NForce 2 Ultra 400?

Anyway, I ran Prime95 overnight for over 12 hours on Torture Test, blended. Average CPU temp was 53 C, case was 28 C. Absolutely no errors found....

Could it be that Panda Titanium 2004 is causing the reboots and crashes and what not? When I installed it a few months back, it would cause BSODS and random reboots. One strange thing was that when I got this chip and overclocked it, when I successfully made it to the desktop, Zonealarm Pro had forgotten what network it was on, and all of the program access list was erased! Argh......I'm going to dump Panda for now and see if the computer runs into any reboots or anything again. I don't get it...should I leave the Vcore at 1.8 in my bios (since it shows up as 1.81-1.87 in MBM5?) Or should I pull it down to 1.75 since it seems to be overvolting by ~.5 volts? Thanks for the help!

Edit: By the way, when I stopped Prime95 this morning, the CPU temp at load went from 53 C to 38 C idle, so the cooling looks to be ok?

 

[Spudd]

Played a game of WarCraft 3 FT, and after 20 minutes of gaming I got dumped back to the desktop with this:

Fatal Error
Exception: 0xC0000005 (ACCESS_VIOLATION) at 001B:6F61EB8E
Instruction at 0x6F61EB8E referenced memory at '0xFF00000C'
Memory could not be read.

What the hell does that mean? Is the Corsair PC3200 memory I just bought bad? Looking at all of my PSU voltages, I notice the +3.3 is hovering at 2.54

 

[pelikan]

Since prime95 is okay but 3D crashes, you have a chipset issue. Your FSB is too high. You should find your max stable FSB. Back it down by 5 MHz at a time, starting at 195 MHz. Keep a low multi that you know works. Play some games for an hour or so. When you find something that seems stable, loop 3dMark 2001 overnight.