Hi, im new to the forums. I was told that i could probably have some of my questions answered on here, so here it gos. I was overclocking my videocard the other day, x1600pro 256mb from sapphire. When i tried to up the ram my vidcard frizzed out and i had to reboot. When i rebooted all my games had no real vizualizations, the screens where all Blurring with no solid forms on them. I uninstalled my card reinstalled and took off ATITool which was my overclocking software. Now, all of my games run half the fps they did, and even when i turn down the graphics to nilch they are getting the same crappy frames. Even some of my movies look like a rainbow, all RGB spectrum is what it looks like. Either way i hope somebody can give me some sort of answer besides get a new vid card.

2.8p4 533FSB
1gig 333 pc 2700
250gb WD hdd 7200RPM
20gb Maxtor 7200RPM (POS)
ati x1600 Pro 256mb

overclocking will burn out cards, memory, and CPUs. This a risk you take when overclocking.

I NEVER recommend overclocking. I learned the hard way.

yah, but if i burnt something out wouldnt it have to mess up the card so i cant actually use it? like with a mobo, if you have bad ram or cpu it wont even post. either way is there any way to reverse my mistake?

No way to reverse internal heat damage, which is what sounds like what happened. The only thing you might try is remove the HSF on the GPU, clean off any paste with alcohol and apply some new ArticSilver, reinstall HSF. It's a shot in the dark.

ok thanks guys for the help, looks like im going to get a new cooler and some hsf so wish me luck.

Read my signature...

Athough OCing is fun, but never do it unless you can replace it.

I think the main thing to worry is NOT to increase your voltage too much.

Read up on how OC can damage your hardware, how the hardware actually works (high school Physics a minimum here) and how overclocking changes these things. Knowing all these will reduce the chance of you doing something that permanently stuffs up your PC.

Actually, knowing how they market chips help in knowing how much you can overclock too. Not sure how much of this information is actually freely available on the net, and I'm not too ready to type out a lecture.

I overclock my comp, and so far I've spent quite alot replacing parts. They didn't fail on me tho, just that I realise that to get a good overclock, one needs good parts

I think OCing the CPU/FSB then underclcoking the mem etc to get 2.4GHz from 2.2GHZ on a 64 CPU is really pointless and on stock cooling is dangerous. I got no performance increase with my OC as stated above measured as FPS in games like CS:S, HL2, COD2, Doom3. I did notice a slight improvement in 3dMark05 when I OC'd the 6600GT 128Mb, PCI-E. stock is 500/1000 and I took it to 575/1125. but after a few hours of play I get the occasional distortion as heat sets in (bc of stock cooling on the card). Now I backed it off to 560/1075 and it's ok now.

A point to remember is the voltage might make it stable but then you are just adding energy to the GPU or CPU etc, which in the end means heat. And it might be ok now,ie cooler months, wait until summer when/if your place warms up (i.e not great A/C).

Cheers.

ok, so i think there is something else the matter. not the video card. when i put in my x700 i get just as bad a frames, even though i used to get great frames in all my games. i never overclocked it, but it seems to have the same problem. so i have to keep investigating. thanks again for all your advice!

Actually, getting a high FSB is NOT pointless at all. Your FSB is like the mail system, the faster mail gets to its destination, the faster the destination can start working.

Therefore, a higher FSB will prove to be much more beneficial than a higher clock.

If I'm wrong, I just blew 200 bux upgrading RAM and another 120 bux upgrading mobo to handle high FSB speeds.

And if I'm wrong, I can't explain the nice 25% increase in 3dmark2k5 points when I upped the FSB, compared with a measly <1% when I upped my clock by 35%.

Then again, I must say it all depends on your hardware. If you're running a GeForce MX2 with a nice Athlon FX-60, doubling your FSB and clock will not show any (if at all) increase in benchmark points or FPS.

chad420: Maybe with X1600Pro OC episode you damaged/overloaded the power supply. Check your voltages with whatever software you like (Everest for example). What are the specs of PSU. Make/model, current rating on the 3.3, 5 and 12V outputs?

Check that you uninstalled the vid drivers correctly for the X1600Pro. Reinstall the X700 drivers.

CMH: I'm not saying OCing the CPU is completely useless, just maybe not worth the risk. The FSB is the system bus clock freq and you're right this will determine the data transfer to all components and this is where things can go wrong. Some mobos are kind to this process others like my @#$%ing world's most expensive motherboard MSI K8N Neo4 SLI Platinum seem to be lame ducks w.r.t this. If I up the FSB, keeping the CPU multi at 11 and take the HT multi to 3 to keep the HTT bus at 1GHz or less, blah blah blah...the increase in FSB seems to screw up the PCI bus so I loose my ethernet. I can't seem to bring my FSB above 218-200Mhz, not matter what I do with RAM timings, HTT bus etc. without crapping out the PCI bus and even the PCI-E bus. So I'm just bitter and all OCer's can go to H E double hockey sticks! :haha: Just kidding. But seriously alot of mobos have these limits and the end result is usually a trade off between an increase in CPU speed and a decrease in RAM frequency. In the end it depends on what the CPU is doing. If it's doing work without alot of RAM use then this will help, but add RAM use, like video stuff and you end up close to where it was b4 the OC. IMO

Cheers.

o.O

I know upping the FSB alot will cause some components to fail (I know my Soundcard is kinda shaky now, and I'm not sure my current internet problem is due to this). I know for sure that SATA drives cannot take anything past 275mhz. You'd be pushing it at 275.

Either way, OC is not something anyone should be toying with if they don't know jack o. Millions of things can go wrong...

Sorry about that mobo, sounds like some clever marketing snagged ya.

Actually I read a raving review for it at HardOC which compared the top SLI boards and especially a head to head contest between the ASUS top brd A8N SLI and the MSI Neo4. I was going with ASUS until I read this acticle and saw the bench marks etc. The day b4 my parts were to arrive, I went bk to read the article again and salivate about my great mobo and they had updated the 1 month old article with a retraction, :unch: They had found probs (undefined) with the production version and OCing that MSI promised they would correct but didn't/couldn't. sigh. Cheers CMH

I wonder where Chad420 went? Did you fix your problem?

Ouch. Thanks for sharing. I usually trust production modules to be better (or at least as good) as test modules. Now I know better.

The risk reward Ratio

Overclocking can be a risky business, but it is a real rush when you hit high clocks, makes you feel GOOD. Runnin' faster than you.....doin' F.E.A.R. @ 12,000 fps...lol all good fun, unfortunately there is the other side of the coin:
The "What's that smell" side. When your computer turns into a toaster and all you cards are just a little too well done. Can you spell RMA? Great if you have a waranty on your stuff, you just go through withdrawal 'till the new stuff arrives.
"What's that"???? "Not covered for that?????" "You won't rma my stuff ???"
Well I needed a new system anyway, this old one was waaaaaaayyyy tooooo slow. Now where can I get the bucks???
Simple decisions on whether or not too push you stuff to the limit. If you have the wherewithall to support the little mistakes that can be costly then go for it. Ya gotta ask yourself"do I feel lucky?" Well do ya Punk????, oops sorry, watching a little too much Clint Eastwood.
Risk reward ratio No risk, no reward, big risk, big reward.
I advocate always going for the big risk, life is too short, eat dessert first and clock it till it sets off the smoke detectors in the room then back off 1%.
And don't forget water cooling, doesn't just cool GPU's and CPU's, it's also useful for putting out fires.
Sorry for the raving today

I think that is not the main risk. It is the gradual failure of parts that you dont even notice till one day they are shot that is the real risk. I burned up my old x600 due to OCing. I tweaked memory timings and such, only bumped up the clock speed a little bit.
I pushed it to the max, it was very stable.
But over time the card started to not work on a cold boot. Then it died.

Got my x800, and have not even thought about getting atitool and ocing.

ok i downloaded everest and i cant find anything about the PSU, everything else seems like its ok though...

spade, don't scare me. gradual failure? I've talked to so many overclockers, and their rigs are all fine after 2-3 years. How long did it take for your x600 to burn up?

And how much did you tweak mem timings? Tweaking mem timings isn't a small overclock as far as I know, I can't tweak my RAM's timings at all, even at 400mhz (and its rated 550).

Chad420: It's in the motherboard section, you should see temps, RPMs of fans and voltage levels and the name/make of the sensors.

I think Spade hit it on the head with

This is EXACTLY how the real world actually works. I will tell you briefly about my previous job with Nortel.(they are a world class telecom company that made THE 1st 10Gbit GaAs laser diode and detector) These are used worldwide in internet fiberoptic telcom. Ok get to the point Kirock.

I was the Reliability Test Lab Supervisor, it's manditory in Telcom (legal, license, blah, blah, blah stuff) to provide lifetime and failure rate data, These are 100 page reports which prove by real experimental data the life expectancy of the product.

How do we do this you might ask? HEAT, HEAT and a little more heat and current! We would apply 3 to 4 times the normal operating current while they were mounted inside a chamber at 100C degrees. This LITERLLY accelerates the failure rate of the device (the laser diode in this case). So something that would normally last 15-20 YEARS in the field in normal use would run it's life expectancy out in just a few 100 DAYS!!!!!!

We would measure performace parameters, like power in vs. light power out etc. to measure and then predict the MTF (mean time to failure). You could plot this data and see it slope away to zero. This is done stastically with populations of a few dozens to over 100 samples. The more devices (samples) you use the Greater your Confidence is in the MTF prediction. Again this is DICTATED by LAW and our legal requirements as manufactures of telecom devices( I mean the sample size and the level of confidence and varies per product). ALL such companies (Alcatel, Siemens, etc) must provide and prove this data.

What the hell is your point KirocK? Glad you asked. Heat is the whole point. The worlds best manufactures in Telecom devices use heat to kill devices in a short period of time so they can provide proof of the reliability of their products.

:wave:

Cheers.

lets see here.....Field Value Core Voltage 1.5 V. Thats under the Chipset. Field Value
Core Voltage 1.475 - 1.55 V, thats the core/IO voltage. Under CPU.