Is It a Good Idea to Go from 2.4 to 2.8-ish Ghz After Keeping it at 2.4Ghz for Over 1 Year?

[geokilla]

My X2 3800+ is clocked at 2.4Ghz. It's been running at this speed for around a year. Since I have some free time before exams start, I want to push it to around 2.8Ghz. However, it's been running at 2.4Ghz for so long would there be a negative impact on the CPU for the sudden increase in clock speeds?

Also, should I run Blend, Small FTTs or Large FTTs? Last time I ran Prime with Small FTTs for 24 hours and Memtest test #5 for 3 hours. This time, I'm planning a combination of something like 12 hours of Large FTTs and 8 hours of Blend, then 3 hours of Memtest test #5.

 

[THUGSROOK]

as long as you DONT add any voltage, it doesnt matter what you do.

large FTTs.

24 hours is a waste of time and electricity.
you can go for 24 hours and still not be 100% stable, so its a total waste.

 

[seanp789]

I'm not an EE but it shouldn't matter than you ran it at 2.4ghz. Running a CPU way out of spec in general will lower it's working life but by then it's time for an upgrade anyways.

 

[TC91]

keep the temps within reason (60c or so) and dont crank the voltage, and the chip should be fine.

 

[LOUISSSSS]

Originally posted by: THUGSROOK
as long as you DONT add any voltage, it doesnt matter what you do.

large FTTs.

24 hours is a waste of time and electricity.
you can go for 24 hours and still not be 100% stable, so its a total waste.

24 hours will show that your system is more stable than if you test for 1 hour, which is about as good as nothing...

24 hours of stability testing is not a total waste and is highly recommended.

if youre worrying about electricity, why are you wasting it on AT?

 

[geokilla]

Well 2.4Ghz is as high as I can go on 1.32V. I'm using a Cooler Master Hyper TX 3 so my temps are fine. Maxes out at around 53C (95% F@H) at the moment.

So I guess the next time I overclock, I should just start off strong assuming I have a good cooler and case?

 

[VirtualLarry]

I think that you can go as high as 1.45v or possibly even 1.5v. My X2 3800+ S939 chip did 2.75Ghz on 1.4v.

 

[error8]

Originally posted by: geokilla
Well 2.4Ghz is as high as I can go on 1.32V. I'm using a Cooler Master Hyper TX 3 so my temps are fine. Maxes out at around 53C (95% F@H) at the moment.

So I guess the next time I overclock, I should just start off strong assuming I have a good cooler and case?

I had this chip and the same mobo as you and kept it at 2,58ghz with 1,48 V, for some 6 months or so until I got my 4000+ X2.
The bottom line is that at least mine, was a crappy overclocker, the mobo too ( stupid 1,95V max on memory).
Keep the temps under 60 C and don't be that scared about voltages, since this is a 90 nm chip and likes high Vcore. If you're maxing out at 53 C at 1,32 V you really need to change your cooler.

 

[betasub]

Originally posted by: geokilla
I want to push it to around 2.8Ghz. However, it's been running at 2.4Ghz for so long would there be a negative impact on the CPU for the sudden increase in clock speeds?

LOL. What, like the gears will seize up?

How long it's been at 2.4GHz doesn't affect its chances of hitting 2.8GHz (possible, but probably requires plenty extra Vcore).

 

[PlasmaBomb]

It will be fine, although you might not get 2.8GHz

 

[Foxery]

Originally posted by: geokilla
Also, should I run Blend, Small FTTs or Large FTTs? Last time I ran Prime with Small FTTs for 24 hours and Memtest test #5 for 3 hours. This time, I'm planning a combination of something like 12 hours of Large FTTs and 8 hours of Blend, then 3 hours of Memtest test #5.

Some time savers - "Blend" means both large and small FFTs. You don't have to add a run to do them seperately. Memtest isn't necessary at all, unless your new RAM settings take it way above its own specs.

As to the original question, I can't see you hurting anything. Either it'll work or it won't.

 

[geokilla]

Originally posted by: error8

Originally posted by: geokilla
Well 2.4Ghz is as high as I can go on 1.32V. I'm using a Cooler Master Hyper TX 3 so my temps are fine. Maxes out at around 53C (95% F@H) at the moment.

So I guess the next time I overclock, I should just start off strong assuming I have a good cooler and case?

I had this chip and the same mobo as you and kept it at 2,58ghz with 1,48 V, for some 6 months or so until I got my 4000+ X2.
The bottom line is that at least mine, was a crappy overclocker, the mobo too ( stupid 1,95V max on memory).
Keep the temps under 60 C and don't be that scared about voltages, since this is a 90 nm chip and likes high Vcore. If you're maxing out at 53 C at 1,32 V you really need to change your cooler.

My case has really bad airflow. I'm using the stock AMD fan as the "front intake" even though there's no air slits or anything like that which would allow the air to go in from the front. I also have a side intake and a rear exhaust. I think the cooler is on correctly, but I'm not too sure. It took a full hour with my friend to put that cooler.

Plus, I say it's running at 1.32V but I actually have no idea what the vcore is because when I set it to 1.32V in BIOS, I get like 1.36 during idle and 1.32V during load in PC Probe. CPU-Z reports a low 1.184V, which is obviously wrong....


My computer seems to lock up at anything over 240 X 10, but if I drop the multiplier down, I can get up to something like 270 X 9 even though I'm giving it 1.4V for both settings. How come it's like that?

 

[THUGSROOK]

Originally posted by: LOUISSSSS

Originally posted by: THUGSROOK
as long as you DONT add any voltage, it doesnt matter what you do.

large FTTs.

24 hours is a waste of time and electricity.
you can go for 24 hours and still not be 100% stable, so its a total waste.

24 hours will show that your system is more stable than if you test for 1 hour, which is about as good as nothing...

youre doing nothing but wasting your time.

Prime95 is not the final word stability, even after 24hours.

 

[aigomorla]

Originally posted by: THUGSROOK

Originally posted by: LOUISSSSS

Originally posted by: THUGSROOK
as long as you DONT add any voltage, it doesnt matter what you do.

large FTTs.

24 hours is a waste of time and electricity.
you can go for 24 hours and still not be 100% stable, so its a total waste.

24 hours will show that your system is more stable than if you test for 1 hour, which is about as good as nothing...

youre doing nothing but wasting your time.

Prime95 is not the final word stability, even after 24hours.

i cant believe an elite is saying this.

Then what is your prefered method of stress testing? When overclocking stress testing is vital. And the reason for prime is because if it can survive prime, it will most likely survive whatever you throw at unless you F@H.

F@H is a better stress program however that can take a week.

Now why is a long endurance imporant? Has to do with the 2 dies on the cores. Ask any quad owner, you dont give your cores enough VTT, it will fail after 2- 4 - 8 hours

So this is why i cant believe an elite is saying this. I own 6 quads, i know how they behave very well. I own 1 yorkie and 5 kents, and all of them definitely need a form of stress testing after you overclock them.

Actually count my AMD spider which is a 9850 that makes 7 quads.

Originally posted by: THUGSROOK
as long as you DONT add any voltage, it doesnt matter what you do.

large FTTs.

24 hours is a waste of time and electricity.
you can go for 24 hours and still not be 100% stable, so its a total waste.

OP, this is half correct. You ram up clock speed, and if it holds, you get increase heat. More commands per clock, so you need to expect more heat.

However you need to apply more voltage to get 2.8ghz. Its rare for a X2 to hit that on stock.

If you want my honest advice, you wont see much difference from 2.8ghz to 2.4 unless you do a lot of encoding and processor intensive tasks. Even then the difference is a matter of minutes. Less then 5 depending on the length of the movie.

 

[THUGSROOK]

Originally posted by: aigomorla
i cant believe an elite is saying this.

believe it.

many ppl have fallen off the P95 bandwagon due to it being unreliable and a waste of time. it does not test everything and it is not the final word in stability. its benchmark is pretty good tho.

 

[betasub]

Originally posted by: THUGSROOK
it does not test everything and it is not the final word in stability.

That's because there isn't a definitive test for stability. P95 provides a stress test for instability i.e. a convenient way of finding CPU/memory instability through extreme stress.

 

[THUGSROOK]

24 hours is convenient?

what do you guys do with new cars? run them at Indy for 24 hours too?

 

[error8]

Originally posted by: geokilla

Originally posted by: error8

Originally posted by: geokilla
Well 2.4Ghz is as high as I can go on 1.32V. I'm using a Cooler Master Hyper TX 3 so my temps are fine. Maxes out at around 53C (95% F@H) at the moment.

So I guess the next time I overclock, I should just start off strong assuming I have a good cooler and case?

I had this chip and the same mobo as you and kept it at 2,58ghz with 1,48 V, for some 6 months or so until I got my 4000+ X2.
The bottom line is that at least mine, was a crappy overclocker, the mobo too ( stupid 1,95V max on memory).
Keep the temps under 60 C and don't be that scared about voltages, since this is a 90 nm chip and likes high Vcore. If you're maxing out at 53 C at 1,32 V you really need to change your cooler.

My case has really bad airflow. I'm using the stock AMD fan as the "front intake" even though there's no air slits or anything like that which would allow the air to go in from the front. I also have a side intake and a rear exhaust. I think the cooler is on correctly, but I'm not too sure. It took a full hour with my friend to put that cooler.

Plus, I say it's running at 1.32V but I actually have no idea what the vcore is because when I set it to 1.32V in BIOS, I get like 1.36 during idle and 1.32V during load in PC Probe. CPU-Z reports a low 1.184V, which is obviously wrong....


My computer seems to lock up at anything over 240 X 10, but if I drop the multiplier down, I can get up to something like 270 X 9 even though I'm giving it 1.4V for both settings. How come it's like that?

Drop your HT multiplier from 5X to 3X.

Set ram to 667 or 533 ( if it works, because mine didn't).

When I overclocked my 3800+ X2, I didn't drop the cpu multiplier and kept it at 10X. The higher it is, the better. You should touch 250 X 10 with 1,4 V- 1,45 Vcore with this chip if you have good temperatures.
This mobo did for me 341 X 8 with an athlon64 3000+ , but single core. I guess it can hit 300 mhz fsb with a dual core, so your only limitation is ram, HT or cpu but not the FSB.

 

[VirtualLarry]

Originally posted by: THUGSROOK
24 hours is convenient?

what do you guys do with new cars? run them at Indy for 24 hours too?

Stress-testing for 24hr is a hell of a lot more convenient than having your computer crash every 5 days because it's unstable.

 

[THUGSROOK]

Originally posted by: VirtualLarry

Originally posted by: THUGSROOK
24 hours is convenient?

what do you guys do with new cars? run them at Indy for 24 hours too?

Stress-testing for 24hr is a hell of a lot more convenient than having your computer crash every 5 days because it's unstable.

is that your only 2 options?
if 24 hours isnt definitive why do it?

there are plenty of programs out there that will error out much more quickly then prime95, find them.

 

[error8]

Well 24 hours of Prime95 seems quite a lot. There is now way in hell I'll ever stress my cpu that hard in his entire life inside my computer. Maybe this is a viable solution if you'll have to keep it 100% full load all the time, but for a regular user that plays games every once in a while , watches a movie or just uses internet explorer, there is no real need for that test.

I find it enough for 6 hours of prime or orthos and then some real life situations like a couple of hours of different games and lots of programs opened at once.

 

[aigomorla]

Originally posted by: THUGSROOK
24 hours is convenient?

what do you guys do with new cars? run them at Indy for 24 hours too?

no but i do warm up my car every morning for about 5-10 min b4 i drive it out.

I believe engines are fault tested for months, and then cars are even given demos wearing under cover bra's on the street, b4 production is finally started.

And Cars have a really nice warrenty. I dont see why your mixing the two as there completely unrelated.

I dont see whats wrong with prime and you have yet to provide us with any proof that prime is wrong.
Your pushing the thermal point on the processor, your checking its thread process to see if the overclock is stable.

Im sorry, but ANY form of testing is better then NO TESTing.

If you think prime is that bad, give us a stress program you think is good?

Should i bust out linpack?


Also you havent answered my statement on the Quads and VTT problem. HOW else you think your gonna figure that out unless you load the processor for long terms. Sorry i do play games longer then 2-4 hours on the weekend if i have the time. If my system cant hold prime for 2 hours, theres a high chance it wont hold during supreme commander.

And to me, none of my machines gets an okey stamp watercooling until its passed at least a 8hour stress test.

Even when i buy parts, the first thing i do is throw it on my torture rack, set it at stock and stress it for 12 hours. You'd be suprised how many ram sticks i got DOA.


Ask Any advance builder, or advance moder or overclocker. They all have a folder for fast access called system tools, or system programs. In it will have these programs:

OCCT <-- sometimes
Prime95
RealTemp
CoreTemp
CPU-Z
GPU-Z <-- if you have an nvidia
Omega Tools <--- if you have ATI
Or ATI Tools

If you guys want to know what i use, i use Knopsis. Its a bootable windows CD which is designed just to trouble shoot without the need of a hard drive. This way i dont need to install XP to start stress test on anything.

Sorry i dont like orthos anymore. Compared to OCCT and Prime, its a joke. Ive had Orthos pass tons of times where OCCT and Prime would catch it in 15 min. :\

 

[Martimus]

Originally posted by: aigomorla
If you guys want to know what i use, i use Knopsis. Its a bootable windows CD which is designed just to trouble shoot without the need of a hard drive. This way i dont need to install XP to start stress test on anything.

Sorry i dont like orthos anymore. Compared to OCCT and Prime, its a joke. Ive had Orthos pass tons of times where OCCT and Prime would catch it in 15 min. :\

Do you mean knoppix? I did a google search on Knopsis, and I just got a list of peoples last name. Knoppix seems to be a bootable CD with many open source applications on it. I'm not sure if that is it though, but I would be interested in d/ling the image if it does what you say it does.

 

[error8]

Originally posted by: aigomorla

Originally posted by: THUGSROOK
24 hours is convenient?

what do you guys do with new cars? run them at Indy for 24 hours too?

no but i do warm up my car every morning for about 5-10 min b4 i drive it out.

I believe engines are fault tested for months, and then cars are even given demos wearing under cover bra's on the street, b4 production is finally started.

And Cars have a really nice warrenty. I dont see why your mixing the two as there completely unrelated.

I dont see whats wrong with prime and you have yet to provide us with any proof that prime is wrong.
Your pushing the thermal point on the processor, your checking its thread process to see if the overclock is stable.

Im sorry, but ANY form of testing is better then NO TESTing.

If you think prime is that bad, give us a stress program you think is good?

Should i bust out linpack?


Also you havent answered my statement on the Quads and VTT problem. HOW else you think your gonna figure that out unless you load the processor for long terms. Sorry i do play games longer then 2-4 hours on the weekend if i have the time. If my system cant hold prime for 2 hours, theres a high chance it wont hold during supreme commander.

And to me, none of my machines gets an okey stamp watercooling until its passed at least a 8hour stress test.

Even when i buy parts, the first thing i do is throw it on my torture rack, set it at stock and stress it for 12 hours. You'd be suprised how many ram sticks i got DOA.


Ask Any advance builder, or advance moder or overclocker. They all have a folder for fast access called system tools, or system programs. In it will have these programs:

OCCT <-- sometimes
Prime95
RealTemp
CoreTemp
CPU-Z
GPU-Z <-- if you have an nvidia
Omega Tools <--- if you have ATI
Or ATI Tools

If you guys want to know what i use, i use Knopsis. Its a bootable windows CD which is designed just to trouble shoot without the need of a hard drive. This way i dont need to install XP to start stress test on anything.

Sorry i dont like orthos anymore. Compared to OCCT and Prime, its a joke. Ive had Orthos pass tons of times where OCCT and Prime would catch it in 15 min. :\

Cra*. You're right. Tried OCCT and it crashed in 4 minutes, but it passed 3 hours of Orthos. Got to up that voltage again. :roll:

 

[aigomorla]

Originally posted by: Martimus

Originally posted by: aigomorla
If you guys want to know what i use, i use Knopsis. Its a bootable windows CD which is designed just to trouble shoot without the need of a hard drive. This way i dont need to install XP to start stress test on anything.

Sorry i dont like orthos anymore. Compared to OCCT and Prime, its a joke. Ive had Orthos pass tons of times where OCCT and Prime would catch it in 15 min. :\

Do you mean knoppix? I did a google search on Knopsis, and I just got a list of peoples last name. Knoppix seems to be a bootable CD with many open source applications on it. I'm not sure if that is it though, but I would be interested in d/ling the image if it does what you say it does.

knoppix.

Sorry, i dont know why the hell spelled it that way.

ugh.. fingers going faster then brain right now.