over clocking my Q9550

running fine - 24 hours of prime 95 stable no errors

leaving it right wear it is

You mentioned this...> I will be posting my settings and if its stable or not and please if you know anything about what im doing please help cuz its my first time also please check back from time to time i will post question if im stuck which im sure wont be to long

I'm sorry, I thought that because you started an entire thread on overclocking your CPU that you might be open to suggestions or advice, apparently that isn't the case? If so I'm sorry about that. What you said in the statement above would lead me to believe you wanted some help with this.

I hope you mean the Prime 95 Blend test because the small FFT test doesn't stress the memory controller or chip set and isn't a good test of overall stablity. Running anything more than 10 hours is a waste of time because the entire test repeats itself over and over again. Just though you might want to know the facts about stablity testing with Prime 95 and also how to achieve a higher overclock with your CPU...

chev65 said:

You mentioned this...> I will be posting my settings and if its stable or not and please if you know anything about what im doing please help cuz its my first time also please check back from time to time i will post question if im stuck which im sure wont be to long

I'm sorry, I thought that because you started an entire thread on overclocking your CPU that you might be open to suggestions or advice, apparently that isn't the case? If so I'm sorry about that. What you said in the statement above would lead me to believe you wanted some help with this.

I hope you mean the Prime 95 Blend test because the small FFT test doesn't stress the memory controller or chip set and isn't a good test of overall stablity. Running anything more than 10 hours is a waste of time because the entire test repeats itself over and over again. Just though you might want to know the facts about stablity testing with Prime 95 and also how to achieve a higher overclock with your CPU...

ahh yes yes my target OC was 3.5 im trying to OC my CPU not burn it up im stiiting at 3.4 close enough to target yes ever test i do is blend i did take suggestions and they worked other then that i dont know what you point is

plus your not OCing a 9550 your OCing a 9650 kinda apples and oranges there huh??
what may work for 9650 might not work for a 9550 could i over clock a 2.8 northwood the same as a 9550 .. no just like the 9650 it OC is different

another point trying to get peck performance which setting y 1066 ram at 900 would defeat that

The only difference between the Q9650 chev has and the Q9550 you have is the Q9650 has a default multiplier of 9 instead of 8.5, so they're practical identical apart from the 170MHz speed difference at stock.

There is also the QX9650 that has the unlocked multiplier.

Everlong said:

The only difference between the Q9650 chev has and the Q9550 you have is the Q9650 has a default multiplier of 9 instead of 8.5, so they're practical identical apart from the 170MHz speed difference at stock.

There is also the QX9650 that has the unlocked multiplier.

beside the only differences that you list the stepping is also differant Q9550 is C0 C1 and E0 were the Q9650 is only E0

plus ever CPU is different along with motherboards

i got an identical CPU and board and when i do what i did to this system i get BSOD

[QUOTE=FCVPI99;779872

another point trying to get peck performance which setting y 1066 ram at 900 would defeat that[/QUOTE]

The actual RAM speed means very little.
FSB speed & Timings will play a larger role.

Once your CPU and FSB is stable, start working a bit with the timings.

Assuming 1600FSb is the goal:

The bandwidth difference between a 400/1066 strap for example @ 5-5-5-15
and
333/800 strap (960ish RAM speed) 5-5-5-12 will be little to none. In fact, depending on your sub-timings and cross channel settings, you may be able to get a bit better bandwidth at these speeds than at 1066.

IMHO, PRIME is OK if you have time to kill and as a secondary test.
A more accurate test for Heat and stability is Intel Burn Test.
Many OCers hate this app because thier precious OC that was Prime stable bites it under IBT in less than 2 minutes.

just my 2cents for what its worth.

wow ran it fine for 100 tests cool im happy

FCVPI99 said:

chev65 said:

You mentioned this...> I will be posting my settings and if its stable or not and please if you know anything about what im doing please help cuz its my first time also please check back from time to time i will post question if im stuck which im sure wont be to long

I'm sorry, I thought that because you started an entire thread on overclocking your CPU that you might be open to suggestions or advice, apparently that isn't the case? If so I'm sorry about that. What you said in the statement above would lead me to believe you wanted some help with this.

I hope you mean the Prime 95 Blend test because the small FFT test doesn't stress the memory controller or chip set and isn't a good test of overall stablity. Running anything more than 10 hours is a waste of time because the entire test repeats itself over and over again. Just though you might want to know the facts about stablity testing with Prime 95 and also how to achieve a higher overclock with your CPU...

ahh yes yes my target OC was 3.5 im trying to OC my CPU not burn it up im stiiting at 3.4 close enough to target yes ever test i do is blend i did take suggestions and they worked other then that i dont know what you point is

plus your not OCing a 9550 your OCing a 9650 kinda apples and oranges there huh??
what may work for 9650 might not work for a 9550 could i over clock a 2.8 northwood the same as a 9550 .. no just like the 9650 it OC is different

another point trying to get peck performance which setting y 1066 ram at 900 would defeat that

I think maybe you have underestimated my abilities by just a tad here. I know exactly what the safe limits are for your Q9550 because I have helped hundreds of people overclock them in the past. And not that 3.4Ghz isn't a good overclock because it's perfectly fine.

The numbers I gave you are well proven and in fact are safe 24/7 limits for your particular CPU. Remember that your the one who started the thread looking for safe limits and answers. I was only trying to give them to you based on my considerable experiance in this area.

Not apples and oranges at all, in the fact the two CPU's can safely run numbers very close to each other.

As for the memory settings. The difference between 1066 and 900 is only 166Mhz which is nothing compaired to the gains you might see by going from 3.4GHz to 3.825Ghz with your CPU. I have in fact measured the difference in these gains and CPU overclocking ALWAYS comes out on top.

I will however have to disagree with Wishmaster as to which stablity test is the best. Prime Blend is in fact superior to the Intel burn test. I've seen so many so called stable Intel burn tests fail Prime Blend within minutes. We aren't talking about Prime small fft's which only stresses the CPU or large fft's also isn't a complete test. Prime Blend is the real deal.

In fact I have dug those Intel burn test guys out so many times over at the EVGA forum that I lost count. I have 37 blue ribbons over there for helping people overclock their CPU's. They come in there crying an complaining that their games are crashing after running the Intel burn test all the time. It's sad a sight indeed.

The fact of the matter is the the Intel burn test is famous for burning up CPU's and isn't worth the risk. This fact repeats itself at overclocking forums all over the internet. It also doesn't stress the chip set or memory controller nearly as well as Prime Blend. It also gets the CPU much hotter than it would ever come close to with normal use. Heat kills CPU's so why take the risk.

I believe the problem is that people do not run the required 10 hours of Prime blend which is a complete loop start to finish and is mandatory for total stablity. Running the test a couple hours isn't even close to stable but people seem to think it is.

Looking at the numbers of my overclocks I would have to say I have at least some evidence to back up my statements.

I run my Q9650 at 4.05GHz 24/7 using 1.36 Vcore cpu-z and can easily run at 4.275GHz for benching with complete stablity, all on air. You can see the numbers in my sig for my other CPU's.

Good points Chev.

But perhaps I should clarify just a bit.

I like IBT mostly because its much faster at spoting errors when looking for a target goal, and it seems to yield better results for me using 8GB RAM. (fills it all up)

I have seen Prime pass for 24 hrs and IBT fail in 5 passes, but also have seen IBT pass 25 passes and Prime fail within 4-5 hrs.

I believe using IBT to find a target goal, and passing 25 passes MAX stress once there is a good start.

At which point, Prime is a good idea for 10hrs or so.

Although theres never a 100% foolproof stability test, Ive found if you can pass both, theres usually not going to be a issue.


We may disagree somewhat on the subject, but what would Seven Forums be if we couldnt debate our experiences and thoughts :)
But you've made some excellent points. Certainly worth re-thinking.

from what i know im over clocking my CPU -- NO its a locked chip you cant over clock the CPU on a locked chip im over clocking my mother board and im siting pretty for my board -- its max stock FSB is 1600 which is right were i am the stock FSB with my Q9550 is 1333 but my board can take the X chips that have the 1600 FSB

i dont want to kill my computer

i have been saying the wring thing this whole time so for that im sorry also im OC'ing using Stock voltage of 1.275