I need help with testing hardware for OCing

Yalpe

Weaksauce
Joined
Oct 24, 2005
Messages
98
Hello guys,

I'm currently building 3 pcs for me and 2 of my friends. I know that they don't know how to/won't overclock their rigs but I plan on doing it for mine. Now the 3 setups have the same hardware :

AMD Athlon 64 3200+ Venice
OCZ DDR PC-3200 Premier Dual Channel CAS 2.5-3-3-7
Radeon x800 pro VIVO (will attempt the mod :D )
Samsung SpinPoint P 200GB - 7200RPM, 8MB Cache, SATA II
DFI LanParty UT nForce4 Ultra-D, Dual Gigabit LAN, SATA-RAID, Firewire, DDR400 (Socket 939)

The other 2 pcs will have the stock fans on the cpu and mine will have this :
Zalman CNPS7000B-Cu Quiet Copper CPU Cooler (Socket 462 / 478 / 754/ 939 / 940)

Now I have to prepare myself to overclock my hardware and maybe their's if I can get a small boost without disadvantages. I have 2 night to build all the 3 because my friends live 2 hours in car away from me and can't pick them up another day.

I'd like to test all the 3 cpus and ram for overclocking and keep the best parts for myself. According to a guide on this forum, I should test at which speed all the memory sticks can go without changing the timings, then test the CPUs and do some final tweaking. I don't think I'll have time to test everything with the final tweaks so can I only test the ram speed and cpu, keep the best of both and hope I can push them further later on.

Can anyone tell on how I could pull this off fast, without encountering too much problems and still have time for the software installs? I'm fairly new to OCing and I know this is the best place to get info!
 
Yalpe said:
2x OCZ DDR PC-3200 Premier Dual Channel CAS 2.5-3-3-7
First of all.. this RAM does NOT overclock very good at all.
And as you wrote it, you want to use 2 kits... 4 sticks of RAM in an AMD64 rig will not overclock well either..
Lastly, what size are they? :confused:
 
Well they are 2x512 per kit (1gig overall). Not 4 (2gigs). My mistake.

I had to work on very constraint budget so I had to cut somewhere. I don't care if it won't give me a godly overclock, just something good enough to make a difference.
 
Can anyone tell on how I could pull this off fast, without encountering too much problems and still have time for the software installs? I'm fairly new to OCing and I know this is the best place to get info!

There isnt really a fast way to test OC's. It takes time to test for stability. The minimum for a stability test is 24 hours of Prime95 on the cpu and 24 hours of memtest on the memory. Your looking at a week or more to find the best OCing pieces out of 3 rigs. Thats a lot of work. If they are all identical and all the parts were made about the same time, the OC's will probably be real close to each other. However, anything involving overclocking is a Mileage May Vary thing. So they could be different, but my guess is they will be real close to each other on thier max stable overclocks.
 
Yalpe said:
Well they are 2x512 per kit. Not 4. My mistake.

I had to work on very constraint budget so I had to cut somewhere. I don't care if it won't give me a godly overclock, just something good enough to make a difference.
Better hope to hell yours like memory dividers, I am almost 100% sure that I can't clock higher than 275HTT due to this crappy RAM.

I recommend getting any other brand, as long as they aren't the same chips.
 
So should I even care about testing the 3? I don't mind leaving the other 2 setups "as is" without overclock. I'd only like to get something out of mine, even if its not considered "very good".
 
Yalpe said:
So should I even care about testing the 3? I don't mind leaving the other 2 setups "as is" without overclock. I'd only like to get something out of mine, even if its not considered "very good".

Give us the first and second line of the OPN on the three chips and we can help you a bit.. but no guarantees..

OPN
The top two lines below AMD Opteron, just incase you didn't know what an OPN was.
 
Yalpe said:
So should I even care about testing the 3? I don't mind leaving the other 2 setups "as is" without overclock. I'd only like to get something out of mine, even if its not considered "very good".

Just depends on how much work you want to do. I would just build all 3 at stock and make sure they work good and then work on overclocking your rig later when you have more time. Overclocking can be pain, and sometimes is not worth the hassle for the gain.
 
Alright then, I'll recieve the parts tomorrow via UPS so I can't give you the 2 lines yet, but I will ASAP. I'm starting to think I'll just say to hell with it and build the 3 without OCing them.
 
Yalpe said:
Alright then, I'll recieve the parts tomorrow via UPS so I can't give you the 2 lines yet, but I will ASAP. I'm starting to think I'll just say to hell with it and build the 3 without OCing them.

If they're all the same (If you order them all from the same place most likely they could be...) then just build them all and don't worry about it...
 
Hi again,

I'm getting the hardware today, and yes they are from the same place. Another quick question :

What if I build the 3 pcs at once, and test them all for their max OC with quick tests. Wouldn't the cap be a good indicative of their OC capabilities? I could then pick the best parts and tweak them a lot more in the future.

Just to make it clear, I do not plan to make this OC an hardcore one, I only want to get something out of it. 2.0ghz -> 2.6... well... even 2.4 would make my day/week/month. Isn't there a way to quickly test for the best parts even if I don't test for stability first?
 
You could just keep raising the fsb and not your voltage and see which chip goes higher and keep that one. Quik way but not too accurate.
 
I usually work up HTT/FSB settings in 5 MHz steps, with Memtest86 as first test for memory stability.

If you want to move quickly to establish memory stability, try this:
- Let Memtest86 run tests 1 & 2 - takes only seconds.
- Press C 1 3 5 Return 0 keys in sequence to run Test #5 for one pass.
- Press C 1 3 8 Return 0 keys in sequence to run Test #8 for at least one pass.
- Eject Memtest floppy, boot into Windows and run SuperPi or Prime95 for CPU stability.

Memtest86 Test #8 seems to a good critical test for memory stability. Skipping the full range of Memtest86 tests can save hours of testing.

Hope this helps!
 
Just run up fsb like is said till you need more voltage and run the ram at 1:1. whichever goes higher keep that. Simple.
 
coolkishan said:
Just run up fsb like is said till you need more voltage and run the ram at 1:1. whichever goes higher keep that. Simple.
That RAM won't handle more than about 220mhz, and that's if he raises voltage and losens timings.

You NEED to run a divider with this memory.
 
OMG Borgschulze read what he wants to do!!!!:mad: :mad:
He just wants to quickly check which one out of 3 identical pc's is best at overclocking. He doesnt want to overclock anything stable at the moment!!!!!!!!!!!!!
Read and then post!!
 
coolkishan said:
OMG Borgschulze read what he wants to do!!!!:mad: :mad:
He just wants to quickly check which one out of 3 identical pc's is best at overclocking. He doesnt want to overclock anything stable at the moment!!!!!!!!!!!!!
Read and then post!!
He won't get far, you read... I have the RAM he has purchased, he will HAVE to use a divider in order to find even near the limit of those CPUs.

The computer will just crash before he knows anything about the overclock without the divider.
 
For the most part, Borgschulze was right. The memory stick cap at ~220-225mhz. But with a 1/02 I was able to test the cpus and found one that was clearly better then the other 2s. It might not be acurate but whatever, I'm glad I tried since I learned a bunch of stuff and at least I know there was no "pearl" in my parts so I won't have it on my mind torturing me :D .

Now I'm shooting for a stable OC. Wish me luck. And thanks for the help.
 
Yalpe said:
For the most part, Borgschulze was right. The memory stick cap at ~220-225mhz. But with a 1/02 I was able to test the cpus and found one that was clearly better then the other 2s. It might not be acurate but whatever, I'm glad I tried since I learned a bunch of stuff and at least I know there was no "pearl" in my parts so I won't have it on my mind torturing me :D .

Now I'm shooting for a stable OC. Wish me luck. And thanks for the help.

You never did tell us the first and second line of the OPN... :(

But yes, this RAM sucks hard.
 
Back
Top