ORIFs OC Blog or something

[orif]

I figure I'll just start writing down what I'm doing and maybe someone will come by from time to time and drop a clue on me.

To start with I'm a Leo (I think this maybe important). I started my tinkering with an XT and I'm sure those that had the pleasure of working with MSDOS know that it was out of necessity. Now I find myself cruising these forums because I simply enjoy overclocking and my system is old enough that it's useful.

While I am looking into a serious upgrade sometime in the next year I figured I'd spend sometime playing with my hardware. It's just fun to see how high I can go.

Here are the basics:
AMD Athalon 64 3500+ Manchester @2.4ghz
Gigabyte k8n Pro SLI
OCZ Platinum DDR400 1gx2
BFG 7600gt oc
Maxtor 300g 7200 SATA
Antec Neo 550

My basic overclock that I boot into everyday:
FSB(Or what ever) 220=2.42ghz HTTx4=880MHz

With this basic overclock I am "rock" solid. Unfortunately I have to drop my HTT to 4 in order to to achieve even an increase from 200 to 205 on my FSB would cause a continuous reboot cycle (Unfortunately I don't think this was always the case).

Currently I'm looking to go from 2.4ghz to 2.5ghz. All I need to do is increase my FSB by 8mhz. So far a cpu increase from 1.35v stock to 1.375v has only made prime95 show it's usefulness. tinkering with htt and ram haven't produced results so I'm going to start bumping my core voltage as it is the variable that makes sense right now and is delivering results.

I figure I'll try 230 on FSB at 1.40 and 1.425vcore and see what I get.

 

[orif]

1.40 froze my pc after about 20 second on prime95 blend. After a reboot I'm still typing.
1.425v.failed probably around 20 minutes on prime95 blend.

1.45v and have no problems at the moment other than my temps 48c/118f. I've ran prime95 blend for 9hrs.Things seem to be going well.

 

[Thanatos421]

That looks good so far, hope 1.45 works for you. That chip should OC decently. Your tempereatures look good for a sligthly high vcore. Run the Small FFT test to ensure that CPU temps won't get out of control.

Good luck!

 

[orif]

Thanks, will small FFT generate more heat? I guess I will try and find out for myself. After, I'll see how many hertz I can squeze out at 1.45v.

 

[orif]

Well I bought a new cooler to replace my stock AMD. I wound up with a Zalman CNPS9500 AM2. The price was reasonable and it fits almost everything on the market. Installing the Zalman was a bit uncomfortable as it was up to me to decide if the cooler was placed properly. After scooting the cooler around for a bit I was concerned about my thermal paste but just clamped it down anyway.

*Temps are all while running prime95 blened*

I have to say that I'm amazed at the improvement. With the fan set at it lowest (1350-1400rpm) I'm at 40c. Higher fan speed will certainly lower my numbers.... .... .... 38c with the fan cranked up all the way.

Best case? I had a -10c improvement by installing a new cooler
Worst case? Its a lot colder in my house today. (My trusty REI compass, thermometer key chain says it's somewhere between 70-75f.

 

[orif]

Pre-Zalman

set HTT= 4 (stock HTT=5)
set CPUv =1.45 (stock CPUv=1.35)

with these changes I would have swore that in order to raise the FSB=230 to FSB=231 you would have to raise voltage.

Post_Zalman install

With nothing changed but heat sink I've been humming along at FSB=235 with Prime95 1040k running for 30minutes with no problems.

I got the Zalman so I could raise my voltage while keeping cool. I didn't expect this. The only change I can see in my system is a temperature change.

Why has a new cooler made a FSb increase possible without raising voltage? I'm going to see how I'll comeback when I find the new edge.

 

[orif]

It seems that I am able to now go from FSB=230 to FSB=235 without raising my voltage. With my temps running lower I'll take a look at raising my voltage a bit.

 

[Thanatos421]

Well, one possability is that the new cooler helps cool the chipset more than the stock cooler? Maybe? I don't know really, just grasping at reasonable straws. If the chipset runs cooler, it would be able to run at higher HT speeds (perhaps) without touching any voltages. Who knows, but at least it's working to your advantage

 

[orif]

Well I thought it was interesting and new it's new to me.

Got another question. Should I worry about volts themselves or just the heat they produce? I ask because I'd like to push my way up to 2.7 and unless my ram craps out I think I can do it by raising from 1.45 to 1.55 or hopefully. This up from +.2v from my stock 1.35. Am I a nervous nelly or am I already in an uncomfortable zone?

If anyone could chime in and let me know how long you've been running over stock voltage that would be great.

 

[Thanatos421]

Well, I believe (correct me if I'm wrong someone!) that upping voltage a few tenths is fine. The problem comes when you over-volt so much that the chip starts suffering from severe electron migration. The heat is only one piece, although usually the biggest, of the CPU destruction puzzle. Once you reach that voltage peak, whatever it may be for your CPU bin, then the "lifespan shortening" accelerates much more drastically than a typical moderate overclock. If 1.35 is stock, I wouldn't go much over 1.55 orr 1.6 IMO, someone please call me on that if I'm wrong.

 

[orif]

Thanatos bringing the good stuff as usual. I was poking around last night around 2.7ghz but not having much luck without raising my voltage over 1.55.

I think I'll retry some of the other variables and see if any improvement shows.

 

[Thanatos421]

Good luck! I didn't have much success overclocking my Opty 185 unfortunately. I did get it to post at 3.0, but it would fail stress testing no matter the voltage. I think that was just my motherboard at the time tho. It was the value version of the Asus A8V S939. VIA chipset and all. I think that had a big part to play in that OC effort.