Sign in with
Sign up | Sign in
Your question

Overclocking Barrier w/ 3700+

Last response: in Overclocking
Share
March 28, 2008 5:05:05 AM

I'm attempting my first overclock ever, so please bare with my noobness. I'm attempting to overclock my 3700+ San Diego.

I first raised my FSB in increments of 5 until I reached 230. Windows booted okay. Once I increased it to 235, Windows would no longer boot normally. So I increased the memory voltage to 2.65 from a stock 2.60. Still no boot. Then 2.7, still no boot. From here, I then tried to raise the chipset voltage to 1.415 from 1.400. Then to 1.425. In each case, I couldn't get the PC to boot Windows and it actually seemed like it got worse as I made those moves. I've now reset the voltages and have the FSB at 220 with an 11x multiplier. Seriously...is that it?

Is there something I'm not doing properly here or is my true cap 230 (which I've yet to stress test) with stock voltages? My temperatures haven't even broken 40C yet.

PC Specs:
Lian-Li PC61
ASUS A8N-SLI Deluxe
Corsair HX520w
AMD Athlon 3700+ San Diego
eVGA 7800GT
2GB G.SKILL DDR400
WD Carviar 300gb SATA 3.0gb/s
March 28, 2008 5:15:35 PM

Okay, I stress tested at 230 for 10 hours without error. I can't break 230 though. I'm not sure why. I thought this morning that it may have to do with my HTT, so I set that back to 4x, but no change. If I even touch my DDR voltages, the PC doesn't seem to like it. The CPU voltages seem to have no effect.

I'm lost. Anyone have any advice. I just find it hard to believe that all can I OC this lady to is 2.53gHz and I haven't even tweaked the voltages yet (well I have, but apparently not properly).
March 28, 2008 5:23:31 PM

The ram? List the ram settings.
Related resources
March 30, 2008 5:36:50 AM

I figured it out. My memory index (divider) was set too high and the memory couldn't handle it. I set it back to 333 and now I'm stress testing at 2.7gHz. So far so good!
March 30, 2008 5:39:10 AM

That was my initial guess, so I asked for your ram settings. Apparently, you didn't say how fast you were running it at.
March 30, 2008 6:15:58 AM

Any idea what temps I should stay within here? It's idling around 42-45 and under full load 50-51. This is only at 2.7gHz. I'm going to try and reach 3.0, but we'll see.
March 30, 2008 7:26:57 AM

Okay, I'm stable at 2.8. I want to push it to 3.0. I have my HTT set at 3x right now. Will I want to keep it here or set it at 4x? The FSB will be at 270 at that point. I'm also thinking I'll need to dial back my memory index to 266 as well, right?
March 30, 2008 6:32:16 PM

Okay, I've reach 2.9 and I am in the process of stress testing. Assuming all goes well, are my settings okay?

http://valid.x86-secret.com/show_oc.php?id=337750

I'm concerned really with my HTT being set at 3x. Can I set it back to 4x which would put the HT Link back over 1000?

CPU-Z shows my Core Voltage at 1.440-1.456...it fluctuates between those two.
March 30, 2008 6:41:42 PM

stabgotham said:
Any idea what temps I should stay within here? It's idling around 42-45 and under full load 50-51. This is only at 2.7gHz. I'm going to try and reach 3.0, but we'll see.

Your load temps are fine, I would stay under 60C. I think 65C would be pushing it.
March 30, 2008 6:45:21 PM

stabgotham said:
Okay, I've reach 2.9 and I am in the process of stress testing. Assuming all goes well, are my settings okay?

http://valid.x86-secret.com/show_oc.php?id=337750

I'm concerned really with my HTT being set at 3x. Can I set it back to 4x which would put the HT Link back over 1000?

CPU-Z shows my Core Voltage at 1.440-1.456...it fluctuates between those two.

Your settings seem fine. As for your HTT, I doubt you can do x4, 1000MHz+ is quite difficult to do. Nonetheless, you can give it a try after you made sure your entire system is stable on x3.

For your Vcore, does it fluctuate under constant load? It is normal to fluctuate when going from idle to load.
March 30, 2008 8:37:52 PM

Yes, that's where I am showing it. At least under CPU-Z its showing the fluctuation while stress testing it. I'm not sure if that is something I should be concerned about. Is it? It does the same thing while idle in CPU-Z.

The HT link is running at 795.4. Is that too low?

Any recommendations on how to break 3.0? My temps are stable only reaching 55 under full load.
March 31, 2008 12:21:48 AM

Good read. I think I'm good for now. While stress testing, my PCU reached 57c. I don't think going for 3.0 would be wise right now without some better cooling and probably better RAM.
March 31, 2008 6:22:35 AM

Wow...this is just way too loud for my liking. I think I'm going to purchase a new HSF and some AS5. The Ximatek S1283 should be a decent one, right?
April 1, 2008 4:29:47 AM

Yes, the Xigmatek should be decent, but make sure it will fit. You have a socket A, right? So the Xigmatek won't work since it doesn't support it.
April 1, 2008 6:01:24 AM

No, I have socket 939 and the compatibility profile says that it will fit a socket 939. Once this thing hits 50c, the fan just goes nuts...way too loud.
April 10, 2008 6:11:58 AM

Okay, so I've received my new Ximatek S1283. On saturday I plan to lap my CPU and install the new HSF. I really, really, really want to break 3.0 with this processor. Right now I'm sitting stable at 2.916. There has to be something that I'm not thinking of that I can do to break 3.0. Even with the stock cooling, the highest my temps reach is 58. I'm very confident that the lapped CPU and the new HSF will bring those temps down drastically.

In theory, I think I should be able to break 3.0 with the stock cooling, but I'm not sure how.

My current CPU Voltage is at 1.435. What voltage is reasonable for an increase? I have not yet touched my DDR voltage which is at 2.5v which is the standard rating for my RAM (2.5v-2.7v). Should I increase that and see if I can push my FSB further?

Any tips would be helpful. I want to get every last ounce of juice out of this CPU since this year will be its last "HURRAH!" in my world as I intend to upgrade in the very near future.
April 11, 2008 12:37:23 AM

ARRGGG! This is driving me nuts. I rock solid at 265 FSB for 12 hours, but if I go to 266, she fails after about an hour and a half. This is killing me here. There has to be a way to go higher.
April 11, 2008 12:49:41 AM

stabgotham said:
ARRGGG! This is driving me nuts. I rock solid at 265 FSB for 12 hours, but if I go to 266, she fails after about an hour and a half. This is killing me here. There has to be a way to go higher.

Your ram is locked on 333Mhz right? In that case, there shouldn't be any need to increase the voltage. It's possible that you have hit a wall...

Anyways, you don't really need .1GHz more do you? I'd be content with what you have, it's already pretty good.
April 11, 2008 2:13:59 AM

what more do you want from the old 90nm chip?and your overclock is pretty good!so i think you should settle with what you got.and the amd cpu is lacking in performance compare to intel.so it would be better just to compare it to its family range.
April 11, 2008 6:58:09 AM

I know...I know. I really don't need that extra .1gHz, but it would be one hell of an achievement if you ask me. 2.9 is great, and I should be thrilled to have reached that on my first OC, I know...I started this whole process though with the goal of 3.0...I'm just so damn close!!!

Yes, I have my memory at 333, but it's not locked. I can move it around, but I can't run 400 at this FSB or else she crashes.

I just finished lapping my CPU and I installed the new Xigmatek S1283 cooler. My temps now are:

Idle - 32
Load - 45

That's a major decrease from my stock HSF which got:

Idle - 46
Load - 58

This cooler is absolutely massive!!! Can't believe it fit in my case, but it did...by mere millimeters. Awesome though, never runs over 1500 RPMs and is dead quiet.
April 16, 2008 7:48:14 AM

Okay...so I just couldn't leave it alone. So...I went back at it. Anyways, in my first attempt to OC further, I dropped my memory index to 266, set my multiplier to 10.5 and the FSB to 280. This OC'd me to 2.945gHz.

Okay, so it stress tested and came through with flying colors. So I'm ready to push further...but, before I do...quick question.

I originally had my memory index at 333 with the memory running at about 208mHz (approx.), now its at 266 and running at 220mHz (approx...since I'm not at home to check). I'm a noob, so does this seem okay? I mean, at some point, will reducing the memory index cause my memory or my system to run slower?

Also, what's better...higher FSB or higher multiplier? At an 11x multiplier, I couldn't break 265 on the FSB, but at 10.5 (with the memory index reduced), I am able to push beyond that 265 barrier on the FSB.
April 16, 2008 12:55:05 PM

if the idea is to get the highest cpu clock speed then it won't matter whether you adjust the "fsb" or the cpu multiplier. faster "fsb" could stress your motherboard and your components more than you like, especially if you don't have locks on the pci bus and pci express speed. but if everything is stable then things are good. it also looks like your memory is slightly overclocked, but again, if it's stable, it's all good.
April 20, 2008 3:24:20 AM

Well, I did it. I finally found a combination to break 3.0. Stable so far (after 4.5 hours) and looking good to make it 12 hours+. Here are some verification links. This OC ranks #75 on hwbot.org for OC's of the 3700+. I'm pretty proud of that.

http://www.hwbot.org/result.do?resultId=725589

http://valid.x86-secret.com/show_oc.php?id=348989

After checking out some more posts from people who have OC'd this processor and looking at some of the other scores, I'm thinking of pressing on even further. I don't really know how much further she will go, but I have to try. I have to find out what she can do! 3.2? 3.4? We shall see...but I'm thrilled about reaching my goal of 3.0!!
April 20, 2008 5:21:46 AM

Wow, that's seriously a nice OC for that old sucker. I'd be thrilled if it can hit 3.2GHz. Heck I'm thrilled it can even do 3.0GHz.
April 20, 2008 7:14:35 AM

I'm at an odd point with this one currently. I've got the following settings:

Memory Index: 266
FSB: 300
Multi: 10x
HTT: 3x
Timings: 2.5-3-3-6 1T
Core Voltage: 1.5v
DDR Voltage: 2.65v

My ram is running at 200 mHz which is right where is should be with no OC. I don't think it's going to allow me to climb any higher with this multiplier. I think I'm going to have to drop back to 9.5. I know this because when I push to an FSB of 305, it fails the stress test. Perhaps this processor can only get to 3.0. Who knows. I'm going to toy around some more and see what I can do. This has been a fun experience.
April 20, 2008 7:24:14 AM

I've never played with memory timings before...any clue as to whether or not doing something like that might help?
April 20, 2008 8:39:31 AM

After doing some more reading, I'm wondering if it's actually my motherboard that has capped out. Any clues on how to test this?
April 21, 2008 1:39:16 AM

stabgotham said:
After doing some more reading, I'm wondering if it's actually my motherboard that has capped out. Any clues on how to test this?

There is, and that's reducing the fsb to lax settings, then attempt to OC the ram. Knowing that all other components are stable, you'll know any instabilities are due to the ram, so work on it.

After you find the max OC for the ram, set the fsb as high as you can, knowing that the cpu is stable. Then obviously any instabilities is due to the motherboard.
April 22, 2008 12:45:43 AM

I talked it over with some guys at Overclockers.com and they all agreed, it seems my motherboard has capped out. In theory, I should be able to get more out of my CPU and memory, but my motherboard won't let my FSB go any higher. Oh well...I met my goal of 3.0, and its stable. To be completely honest, I don't notice any difference from the stock speeds, but maybe I'm just not doing anything strenuous enough to really strain this single core sweetheart!
!