Just Putzen Round

Thanks I.M.O.G, that's useful info and nice to see you drop bye Just Putzun.

OK, I looked closer at your screens - they were fuzzy on the forum app, I could enlarge them on hwbot.

A few notes:
-1.8v
- if you can, turn off those extra cores for cpuz and spi.
-1.81v
- More cpunb for sp32M. It should at least go to 3.4Ghz, or something isn't right in settings.
-1.82v
- good start

Last question I have - what clocks/volts are you booting at in bios? I am real fuzzy here, but I recall getting cold and breaking the system in a little, then setting higher settings in bios (5.5-6g, 1.6-1.65V)... Don't quote me, but I think by starting higher out of bios, I could get higher from software in windows. So for instance, booting at 4Ghz and going to 8Ghz never worked. But booting at 6.2 and going to 8Ghz would. Dunno for a fact why, but I remember it as an observation and something I stuck to - perhaps this was causing other low level settings to different initial values I wasn't aware of, and changes by OC software after boot don't affect those same settings. I don't know, except booting at a higher baseline helped enough that it was part of my routine of working the clocks up.

Speaking of that, my best clocks and benching was often hours in. It takes patience sometimes. Once I figured out what worked, I still often had to work my way back up to it in subsequent sessions - I couldn't just jump back to 7.5Ghz using the same settings as before once it was full pot cold. The same settings would work, but getting back up to those clocks often required working it up to it.

Edit: Hi Mandrake. I haven't been around much for a while, trying to get back. Good to see you.

Blaylock said:

I think this Sempron is a pretty nice chip. Everytime I expect it to BS it just keeps goin. Eventually I'll need to bump the volts, but for now it seems to be doing fine. Probably the most surprising is the CPU-NB volts. I think I'm close to 3000Mhz and it's still at the stock 1.175v.

Click to expand...

Be sure to hand some of the credit to your MB.
Not all boards will push a Sempy to their limit due to the rather low locked multi, I think the highest I ever got my 145 was around the 332 FSB mark with my CHV-Z - The 990FX chipsets aren't as friendly about revving up as other chipsets are.
Good for handling larger multi-cored chips running multiple GPUs but from my experience, not so much for winding these little chips to their limit, that being the only real drawback they seem to have overall.

Some will clockup fine but the ones I have here just won't do it.

Nice work man!

I.M.O.G. said:

OK, I looked closer at your screens - they were fuzzy on the forum app, I could enlarge them on hwbot.

A few notes:
-1.8v
- if you can, turn off those extra cores for cpuz and spi.
-1.81v
- More cpunb for sp32M. It should at least go to 3.4Ghz, or something isn't right in settings.
-1.82v
- good start

Last question I have - what clocks/volts are you booting at in bios? I am real fuzzy here, but I recall getting cold and breaking the system in a little, then setting higher settings in bios (5.5-6g, 1.6-1.65V)... Don't quote me, but I think by starting higher out of bios, I could get higher from software in windows. So for instance, booting at 4Ghz and going to 8Ghz never worked. But booting at 6.2 and going to 8Ghz would. Dunno for a fact why, but I remember it as an observation and something I stuck to - perhaps this was causing other low level settings to different initial values I wasn't aware of, and changes by OC software after boot don't affect those same settings. I don't know, except booting at a higher baseline helped enough that it was part of my routine of working the clocks up.

Speaking of that, my best clocks and benching was often hours in. It takes patience sometimes. Once I figured out what worked, I still often had to work my way back up to it in subsequent sessions - I couldn't just jump back to 7.5Ghz using the same settings as before once it was full pot cold. The same settings would work, but getting back up to those clocks often required working it up to it.

Edit: Hi Mandrake. I haven't been around much for a while, trying to get back. Good to see you.

Click to expand...

I appreciate all the info you've put out here I.M.O.G. I was booting in at 5.7-8 and could run harder benches up to ~6 and managed that one R11 t 6.2ish. My biggest concern was the apparent throttling I was getting. At this point I'm hoping the problem was my BIOS settings. Next time I try this chip I will make sure that T,Probe is set to extreme and I really hope that fixes it. With a full tank I can run for about 8 hrs most likely so testing the cores could be possible. This time I was trying to run certain benches for a comp at the bot and R15 was one of them. I spent hours trying to get it to run above 6.0. I tried high NB voltage high CPU voltage. Believe me I'm no stranger to high power. I have a cold water setup here and have run 1.7-8 more than one occasion. That's what was really throwing me off this time. Anytime the voltage went up the damn thing froze on me. I know the CPU takes it, I just didn't figure out what was holding me back. The Spi runs were a last ditch attempt to get something out of it Tuesday night , I was down to my last liter or two of LN2 and wanted to at least get something to post. That's why the cores are all on and I knew I wasn't going any higher without more voltage but more voltage just shut me down. Even the CPU_NB volts if I put them much over 1.4v the system would run for a bit then lock up again. I realize this was my first attempt at such speeds but I'm certain something was holding me back. I couldn't even run NB speeds that I had run on water, hence the 2800 you see normally this chip sails at 3100.
I for the most part use Cinebench r11 as a semi test for stability at certain settings and my intention was to work my way up a bit at a time. I have learned that any time a bench actually finishes to save it even if it will run faster, you never know when something might die on you. I've already had a couple of those myself so I'm no stranger to the "WTF I wasn't even beating on it yet" black screen , dead soldier.
Thanks for the quasi limits on the NB volts, I knew that CPU could likely go 1.9-2.0 but you never see any NB volts in the screens at HWBot so I wasn't sure where the limits were on that one. I wouldn't be surprised if this one would run close to 4 at that temp if I can just figure out where things went wrong. It was so frustrating sitting at 6.1G for hours trying to get the same bench to run . Lock up, hard restart and try again. Over and over again. All the time knowing it should go faster but why won't it take the volts?? I know in regular use I can set high voltage like 1.65 for 5.0 and it'll run fine even though it only needs 1.5v. I tried up to 1.85v at 6.1 and the same results as 1.7. I know it wasn't a speed wall, it was more like a "work" wall. Anything that put stress on all 8 cores would lock it up. This is leading me back to some kind of OCP. I was using a 1000w rosewill capstone PSU which is built by Superflower, So I figured that shouldn't be holding me back in any way, it shouldn't be kicking in it's OCP.

What are your digi VRM settings? All safeties/over volt should be off. I used extreme for all the rest. I think.

Click to expand...

I'm really hoping that this is where I went wrong. I'll make sure all settings here are completely maxed for my next attempt and hopefully this all goes away.
Thanks again Matt.

P.S. I've been perusing your scores and see that must have been before the erata fix the Stilt put out. My Spi is the exact same as yours but almost 1G slower. It's surprising what a difference it can make.

So got bored and started re doing my overclock on my 6350. as of now I am stable at 4.8 and only 1.4v. Before I could only get stable at 4.65 so not sure if I just know more this time and doing alot better or if the temp in the house is just that much lower than before but its doing awesome right now. I'll post up pics in a little bit when I stop at whatever I make it to.


edit: posted a little too soon earlier. Going on an hour and a half at 4.715 @ 1.4v in the bios but cpu-z shows 1.5.
Max temp has reached 59 so if im stable here I will probably leave it for now. My new toys to torture should be in soon and I am finally getting my a$$ in gear for my custom loop. Soon to be ordering all my parts and will probably redo my build thread.

just trying to find a GPU waterblock that will fit my card.

Sounds good Ray, it's likely that you are just learning more about Ocing. Benching is one of the best ways to really learn the ins and outs of your hardware, cooler temps also help though.

Yea its officially 2hrs stable at 4.7. not much of an improvement like I thought I was getting but ill get it higher soon hopefully.

re-add "return step attenuator" to exterior side of cable internet inlet..... welcome back http request load time! why was it I removed that 3 months ago?......

also starting to believe I am reaching temp limits causing no higher than 5.05 stable...... try again with ice in the reservoir soon!


3.5 to 4 gallons

noticed the buy/sell classifieds on the main forum page.
requires 100 posts and a domain!

Very nice!

rayraysfunhouse said:

Yea its officially 2hrs stable at 4.7. not much of an improvement like I thought I was getting but ill get it higher soon hopefully.

Click to expand...

Nice Ray, If the voltage jumped that much I'd look at my LLC settings.

unsafesteagle said:

noticed the buy/sell classifieds on the main forum page.
requires 100 posts and a domain!

Very nice!

Click to expand...

It helps keep the rif raf out

I picked up an 8320 yesterday since it hasn't been benched by anyone here yet and it was cheap. All I can say so far is wow. This is a pretty sweet chip I must say. I ran IBTAVX and then P95 for 45 minutes @ 5.0 the volts went between 1.41 and 1.428 mind you I am putting 1.275 to the CPU_NB http://valid.x86.fr/deurmg

Holy crap, that's some good voltage for 5.0, I'm assuming that's chilled water? Please don't tell me it's on the Thermaltake, if it is...Did I say wow!!!

I'm jealous of that chip lol I want it. I ma aged to make it stable at 4.8 by slowly raising the fsb. That's where I was wanting to get to. Pushing the limits with the temps though. It's right on the line for cpu

Mandrake4565 said:

Holy crap, that's some good voltage for 5.0, I'm assuming that's chilled water? Please don't tell me it's on the Thermaltake, if it is...Did I say wow!!!

Click to expand...

Yes that's chilled water which was warming as I tested, once the chip got into the 30s it crashed P95 about an hour in. No idea what it will take under "normal" cooling but definately more.

rayraysfunhouse said:

I'm jealous of that chip lol I want it. I ma aged to make it stable at 4.8 by slowly raising the fsb. That's where I was wanting to get to. Pushing the limits with the temps though. It's right on the line for cpu

Click to expand...

There's always a balance Ray. That's the worst part when you have no room on temps any longer.

lol yea I know. But I am actually ordering my custom loop parts so maybe I will have a little leway once I set that up

It'll definately help, no doubt. Unless you just have a quitter for a chip

Yea hopefully not. Been wanting to upgrade to the 8xxx series but need to quit procrastinating on a loop

Ya that can take a while to get all the pieces together.

Yea I've got the pump, cpu block, tubing and distilled water just waiting on a radiators for the most part. Really want a GPU block as well but my gpu doesn't need it

Ya GPU really isn't necessary

Lol I like how it looks with a gpu block thats really the only reason. It'll end up waiting till I get a better one