New to overclocking: Have some questions about my 6300

[Mandrake4565]

then when i tried for 4.5ghz the core kept failing again. it seems no matter what i do my cpu will not budge past 4.4ghz. and raising the bus seemed to only raise my temps. i was able to hit 4.4ghz at a lower cpu multiplier but i ran about 10C hotter than usual. if I run at my normal 22x 4.4ghz with no bus modifications, and vcore at only 1.42volts i can run stable all day at around 48 degrees C. id really like to get up to 4.6ghz though

Welcome to the wonderful world of overclocking Ben. This is the point where you either give up or really start to learn the inns and outs of your system. What works and what doesn't, it is going to take experimenting and patience. I suggest taking detailed notes, to remember what settings you changed. Also you may have hit the point, on your Chip, where the V Core increase doesn't justify the performance increase. RGone tested a 6300 and needed I believe 1.5 Cpu V core to get to 4.6, not all chips are the same but you may really need to start pushing the V core to get to 4.6, which may not be worth it for you.

 

[RGone]

Mandrake4565 told it just about like it is and what I am going to write is just to sort of fill in any blanks.

Forum Overclocking : Water Cooling [Solved] COOLER MASTER Seidon 120M vs Corsair H60 vs Antec Kuhler 620

I don't have to have a full-blown review to see the general, overall direction of some component or system. These are just plain everyday users responding to a question about the mini-closed loop cooling systems and how they don't really seem to do much better at all compared to a well engineered high end air cooling solution.

CM Seidon 120M is just not much of a cooler. NOT if you want to push a 6 core or 8 core FX processor.


Is it a Contender?
In that fairly long thread I was testing an Asrock Fatal1ty 990FX Professional motherboard to see if it could begin to measure up to my CHV motherboard in pushing a Vishera based CPU. It did pretty doggone good. I would surely buy it over the Gigabyte UD3 based motherboards.

But the thing I found off-putting was the amount of Vcore the FX-6300 needed to run fast. I was using a borrowed FX-6300 which was one of three a friend of mine had. He had tested all three of the FX-6300s he had and all on good CHV motherboards and all three tested the same to him. To me that means everyone of his FX-6300s needs a lot of VCore to run fast. Well a 'lot' of Vcore translates to a lot of heat. That did not bother him much as he had a 3 fan radiator system and not a cheap mini-water loop system that has little reserve to push a late model FX type processor.

So you are where you are speed wise and likely you will not go further with your cooling and the way that UD3 motherboard works its' LLC circuit. By logical and methodical trial and error you may reach greater than 4.4Ghz stable in Prime95 Blend mode but it will take you trying many options in the bios. And the major change most likely will be in the amount of Vcore you push to and thru the processor and the result of that will be HEAT your mini-water loop will not be able to handle.

Sadly or realistically there are limits to just about everything. In the case of your configuration being able to really push an FX processor, you are likely at the end. Change the system configuration and you can likely go faster as I did with an FX-6300 but at what cost? The cost probably makes no sense just to increase the cpu speed by 200Mhz. It would likely take a more friendly overclocking motherboard and at least in round numbers about a $250 to $300 cooling system. I just don't see that being a good return on investment.

My super cooling is at least 7 years old and hand-configured by myself to really cool on water when I was part of a benchmarking team. It has the beans and brawn to cool an FX processor, by and large the mini-water coolers don't have that much brawn or power. That is about where you are.
RGone...

 

[ssjwizard]

The UD3 is a platform he will be able to push that CPU pretty close to its real limits, but as you mentioned hes limited by cooling capability at this point.

**I dont think everyone should do this**
I would make note of this though. I have run every single one of my FX CPUs up to 70C while under sustained Folding @ Home loads for months on end without a problem. Instability above 55C is generally a sign that your CPU voltage is a tiny little bit to low for the clockspeed. A little bump will both fix the stability issue and raise the temps another 1-2C.

I have one 8320 here that I am very surprized manged to get binned as an 8 core. 3 of 4 modules will drop AT LEAST 1 core beyond 4.4Ghz unless I let the voltage wide open. This CPU has been running F@H at 4.8Ghz on 1.65V with maybe 2-3 reboots total and ZERO crashes for about 6 weeks now. I have seen a max temp of 73C during this time, and an average load temp of about 60C.

 

[EarthDog]

Why are you surprised? 8 cores were binned at 4Ghz... thats it. Anything above that is gravy. But it did 'barely' make the cut.

 

[caddi daddi]

and don't think you might run a khuler920 or one that size either, I have one and it will only just cool my 965 at 4.2.

when my 6300 gets back from it's sun tanning, beer guzzling vacation it's lazy butt is going into my gamming rig at about 4.4.

 

[ssjwizard]

Why are you surprised? 8 cores were binned at 4Ghz... thats it. Anything above that is gravy. But it did 'barely' make the cut.

Thats true. I suppose surprised isnt the right statement. What Im surprised about is how close to the line that the chip is. It does 4.2Ghz on stock, 4.4Ghz on +0.05v core, but then it gets out of control.

4.5Ghz on 1.5V and module 4 will drop both cores, and 2 / 3 will drop one each within minutes. At 1.55V module 4 will drop one core. At 1.6V I can take it up to 4.65 without dropping cores or crashing. 4.8Ghz requires 1.65V to stay stable.

That 4th module is a nightmare, if I disable it I can OC a solid 200-300Mhz further on the same or less voltages.

 

[RGone]

"ssjwizard" is that wonky cpu the first you got or the second? With those of us that can push even the PD cpu, it is just plain beginning to be as much hassle as the BD's were to get to go fast and maybe even somewhat worse.
RGone...

 

[ssjwizard]

Its the second one. I still have #1 in a clam here and I suspect its going back in anyday now.

 

[RGone]

Humh...I see. That tells me more about the outlook overall.
RGone...

 

[caddi daddi]

I have had cpu's and memory that would clock easy and clock high and i have had stuff that it was what it was.
right now I have some 1600 memory, it's darn fine 1600 memory and that's all it will ever be. I'll keep it because i can depend on it.

 

[Ben62884]

Mandrake4565 told it just about like it is and what I am going to write is just to sort of fill in any blanks.

Forum Overclocking : Water Cooling [Solved] COOLER MASTER Seidon 120M vs Corsair H60 vs Antec Kuhler 620

I don't have to have a full-blown review to see the general, overall direction of some component or system. These are just plain everyday users responding to a question about the mini-closed loop cooling systems and how they don't really seem to do much better at all compared to a well engineered high end air cooling solution.

CM Seidon 120M is just not much of a cooler. NOT if you want to push a 6 core or 8 core FX processor.


Is it a Contender?
In that fairly long thread I was testing an Asrock Fatal1ty 990FX Professional motherboard to see if it could begin to measure up to my CHV motherboard in pushing a Vishera based CPU. It did pretty doggone good. I would surely buy it over the Gigabyte UD3 based motherboards.

But the thing I found off-putting was the amount of Vcore the FX-6300 needed to run fast. I was using a borrowed FX-6300 which was one of three a friend of mine had. He had tested all three of the FX-6300s he had and all on good CHV motherboards and all three tested the same to him. To me that means everyone of his FX-6300s needs a lot of VCore to run fast. Well a 'lot' of Vcore translates to a lot of heat. That did not bother him much as he had a 3 fan radiator system and not a cheap mini-water loop system that has little reserve to push a late model FX type processor.

So you are where you are speed wise and likely you will not go further with your cooling and the way that UD3 motherboard works its' LLC circuit. By logical and methodical trial and error you may reach greater than 4.4Ghz stable in Prime95 Blend mode but it will take you trying many options in the bios. And the major change most likely will be in the amount of Vcore you push to and thru the processor and the result of that will be HEAT your mini-water loop will not be able to handle.

Sadly or realistically there are limits to just about everything. In the case of your configuration being able to really push an FX processor, you are likely at the end. Change the system configuration and you can likely go faster as I did with an FX-6300 but at what cost? The cost probably makes no sense just to increase the cpu speed by 200Mhz. It would likely take a more friendly overclocking motherboard and at least in round numbers about a $250 to $300 cooling system. I just don't see that being a good return on investment.

My super cooling is at least 7 years old and hand-configured by myself to really cool on water when I was part of a benchmarking team. It has the beans and brawn to cool an FX processor, by and large the mini-water coolers don't have that much brawn or power. That is about where you are.
RGone...

Actually I quite like the Seidon 120 so far. I "upgraded" from the Cooler Master Hyper 212 EVO and under load I would get a max of around 55C. With the Seidon I now get a max of no more than 48C at load for long periods of time. This is why I'd really like to push further, because temps are no issue at all right now.

Only reason I'm overclocking this ystem in purely for gaming... I seem to be able to get a bit higher FPS the higher core speed I'm running. I can't say I'm not perfectly happy with this 6300 at 4.4ghz running so cool, but it would always be nicer to go a bit further. If i'm not able to push it any further I would like to thank everyone who took time out of their day to reply to my questions and anyone who made suggestions for me. I really appreciated it, and I wouldnt have been able to get this far without your help. Thanks alot guys!

 

[ssjwizard]

Your at the point where you have to decide which of two ways to go with your OC. Theres the guess/check method trying to finess the last little bits out of the chip OR you go after it brute force. I know my cooling can take more than my CPU can so often I start with the brute force type of overclock.

You already know your on the edge of what your CPU likes without voltage. I would suggest you stop and contemplate just how high your willing to turn it up. The next step would be to set the vCore at whatever the number you just came up with and see how far that will take you on multi alone, and then repeat with a mix of HTT and multi overclocking.

Thats going to show you what the max OC your going to get out of the chip without either upgrading your cooling or shelving your concerns about its safety.

 

[Ben62884]

What is the max safe voltage I can put to this chip? BIOS says 1.55 but ive read about people taking it further. Alsol, what is the MAX temp I can have? I think its 70C on the CPU and 60C on the "Package" temp, correct? Also, say my max safe voltage is 1.55, is that the max it can SPIKE to, or the max it can be set to? Because I know if I were to set it at 1.55 I would probably get spikes up to 1.6-1.7

 

[RGone]

What is the max safe voltage I can put to this chip? BIOS says 1.55 but ive read about people taking it further. Alsol, what is the MAX temp I can have? I think its 70C on the CPU and 60C on the "Package" temp, correct? Also, say my max safe voltage is 1.55, is that the max it can SPIKE to, or the max it can be set to? Because I know if I were to set it at 1.55 I would probably get spikes up to 1.6-1.7

You have the idea of temps pretty close as those are the temp numbers I would consider okay for ME.

Now you want the truth about cpu voltage or you want some off the wall answer? Okay then. There is no safe over-voltage as such. If you read the link I showed for the testing I did on the FX-6300, I used 1.71 Vcore to reach 5.5 CPUz validated only. I did it expressly since the owner said >> REACH 5.5Ghz and I did. One hale of a lot more cooling than you have. Now back to safe Vcore. Most get by with about 1.525V as a workable day in day out Vcore. We have not nearly enough time in to see if the FX processors are going to tolerate high Vcore for long periods of time. I know of one FX-8120 that has seen at least 1.6Vcore for long periods of time and it is still rocken along.

The problem is that it depends on the traces in the cpu. If there is one marginal trace then the over-volt begins the electron migration toward failure. This failure though is not so very prominnent today so far as we have seen it. Only one person that I know of has posted in this forum section wondering if his cpu was dying from over-voltage.

I think I may have made this statement already in this thread but I know I have said it already in the forum. I have NO intention of using high Vcore to reach some of the clocks I first attempted with my FX-8350. The particular cpu I have is just too good at reaching 4.8Ghz where my video editting times to completion are shortened by the 4.8Ghz. So I see no reason to 'risk' my cpu anymore.
RGone...

 

[Ben62884]

Can I ask what it means when the CPU multiplier starts dropping speed in CPU-Z during a prime95 test? Seems like no matter which way I try to reach 4.5ghz, i can reach a "stable" one where the cores wont drop out on p95, but the multiplier value in cpu-z will drop from 22.0x to sometimes around 10.0x for a second or two. Is this because the CPU doesnt want to go that fast? Or does it need more vcore or something?

 

[RGone]

Is every green setting turned off/Disabled? Cool N Quiet, C1E, C6, TurboCore and APM. Windows power manager set to Performance mode. Those can cause cpu throttling up and down.

Have you got a fan blowing on the VRM heat sink? You now have water cooling and there is no over-wash of air helping to cool the VRM circuits. The VRM Overcurrent protection circuit will throttle the cpu down if the VRMs get too hot from supplying power to the cpu.
RGone...

 

[ssjwizard]

That sounds like you have "APM Master Mode" set to something other than disabled.