SB Overclocking Thread

[nOOky]

I have the Asus P8P67 PRO and a 2600K coming along with new memory and a 212 cooler, I can't wait!

 

[Arkainium]

My 2600k is doing 4.4 GHz at 1.3 V on an Asus P8P67 Pro. I had to set LLC to high (50%) to get it completely stable. It doesn't completely eliminate vdroop, but it can maintain a load voltage of 1.28 V. I usually hesitate to use LLC because of the theoretical risks, but I really wanted to get at least 4.4 GHz. I definitely have no intention of going over 1.3 V.

Never mind. Setting CPU affinity to one core and running prime95 with one thread brings the system down. It appears that having one loaded core doesn't bump up the voltage as much as it does with four loaded cores. I guess leaving voltage on auto is not an option after all.

 

[jacktesterson]

I'm using an intel Matx p67 and 2500k and got 4.3 stable at 1.275v

Gonna stop here don't need anymore juice

Who says you can't o/c intel boards.

 

[jacktesterson]

I'm using an intel Matx p67 and 2500k and got 4.3 stable at 1.275v

Gonna stop here don't need anymore juice

Who says you can't o/c intel boards.

4.5 GHz @ 1.30v

Done now

 

[Dadofamunky]

4.5 GHz @ 1.30v

Done now

Way to go! Any IBT or Linx verification?

 

[jacktesterson]

Reverted back to 4.3 as 4.5 caused an error at 1.30v after 13 hours of prime 95.

To be honest, I am not a big believer of stress testing past 4 hours. How often is a CPU 100% stressed for more than this unless your folding?

Regardless, I am super happy with 4.3 and with what I do with this PC.

 

[StinkyPinky]

Wow, 13 hours on Prime. That's dedication

 

[HumblePie]

Wow, 13 hours on Prime. That's dedication

I thought the golden standard was 24 hours?

 

[beyoku]

"Fuggit"..............just got a Noctua NH-D14, I will let you guys know how it goes this weekend.

 

[StinkyPinky]

I thought the golden standard was 24 hours?

Possibly. I'm not that dedicated though.

 

[exar333]

I thought the golden standard was 24 hours?

Some people think a few hours is enough, others want a week (or month) of stable peak performance without errors. I lean more toward the latter; a couple days at 100% usage without errors tells me I am stable. That should be with less than ideal ambients too....testing in the winter with the windows open doesn't really accurately tell you how you will run when it's summer and the ambients are 30C+ (potentially).

 

[jacktesterson]

Bios is set to 1.25v @ 4.2GHz. This is where I decided to stay as I get max core temps around 60 here, and I don't notice the difference any faster.

Why does CPU-Z say 1.296v?

 

[jacktesterson]

 

[slag]

Bios is set to 1.25v @ 4.2GHz. This is where I decided to stay as I get max core temps around 60 here, and I don't notice the difference any faster.

Why does CPU-Z say 1.296v?

You have voltage fluctuations probably. I'm stuck at 4.5ghz, 1.3v. I could probably go higher to 4.6 or 4.7, but I know i would need to increase the voltage and 4.5 is a good even number. 10 hours of IBT is enough for me to consider it stable.

 

[mclaren777]

I'm still a little baffled by OCing with Sandy Bridge. This is my current understanding of it...

A stock 2600K will run at 1.6GHz when idle and 3.4GHz under load, but it can also jump up to 3.8GHz in Turbo Mode when some (currently unknown) amount of additional load is placed on it. So effectively it can scale from any value between 1.6GHz and 3.8GHz depending on load, and it falls out of Turbo Mode when it reaches a set thermal limit.

When you OC using just the Multiplier, you are only raising the maximum ceiling in Turbo Mode. Otherwise it still runs at 1.6GHz when idle and 3.4GHz under load. OCing with just the Multiplier does increase the maximum Turbo speed, but it also decreases the time until thermal limit is reached, at which point it'll fall back to 3.4GHz.

Is that correct?

 

[jacktesterson]

My motherboard has an option for running turbo mode at all times, meaning when any kind of load it goes to turbo. I've
Never seen my 2500k @ 3.3 under any
Kind of load

I'm still a little baffled by OCing with Sandy Bridge. This is my current understanding of it...

A stock 2600K will run at 1.6GHz when idle and 3.4GHz under load, but it can also jump up to 3.8GHz in Turbo Mode when some (currently unknown) amount of additional load is placed on it. So effectively it can scale from any value between 1.6GHz and 3.8GHz depending on load, and it falls out of Turbo Mode when it reaches a set thermal limit.

When you OC using just the Multiplier, you are only raising the maximum ceiling in Turbo Mode. Otherwise it still runs at 1.6GHz when idle and 3.4GHz under load. OCing with just the Multiplier does increase the maximum Turbo speed, but it also decreases the time until thermal limit is reached, at which point it'll fall back to 3.4GHz.

Is that correct?

 

[jacktesterson]

Sticking eith 4.2 and cslling it quits for a while, as funbut this is all i need 24/7

 

[Dadofamunky]

You have voltage fluctuations probably. I'm stuck at 4.5ghz, 1.3v. I could probably go higher to 4.6 or 4.7, but I know i would need to increase the voltage and 4.5 is a good even number. 10 hours of IBT is enough for me to consider it stable.

10 hours of IBT! Awesome!

 

[Arkainium]

It appears that having one loaded core doesn't bump up the voltage as much as it does with four loaded cores.

Does anyone have a solution to this that doesn't involve disabling EIST and C1E? Or do I need to apply more voltage despite being fully stable when all four cores are loaded?

Edit: I have a hunch that it has to do with LLC activating only when all four cores are loaded.

 

[Axon]

I'm still a little baffled by OCing with Sandy Bridge. This is my current understanding of it...

A stock 2600K will run at 1.6GHz when idle and 3.4GHz under load, but it can also jump up to 3.8GHz in Turbo Mode when some (currently unknown) amount of additional load is placed on it. So effectively it can scale from any value between 1.6GHz and 3.8GHz depending on load, and it falls out of Turbo Mode when it reaches a set thermal limit.

When you OC using just the Multiplier, you are only raising the maximum ceiling in Turbo Mode. Otherwise it still runs at 1.6GHz when idle and 3.4GHz under load. OCing with just the Multiplier does increase the maximum Turbo speed, but it also decreases the time until thermal limit is reached, at which point it'll fall back to 3.4GHz.

Is that correct?

More or less, but you can shut all that crap off and just run at your OC rate. I disable real time change in OS and Turbo.

 

[scrubman]

You have voltage fluctuations probably. I'm stuck at 4.5ghz, 1.3v. I could probably go higher to 4.6 or 4.7, but I know i would need to increase the voltage and 4.5 is a good even number. 10 hours of IBT is enough for me to consider it stable.

Curious what max load temps you are going with. I won't judge, promise I kept mine to 70c but I personally don't think going up to low 80's would be too bad since I will not put that kind of load on it during my normal use.

 

[smakme7757]

My 24/7 overclock:

2600K 4.5Ghz 1.265v in bios, 9 Hours Prime Stable (Small FFT's)

Idle temp: 28c
Load temp: 48/9c

 

[Dadofamunky]

I'd like to see what apps people are using to measure their temps when they report them. Those things tend to be notoriously inaccurate. I tend to lean towards using the BIOS-based temp measurements as being the closest to reality. Of course that's not feasible when you're running stress tests.

 

[jacktesterson]

Coming from amd for the first time in years
To intel, I'm still confused about the temp readings.

My cores max out around 68 Celsius according to core temp and hwmonitor.

To be honest I Like how amd is setup better for overclocking in there boards bios'

 

[smakme7757]

I'd like to see what apps people are using to measure their temps when they report them. Those things tend to be notoriously inaccurate. I tend to lean towards using the BIOS-based temp measurements as being the closest to reality. Of course that's not feasible when you're running stress tests.

I use RealTemp and Aida64.