SPCR

So, I got CrystalCPUID yesterday to simulate Cool N Quiet on a mildly overclocked processor (+200mhz with a -0.025 undervolt, stable on orthos overnight). I read through the guide to it on here and looked around for other guides.

So far I have it changing the multiplier. It thinks that it is changing the voltage, but SpeedFan and CPU-Z tell me otherwise. They always showed me when Cool N Quiet changed the voltage.

I have Cool N Quiet set to auto in my BIOS and have tried it on both minimal power management and always on and it still doesn't change the voltage. What am I doing wrong? Does Cool N Quiet need to be turned off in the BIOS?

Ch0z3n wrote:So, I got CrystalCPUID yesterday to simulate Cool N Quiet on a mildly overclocked processor (+200mhz with a -0.025 undervolt, ...

What am I doing wrong? Does Cool N Quiet need to be turned off in the BIOS?

There that is the problem. You need to do the undervolt with CrystalCPUID instead of the motherboard BIOS.

Haha, thanks. I saw your PM right after I posted this. Thanks for all your help by the way. Maybe you should make a comprehensive guide for CrystalCPUID for people like me who make tinker with the bios settings before getting everything on the front end working.

On a side note, it is fairly disheartening to me that the 5400+ BE seems to overclock equal to or, rather, more poorly than the 5000+ BE. Everywhere I have seen, people have no problem with the 5000+ getting stable at 3.2ghz, which is a 600mhz, with either a very small or no voltage increase. Most people seem to only be able to get a 5400+ to about 3.2ghz which is only a 400mhz overclock. Mine on the other hand, 3.1ghz isn't even stable with a 0.025v increase and going to 0.05v only gets me to 3.2ghz and I don't want to run that hot. It is a personal preference because I am using a Ninja Mini and the Brisbane temperature sensors don't give useful outputs. So I have it at 3.0ghz right now and I could use a 0.025v decrease and still be stable. But, that also only means a 200mhz overclock.

Before you ask, yes I reduced the HT multiplier and increased the voltage on the ram slightly and reduced it's speed and loosed the timings when I was testing for CPU stability.

Ch0z3n wrote:On a side note, it is fairly disheartening to me that the 5400+ BE seems to overclock equal to or, rather, more poorly than the 5000+ BE. Everywhere I have seen, people have no problem with the 5000+ getting stable at 3.2ghz, which is a 600mhz, with either a very small or no voltage increase. Most people seem to only be able to get a 5400+ to about 3.2ghz which is only a 400mhz overclock.

All the recent Brisbanes I've had could manage ~3.2 GHz at some voltage. The X2 4400+ I kept was actually the one that needed the least voltage.

I don't really need CPU speed on this system, and having a fast HT speed gives the biggest real boost (w/ the 780G), so I have the HT 5x240 MHz with a 10x CPU mutliplier. I haven't gotten 5x300 MHz HT working yet.

I thought someone wrote a guide on here to using CrystalCPUID.

Why is the HT so important? is it an IGP?

That is exactly my point, it seems like all the 65w Brisbanes are the same chips just with different default multipliers.

Ch0z3n wrote:Why is the HT so important? is it an IGP?

Yes, only important for IGPs.

That is exactly my point, it seems like all the 65w Brisbanes are the same chips just with different default multipliers.

Yes, the same chip is sold for $20 and $1000 at times. The 45W Brisbanes are also the same as the 65W versions. How good (voltage required for stability) your chip is luck-of-the-draw, but the high clockspeed, low TDP chips should be the best. I just buy whatever is cheapest.

Ah, my processor came with my case, so free was definitely the cheapest option.

I kind of would like to have an IGP for hybrid SLI, but it isn't that big of a deal.