1090T OC - can't reach 4ghz - very close to it

[Johan45]

Once you find your max clock you can set things up with offset voltages to allow for all the green stuff to be turned back on. Then your CPU will cycle up and down with load.

 

[bassnut]

1.5 VCORE for a daily oc + high temps is not something I would recommend. For what you gain in performance it isn't worth the effort and possibly damage to your chip. Not sure where you live but winter here is soon coming to an end and with that warmer temps, if you are right up at the max limits be prepared for the possibility of temp related issues. I don't think any one of the guys here push their chips full out all the time, I for example run mine a good 175mhz lower then my max stable oc 4.9 GHZ instead of 5.075GHz. Leave yourself some wiggle room.

 

[bogdans06]

Yes, you're right. I would be on the edge.

With the current OC settings (CPU 3.9ghz @ 1.45v, CPU-NB 2.8Ghz @ 1.25v), the max temp is 55C (reached for a couple of seconds) and it's winter, but my PC is very close to the room radiator which is hot. When the room radiator is stopped (cold), my max. temperature is about 50C after 7 hours of prime95.

 

[bogdans06]

The 1090T OCed at my settings is amazing, I run any game at the highest settings without feeling bottlenecked (with R9 280x vapor-x). The overall performance is great in benchmarks too.

I just confirmed that my settings are stable for 10h+ in P95 Blend mode.

Thank you all again.

BTW, I have attached I new picture with my system.

 

[Johan45]

Looks good and you're right the system is running well right where it is. The extra 100 MHz just isn't worth the hassle.

 

[bogdans06]

Looks good and you're right the system is running well right where it is. The extra 100 MHz just isn't worth the hassle.

Thanks, I agree with you.

I bought an Intel 530 SSD 120GB and I upgraded to Windows 8.1.

I also optimized the cable management a little bit.



Regarding the OC, there is 1 more thing I should try: check if the system is stable with less voltage on CPU, CPU-NB or RAM.

 

[bogdans06]

Given my CPU cooler's position (airflow going down -> up), my rear 120mm fan was useless. The air which was exhausted from it was always cold, while the cooler from the top exhausted only hot air. I decided to remove the rear fan and install it on the case (side) window, pushing air between my video card and the CPU cooler, a perfect opportunity for my cpu fans to pull fresh air.



This, the fact that I did a little fixing on the cable management and the fact that I reinstalled the thermal paste a few days ago (by only spreading a thin layer of AS5 on the CPU, instead of putting a lot) - resulted in a max. temperature of 45C in Prime95 max cpu temp settings. The test was done with the current OC (CPU 3.9, CPU-NB 2.8 etc.). Impressive gain I'd say. Idle temp: 23C.

Guess what I'm going to do now.

LE: 4ghz @ 1.475v - max temp 50C, but it is stuck more at 48C. The test is still running (40 min so far) and I hope it's stable during all night. Maybe I try then with 1.45v.

 

[bogdans06]

4ghz needs probably 1.5v, because it's not stable with 1.475v (with very good temps). I decided to leave it at 3.9Ghz for now and find the lowest stable voltages.

The lowest voltage for the CPU is 1.425v (down from 1.45v) - tested for 8h P95 blend mode. 1.4v crashes the 3rd core.

Regarding the RAM, I'm surprised to see that the system seems stable with 1.5v. My ram is 2x Kingston HyperX blu 4GB, DDR3, 1600MHz, CL9, 1.65V. I run it with exactly these settings (1600mhz, 9-9-9-24) and it is stable at 1.5v for 2 hours in P95 blend mode and 1 pass memtest86+. How come? I can't believe it.

Will I lose performance if I try to get the lowest possible voltages? My current "philosophy" is to get the highest frequency with the lowest voltage.

BTW: idle CPU temp: 21C in memtest. Which is again, amazing. My CPU fans just needed more air to push through that heatsink.

 

[RGone]

Nothing wrong with lower voltages. Not if you are sure of stability. That determination is up to you.

A number of schools of thought on voltage anyway. I personally have zero issue for myself in getting to what " I " consider stable and then maybe bumping up the voltage a single notch since my board has very fine adjustments for most voltages. Plus I take into account how far below what I feel are safe temps I really am. I do video editting for which I have zero desire to lose any information. So to me I ensure myself in a manner that is not perhaps how others do their own setups.

A number of years ago I had to USE my computers for work. A couple of older IT persons said that there were enough possible issues with computers that had to work without overclocking them. I took that to heart and never overclocked the rigs that had to do with my daily work. Work that supported me and family. That main rig ran day in and day out for the 6.5 years I needed it everyday. I was glad of that.

So someone stamping my procedures as okay are not a concern. What is a concern is knowing about one's own rig and use and acting accordingly. Then it becomes how to fish and not a fish dinner.

Luck man.
RGone...

 

[madmeatballs]

I guess 3.9 is the sweet spot for the 1090t. I had that before too. Managed to get to 4.0 at 1.48v stable. I had a lousy Corsair a40 cooler that time.

http://www.overclockers.com/forums/...-time-oc-1090t-on-Asrock-990fx-Extreme4/page6