• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

***official amd 83x0 overclocking thread***

In the UEFI load screen there are three large boxes bottom right to select these quick modes. Is on my Sabertooth 990FX r2 board anyway as I would think on the Crosshair too. Check it out.
 
Hopefully this will be my last question before I go and give it a proper good go and hopefully come back with some good results :)

Anyway, should I be working with my highest FSB (in my case 285) that let me load into windows or should I back off a bit to give me easier to manage numbers on the mem speed and north bridge speeds and the like?
 
I just used my highest fsb and then used ram speed closest to the rated ram speed, then manually set NB at between 2300-2400 depending on what the fsb would let me do
 
jimbaw said:
So if I change my fsb will it change the NB? And do I need to compensate for it?
Click to expand...

Some motherboards have the nb/ht set on auto and will change according to fsb, some will adjust it to its nearest stock frequency depending on the settings. Quickest way to find out, would be to just test it and check if it changes in the bios.
 
Last edited:
Avenged7Fold said:
Some motherboards have the nb/ht set on auto and will change according to fsb, some will adjust it to its nearest stock frequency depending on the settings. Quickest way to find out, would be to just test it and check if it changes in the bios.
Click to expand...

I'm pretty sure I've got them set to their stock value. I'll double check later (I've got a baby asleep on me right now).
Will I get anything if I up nb or ht manually?
 
I remember some vague recommendation that going above 2400Mhz doesn't give anything significant. Would also like to know myself!
 
Righteo so I'm now at this point

ap0h4b





Not sure if I need rights to post piccies and stuff so have done both to be safe.

Basically got to 4.9ghz (Woohoo!!!!) with max temps at around 57/58C in Linx and Cinebench. My question is now this, I've had a play around with my memory to get it up to 1866mhz which is my next available choice using an FSB of 280 (next choice down is 1491mhz which is below my 1600mhz rated speed) and loosened the timings as above from 9-9-9-24 as suggested but how do I test for stability in this and how do I go about stabilising if its not stable? (A lot of stables I know!) I'm assuming doing the small 772mb runs in Linx is for the CPU stress so should I just increase the size of the runs?

Appreciate your help as always guys and am Pretty happy with the results so far :D
 
Right my fsb is up to 225. It bsod after 10min on aida so I'm guessing I need more vcore?
Also my ht link is set to 2600mhz in the bios but in cpuz it reads 2900 ish. Should I set it to something lower in the bios?
 
actually my bsod error was "paged fault in non paged area". is this linked to my ram? would it help if i loosened my timings?
 
BTW guys, in case you are planning on sitting in front of a memtest screen for a while:

Depending on which version of memtest you have, you can allocate different amount of ram for testing. Obviously you want to test it all but a single run takes aaaages, so a speedier way to do it is by opening up several instances of memtest and dividing the memory you have to allocate up, by the number of instances. Each instance will put a core under load. Divide memory by four and it will complete tests four times faster!

If you run it from inside an OS, remember that you wont be able to test the Memory currently in use.

Have had a lot of memory issues with a large set of RAM recently given to me for clocking. Such a pain to have it run memtest to check for errors when you have 32gb to get through, so i am bloody grateful that memtest can be run in multiple instances.

If i was problem solving i would only use one instance from a bootable drive (due to OS taking up a portion of the memory to run) but if it is for error finding during clocking, i really abuse multiple memtest windows.

This is mainly to get any hardcore errors off the bat. Gaming/encoding/rendering or whatever the PC is used for should be how you finally check stability.
 
Right new update: set the ram timings to 10-10-10-30 and everything is fine.
However I've read everywhere that nb is supposed to be higher than Ht. But my NB is at 2474mhz and my Ht link is at 2925mhz. They're both set to their stock values in the bios so what should I do about it?
 
jimbaw said:
Right new update: set the ram timings to 10-10-10-30 and everything is fine.
However I've read everywhere that nb is supposed to be higher than Ht. But my NB is at 2474mhz and my Ht link is at 2925mhz. They're both set to their stock values in the bios so what should I do about it?
Click to expand...

Sounds like they are set to auto and they are clocking up with the FSB. lower them to a different multiplier and bench for comparison with them currently. Keep it at whatever works. If you see no gains, lower it for stability sake.

What are they running at in CPU-Z?

I thought that it physically wont allow HT to be higher than nb frequency.
 
Avenged7Fold said:
Sounds like they are set to auto and they are clocking up with the FSB. lower them to a different multiplier and bench for comparison with them currently. Keep it at whatever works. If you see no gains, lower it for stability sake.

What are they running at in CPU-Z?

I thought that it physically wont allow HT to be higher than nb frequency.
Click to expand...

that's what they are in cpuz. i'm a bit confused because I thought they were set manually. I'll have another go and see
 
Cheers Guys, have been playing BF4 for a good couple of hours with no crashes so I'm happy with my system as is but just have to wait and see how it pans out day to day now :) Think I'm getting a score of around 776 in Cinebench so is that pretty good for a chip being run on air?
 
You must log in or register to reply here.
Back
Top Bottom