SOLVED OC in BIOS doesn't take? Asus M5A97 Phenom II 975BE

[gearhead2578]

Hi all. I'm no expert on this (hence I came here) Trying to OC an Asus M5A97 with a Phenom II x4 975 BE (deneb) in BIOS. I turned the bus up modestly to 205 and set the multiplier at 20 going for 4100MHz. Problem is, in testing with OCCT or Prime95, CPU-z (and the Asus AI monitor) shows the clock only reaching 3200MHz - the bus shows 205 but the multiplier will only go to 14.5.... Heat is stable, I've installed the Antec H2o 620, so it's not throttling (47C max) what gives??

 

[madman7]

Did you save the BIOS settings on exit? Do you have Cool N Quiet and C1E disabled in bios? If not, do so. Also, go into Windows Control Panel Power Options and set the power scheme to High Performance. That should turn off all the green, throttle-down energy saving stuff that messes with overclocking in the higher ranges. Download and run CPU-Z and HWmonmitor.Also post screen shots of HWmonitor and CPU-Z(showing memory and SPD tabs)

 

[Frakk]

I know anything later than 0902 and it cancels the CPU NB overclock, but i have never seen one where it will not let you set the CPU multiplier above 14.5, Which is also to low for a Phenom II x4 975 BE, that should be @ 18x

14.5 sounds more like a 1055t, in which case you would not be able to turn up the multiplier, as that's a locked CPU.

Please install this

And give us a screen shot like you see below.

 

[gearhead2578]

That did the trick, madman - it was cool'n'quiet tryin to keep it pedestrian! Got Prime95 running now and testing the radiator......


oc4 by welder3078, on Flickr

~EDIT~ Hits 54C within 15 min but seems stable there - going to test more tomorrow when I can babysit it. Thanks for the help, guys!!

 

[madman7]

Glad you got it. With the vcore that high, just watch the temps. What program did you monitor the with?

 

[Frakk]

 

[gearhead2578]

I'm monitoring with coretemp as well as occt when running stress tests with it. Coretemp seems to be more accurate than the asus monitors - they monitor at the board rather than the actual cpu core and read about 4deg C cooler. I'm going to be stepping the bus speed back down to get more stability, the sidebar (desktop gadgets) is shutting down repeatedly while stress testing. Other than that it's keeping just cool enough at 54C

 

[scrambler]

There is not much difference between the 975 and 980 is there? Why do the volts need to be so high to achieve 4.1 with the 975? I wish I could use 1.48 volts. It would put me in the realm of 4.4/4.5Ghz.

 

[madman7]

My 980BE has to have 1.475v to get 4.1Ghz. I wish mine had a lower stock vcore than it did(1.4375v). What did you get out of your 980BE scrambler(vcore and ghz)?

 

[scrambler]

My 980BE has to have 1.475v to get 4.1Ghz. I wish mine had a lower stock vcore than it did(1.4375v). What did you get out of your 980BE scrambler(vcore and ghz)?

4.1Ghz at 1.41V. That's under load.

Wow, my stock voltages are 1.35V :/ I stand corrected it seems. It is me who is not ordinary I take it?

 

[gearhead2578]

None of us are really ordinary - we take perfectly good parts and tinker until they're on the brink of meltdown...... and like it!

I'm after on more bit of advice/opinion - is it better to have the bus clock @ 200 MHz and a multiplier around 21-22x (which I did run and test with OCCT with less heat) or to have the bus ~ 205-210MHz and a multiplier around 19.5? Either one will get the same 4.1-4.2 GHz but often it's not the result, but the method that's important. This of course coming out of concern for chipset temps and loading - like I said in a post above, I'm no expert on this...

 

[scrambler]

None of us are really ordinary - we take perfectly good parts and tinker until they're on the brink of meltdown...... and like it!

I'm after on more bit of advice/opinion - is it better to have the bus clock @ 200 MHz and a multiplier around 21-22x (which I did run and test with OCCT with less heat) or to have the bus ~ 205-210MHz and a multiplier around 19.5? Either one will get the same 4.1-4.2 GHz but often it's not the result, but the method that's important. This of course coming out of concern for chipset temps and loading - like I said in a post above, I'm no expert on this...

Overclocking by bus, overclocks the Ram as well. Something to consider.

 

[trents]

Overclocking by bus, overclocks the Ram as well. Something to consider.

And the HT Link and the NB.

 

[click4dylan]

sorry to bump this thread, but my 980 came with a stock vcore of 1.26v! 1.4375 stock, are you serious? did i get an insane chip or something? i run mine on a zalman 9900 cnps from 2008 (or 2009?), and i can achieve 4 ghz at 1.35V stable. right now i run it at 4.2 ghz at 1.43V just to be safe. i'm sure it could run at 1.4 and be fine.

my CPU-NB is running at 2400mhz at 1.15v.

my motherboard is an asus m4a89gtd pro. for months i thought i had a lemon because I tried every voltage possible, but my system was not stable at anything higher than 3.8 ghz. i looked deeper into my bios comparing the voltages at stock and setting it to 4 ghz, and then i found out it was increasing the NB voltage (not cpu-nb) by like 60% (i can't remember the value right now). what i ended up doing was forcing the NB voltage back to the stock value, and that's how i achieved these clocks. it's been running 24/7 for weeks without a single crash. haven't tried higher than 4.2 ghz

the temps never go above 53C during gaming. stress testing reveals up to 60C. at stock voltage and clocks it runs at 40C during gaming and 48C during stress tests. and i live in the oven that is florida

 

[keny]

sorry to bump this thread, but my 980 came with a stock vcore of 1.26v! 1.4375 stock, are you serious? did i get an insane chip or something? i run mine on a zalman 9900 cnps from 2008 (or 2009?), and i can achieve 4 ghz at 1.35V stable. right now i run it at 4.2 ghz at 1.43V just to be safe. i'm sure it could run at 1.4 and be fine.

my CPU-NB is running at 2400mhz at 1.15v.

my motherboard is an asus m4a89gtd pro. for months i thought i had a lemon because I tried every voltage possible, but my system was not stable at anything higher than 3.8 ghz. i looked deeper into my bios comparing the voltages at stock and setting it to 4 ghz, and then i found out it was increasing the NB voltage (not cpu-nb) by like 60% (i can't remember the value right now). what i ended up doing was forcing the NB voltage back to the stock value, and that's how i achieved these clocks. it's been running 24/7 for weeks without a single crash. haven't tried higher than 4.2 ghz

the temps never go above 53C during gaming. stress testing reveals up to 60C. at stock voltage and clocks it runs at 40C during gaming and 48C during stress tests. and i live in the oven that is florida

Stable, in the oc world means "prime 95" stable, is that prime 95 stable at those voltages, if it is then you got a good one but if you have not tried prime 95 blend test for at least two hours then i would check it out and see how you get on

 

[click4dylan]

i haven't tried prime 95 since i use the machine as a server as well as a gaming machine. downtime is not acceptable. i have run linX on it for several hours without a problem though. and i just got mine to boot into windows and upload this screenshot at 4.5 ghz at 1.43V

. i tried 4.6 ghz but it crashed. more voltage might do the trick though. 1.43 is awful low for even 4.5 from what i've seen so many others run theirs at just to get to that speed (if they even can)

this cpu is also under the stress of 4 memory modules (12 GB) of different sizes. 2x 4 GB, 2x 2GB. they're 1600mhz 9-9-9-24 ram running at 7-7-7-18-1T 1333mhz. (i can't ever get 1600 working stable on any amd cpu)

 

[keny]

1600 ram has the best chance of running at that speed on a C3 (which you have), prime is the only thing to stress it for true stability, but i dont like using it myself, if it can do what i want it to do without crashing then that is "keny" stable! and good enough for me, that's a nice chip if it will run your'e needs at a nice low vcore, the imc will be stressed running them sticks, have you tried running the cpu-nb at a few notches above stock to see if it helps with running them at 1600?

 

[click4dylan]

i run the cpu-nb at 2400mhz at 1.15V. i tested 2600 and it worked but it increased temps too much (already borderline 55C). i had a 955 C3 that wouldnt pass 3.6 ghz at any voltage, and it too couldn't handle all 4 sticks at 1600mhz.

only when i use 2 sticks is it stable at 1600mhz

i noticed a long time ago when I run the memory at 1600mhz 9-9-9-24 and try and game with it (it will blue screen after a random amount of time, usually within 1 hour), everything seems so much smoother than at 1333 7-7-7-18. the frame rate itself was basically the same but the general feel of mouse movement was much smoother.
i read online that setting the memory to ganged mode can also increase the smoothness of games. so i'm considering trying out ganged mode at 1333. i don't know if the mem running at the current speed is faster than the stock at 1600, but i'd love getting it stable at 1600. i just don't know what to do to get it so. cpu-nb doesnt help it at all, in fact, the higher i go the more unstable it gets. memory voltage doesn't seem to do anything either (they're all 1.65v chips). 1600 almost seemed to eliminate screen tearing with vsync off. it was that butter smooth.

 

[keny]

That voltage is nothing for the cpu-nb i would not be shy in running it up to 1.3v to achieve stability, but it will add heat to the chip the more voltage you slam into it, and you say your'e temps are high already so maybe thats not a option for you

 

[click4dylan]

i tried up to 1.35V on the cpu-nb on both the 955 and 980, still not stable at 1600