Overclock.net › Forums › AMD › AMD - General › Need help OCing with AMD A8-3870K + ASRock A75M-ITX
New Posts  All Forums:Forum Nav:

Need help OCing with AMD A8-3870K + ASRock A75M-ITX - Page 3

post #21 of 25
Thread Starter 
Quote:
Originally Posted by Yeroon View Post

I have same board, same issues. But i got a 3850 so little more tricky.

Yes, the asrock extreme tuning seems to make "something" stable.

The NB defaults to different settings depending on your ram settings. For 1866 it defaults to 720/1.2v, which is too low/high for this.
900mhz and 1.15+v should get you a good starting point. Should work good with the K as you don't have to use base overclocking.
Run as little volts with that board as possible. I'm folding at 3.559ghz (123 baseclock) at 1.39-1.4v but it just blacksreens with prime95 at the same settings. This is also getting warm for my low profile cooler, with core temps at 48-50c, and socket in the 55ish range. Running lower frequencies can easily get less volts, and should make the screen dropping out go away. See how low volts you can get it stable @ stock, then work your way up the multi. You should be able to get to at least 34x wihout having the blackscreen issue. I have axtu open on startup/auto apply, but going up in baseclock you have to turn off auto apply if you don't want it to run some questionable overclock. its happened to me.
3dmark11 shows nb issues along with ram issues. I take it those are set properly though?

I wish i saw this sooner!! I tried what you guys said perhaps 8 months back about lowering the volts and my motherboard DIED after like 2 weeks of it!!! and I had to use the RMA on it, and it took a month or so to just get it back.. after that I pretty much said forget about the problems and didn't even use the integrated at all;except for folding purposes.

You guys are right, because when I don't use the videocard at all, or use it just for folding, my CPU can get all the way up to like 90+C and still run ok... but about to crash..

so thats 50 degrees C higher when its not having to deal with GPU integrated...

ummm

So doing manual setup and putting it on 900Mhz NB with 1.15 volts right? I am about to try it and see if its stable, another thing.. do I leave it running at 1866 ramspeeds? or bust it down to 1600 ram? THAT would explain why when i used to turn it down to 1600 ram speeds it would be more stable.. Currently Iam not using the GPU Integrated at all and Iam letting it go at 1866 and its been like this for several weeks and it is fine.

Iam excited now! I'll report back to see if all of this works.... As for the Processor... you got 3.5 stable from those volts?? I am going to give CPU overclocking a try again soon.. I just want the graphics part dead stable on DX11 running the Dualgraphics CrossfireX.

Glad other people are running into the same problems and people knowing how to fix it!
Edited by Vash the Stampede - 2/11/13 at 1:41pm
post #22 of 25
Thread Starter 
Quote:
Originally Posted by TinoArg View Post


I also tried with the VDDP (which I don't really know what is it), lowering and rising, and it didn't affect anything.






Yeah that thing is confusing as heck.. because it says APU PCI Express or something.. but I believe that one is for the PCI express port it self and it doesn't change the VDDC on the Dedicated OR the integrated at all confirmed by Multiple diagnostics like GPUZ and Kombustor, and others.

Theres also one in the Manual CPU overclock settings, but i believe that is for the north bridge.. because the name of it says NB on it as well... so I think thats the voltages to the NB only which is built into the CPU...

Because changing any of these settings doesn't change that GPUZ says or HWMonitor or anything... on EITHER the Manual CPU Overclocking part and the PCI Express part

so far... it seems like actual integrated VDDC volt change abilities is a question mark...

and again.. Iam glad you 2 are here.. and both having the same problems.. with the same motherboard and with DIFFERENT CPU's!!!

RESULTS AFTER TRYING WHAT YOU GUYS RECOMMENDED:
SAME THINGS WOULD HAPPEN... I think my theory still stands on this.. BUT YEAH I can get a much higher stable CPU Frequency now!!! So There are some good things.. and without your guys help.. i would have never known or found out.

I think it is dependent on heat... that is my theory on it.. keep it below 50C and it won't crash.. simple...

Seems like the integrated part of this card is just completely unstable at higher heat ranges.. at 50-55C is crashing point.. kinda sad.. because the CPU itself can go up to 90C+ and still not crash on stock clocks and voltages

the other part about this is that things SEEM stable.. and for DX9 games the integrated card won't crash no matter what the heat is.. and no matter if its folding... BUT on DX10 and DX11 games, if the heat is abover 50C you can be sure it will crash... it will crash in DX9 as well.. but DX10 and DX11 are almost sure things when it hits those kind of temps...





I am just you like you 2... I WISH THERE WAS ANOTHER FM1 Mini-iTX because iam soooo deep in this case and i love the case itself and mysetup that i don't wanna change it...

on a good note they do make FM2's in this now.. but ASRock makes it again.... and i think its too new for people to come up with the unforseen problems... like the ones we have now so I don't wanna get a new mobo and a new CPU just to MAYBE fix a problem like this... (a 200 dollar investment)

iam sure all other FM1 motherboards that are MicroATX's and bigger have no problems at all.... they are designed differently with different chipsets and voltages.. and bio's abilities... so for us unlucky ones with this motherboard.. i think we just have ourselves here to count on with what works and what does...

I am sorry guys... but the only real fix to this is to just not use the integrated videocard.. I have gotten a dedicated videocard and thats why I haven't returned to this thread much.... it solved everything... and I can let my CPU loose to whatever temp... (90C+) and because we can't disable the integrated GPU, I just let it run Fold@home 24/7 and it never crashes.... Fold@home must use like a DX9 or something because I been folding for months now on end on this integrated and it hasn't crashed..

Overall besides not being able to OC really well and not taking advantage of both integrated and dedicated in Dual Graphics/CrossfireX... I love my PC... i have noctura fans.. and a fanless PSU and a low spinning CPU fan... my PC is dead silent.. and in a mini-itx Lian Li case... I am very happy with what I have... even with this problem..

and again... Iam glad finally 2 other people with the same exact motherboard are running into the same problems... for a while i thought it was perhaps the APU that was the problem.. but iam pretty sure its the motherboard... The reason iam glad you 2 are having the same problem is that there is a lesser chance that it is a CPU/APU problem because what are the odds that all 3 of us have a bad CPU??!?! its really slim... and it is most likely just this ASRock Mini ITX FM1 motherboard.

iam pretty confident that 100% of these boards have the same problem... we only see it because we actually run DX10 and DX11 games to actually see it fault... not many people play DX10 and DX11 games to ever really see this problem.. and without many people complaining to ASRock of the problem.. they don't really regard it as something they need to fix... it sucks.. but I kinda don't trust ASRock and AMD as much after this.. I bet there is a reason they are the only ones that make Mini-ITX boards in the FM1 and FM2 socket... Asus and Gigabyte and everyone else knows that they can't get it to run stable enough to have a good product.. so Asus and Gigabyte and others pulled ALL of their Mini-ITX from FM1 and FM2 shortly after initial release.... so who knows if its a CPU/AMD's problem.. or the motherboard/ASRock's problem.... it could be both... they should have a Mini-ITX standard on both ends so that problems like this would not occur... i don't think their is any industry standard on Mini-ITX.. all companies make them differently from my understanding.

ON a good note, running NB at 900 but iam keeping it at 1.2volts instead of lowering it down to 1.15 like you guys... for me heat isn't an issue.... i don't even care about the heat anymore as long as it doesn't crash with it. THANKS to YOU guys, didn't realize the the higher speed ram allowed for the 900Mhz BUT the it doesn't set it on there automatically for you, you have to do it youself.

as for the CPU frequency..THANKYOU so much guys.. it does work just like you guys said it.. running 3.3Ghz.. iam scared to go higher because I don't wanna go in and crack open the case to move the jumpers to reset the bios lol. I am not using the ASRock utility, iam using the bios and then saving, and then prime95ing... it seems to be the best stable way... when i use the Utility sometime it changes the bus frequency around sorta like AMD Overdrive.. it doesn't always change the right things to get the right OC, so sometimes your running at stock or lower.. or sometmies you are actually running at default... so yeah.. iam sticking to doing it with the bios.

SO! the final clock for me is 3.3Ghz@Stock volts.. the reason I don't wanna go higher then stock volts is because it scales right now with the Cool n quiet feature.. so the lower the clock, the lower the volts go as well.. once you change the volts manually, it doesn't scale anymore on this Motherboard... and I didn't want it to run at max frequency and max volts all the time because i never turn it off... and i fold on my integrated GPU anyways.. and i have a dedicated GPU that i play videogames on.. so yeah.. my system has a large work load on it all the time... and the CPU is always overwhelmed.. I generally turn off fold@home when i play videogames tho because it takes to much CPU power to feed the GPU.
Edited by Vash the Stampede - 2/11/13 at 2:34pm
post #23 of 25
Thread Starter 
Quote:
Originally Posted by TinoArg View Post

Hi, could you finally solve your problem?

I have almost the same config (A75M-ITX + A8-3850), and the same problem. At first, with a little OC, video signal went black after some minutes of gaming or GPU stress programs, though with CPU stress programs was stable, but now it do the same even at stock speeds....

But the problem is not the APU, I had another motherboard (Biostar A55MH) and it works fine (I only tried stock because it hasn't any OC option in the BIOS), neither the temperatures, mine only reachs 43°C at full stress with a CM GeminiiS.

The problem is the motherboard VRMs. They get very hot and Asrock motherboards has a protection which turn off the VRM to prevent damage. But I don't know why this happens at stock setting when it supports 100w APUs.

I'm thinking of making a heatsink for them, but if the problem is VRMs can't provide enough power, it will be the same.

any idea which heatsinks to buy that will fit our VRM's? those things are tiny!!!

will these be too big?

http://www.newegg.com/Product/Product.aspx?Item=N82E16835708012

EDIT:

BOUGHT! hopefully they will come soon, i really hope it fix's my GPU problem, the the GPU part in the APU.

the reason why i didn't do this early was because if its the VRM overloading, wouldn't it crash with the discrete card and CPU alone as well? the system is 100% stable with only discrete and CPU alone.. and UNSTABLE with APU's GPU and CPU working together.. or any time the APU is playing a DX10-DX11 game.... well... it was only about 20bux for me to buy those heatsinks.. I just hope it fits my VRM's .. the initial ones i wanted to buy from wayyyy back are now out of stock and i have to buy the bigger ones.
Edited by Vash the Stampede - 2/14/13 at 5:51am
post #24 of 25
Thread Starter 
Ok I ended up buying this as well from Xoxide, because I think the C10's might be too big.. but who knows... if they aren't too big then i will use it on the VRM!!
http://www.xoxide.com/swiftech-mc21-mosfet-cooler.html
(I think i bought the last one.. because it says its out of stock now)

BUT you guys gave me a huge hint! you said that when you do 1866 ram it becomes unstable... it does it for me too when iam using GPU... so the GPU is not stable at 1866.. and it is SOMETIMES stable at 1600...

I did something I never thought would work.. but i dropped it down to 1333 ram speeds... and yeah.... its dead stable.. at the cost o GPU performance on the integrated... BUT i have these motherboard heatsinks coming in, which of those mosfets are the VRM of the memory??? Because I am pretty sure our crashing is memory related.. and not so much CPU VRM related. does memory share same VRM as the CPU? or is that the VRM Mosfets on the top of the motherboard next to the RAM? and should i cover just those 2 parts? or the the 2-3 other mosfet looking chips next to the RAM slots as well?

Hopefully those heatsinks are coming in soon, iam excited to perhaps fix or solve this problem for all of us with this crappy motherboard.

heres a pic if someone can point it out to me:
http://www.newegg.com/Product/Product.aspx?Item=N82E16813157273
post #25 of 25
Thread Starter 
FIXED IT! can't believe somethign like the the VRMs was a problem! had this problem for over a year now and something as simple as heatsinks fixed it.. iam going with ASUS for now on... something that comes with VRM's stock.. i can even OC higher as well!! this is such B.S. that VRM's was a problem this WHOLE TIME!!!





I bought the Enzotech MOS-C10's and they DID NOT FIT!!!!! they are wayyyy too big, and I ended up buying the SwiftTech MC21's and they fit PERFECT!!! i have 2 extra's after i put the heatsinks on. so they came with 4, and 2 of them covered them perfectly, 1 of the Swiftech MC21's fit over 4 VRM's, and theres 8 total as you can see. so yeah!!!! everytrhing works now!! i can't believe it.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD - General
Overclock.net › Forums › AMD › AMD - General › Need help OCing with AMD A8-3870K + ASRock A75M-ITX