• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

project stabilize sabertooth fx990 with amd fx9590

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Good chats on overclocking.

So much to mention like spread spectrum. Bios states having vrm ss turned on brings stability while off yields higher oc. Mines off.

DOCP and this particular set up doesnt work for those with inadequate cooling.

This chip is ready for big ram speeds though. 2 sticks any amount 2400mhz isnt a problem. Nb/HT voltage on auto, should see readouts at 1.4v, I use 1.45-1.50v. Nb core or chipset voltage if you will, Dont need more than 1.25v in most cases. HT v @ 1.3v.

Funny Ive used these figures since phenom one release with much trial and error along the way. These are my starting voltage figures minus the cpu. That takes testing.
 
hey again,
sadly, i've encountered 2 freezes and 1 BSOD (SYSTEM_THREAD_EXCEPTION_NOT_HANDLED - 0x1000007e - caused by dxgmms1.sys) since i put my fx9590 back into my computer...
i've set the NB voltage offset so that it reaches 1.3V and disabled DOCP but set the memory frequency to match my RAM (DDR3 1866) now.
i'm concerned that the cpu temperature is too high though. where do i look in hwmonitor to see my cpu temp? the cpu temps under "temperatures" in the top half or the temperatures under AMD FX-9590?
my cooler is Corsair Hydro Series H90 btw. the pump is plugged into the cpu fan opt slot on my mainboard (it's a 3-pole plug) and the fan that is mounted on the radiator is in the cpu fan slot (4-pin plug).

do you think it's still the bios-settings (the voltage?) that causes those problems or maybe temperature problems?
what should i do?

here's some additional info from hwmonitor and cpu-z if you need that. if you need other information from me, tell me pls :)
i still have an air cooler here but i don't think the temperatures vary much.


Untitled.png cpu.png gpu.png mb.png ram.png spd.png
 
Last edited:
I would drop the ram down to 1600 Mhz to try and rule out it possibly causing the instability. With the Cpu Nb voltage at 1.3 and the timings as loose as they are I find it hard to believe the ram is the issue but you never know.
 
i don't think it's my RAM, too. my pc runs perfectly when i switch out the fx-9590 with my old phenom II x6 1090T. i'll try that.

edit: can you tell me which of the temperatures hwmonitor lists are the right ones to look at?
what should be the maximum temperature for my cpu to deal with?
 
Okay looked at post #22...

...and went back to the thread you said was the original one you linked in your #1 post and this is where things seem to be now.

Okay HWMonitor shows Current Cpu Voltage as 1.417V > and HWMonitor shows that Cool N Quiet is still enabled since Minimum Cpu voltage was as low as 0.875V > and that the Maximum Vcore was at some point in time 1.464V.

At the time that CPU-z showed 4715Mhz the Cpu voltage in CPU-z Cpu Tab was showing as just 1.428V. That would have to be one h*ll of an awesome processor to work with that little Cpu Voltage. HWMonitor does not show that there has been enough Cpu Voltage to the processor and neither does CPU-z show enough Cpu Voltage, at least not from the FX-9590s we have seen come thru the forum.

And if the Maximum Vcore was 1.464V when under the Load of P95 Blend mode...well that is most certainly 98% not enough voltage.

I went back up to the top of post #22 to RE-read what was written before I hit the submit button to post this and found this in conjunction with the question of "what should I do"?

>> my cooler is Corsair Hydro Series H90 << Unless, again, you just have one h*ll of processor that cooler is not going to be enough for that FX-9590 when under a good load.

IF I wanted to actually know what was going on, I would go into bios and change n0thing but the multiplier to x20 save and reboot and see if some of the crap acting goes away. If it does you don't have enough voltage to the cpu or not enough cooling or a combination of same.

In addition I looked back even to that older orginal post and have yet to see what sort of power supply is in use and what case the stuff mounted inside of. We need a Signature with the informatin shown in my signature. So we can tell what in h*ll is in use.

Plus I notice that you have 32Gigs of ram in the board and I doubt for me if I were having issues with BSODs and such crap that there would still be 4x8Gigs of ram in the mobo loading down the IMC. I would already have looked in my manual and found the two primary ram slots and only have 2 of the 8 gig sticks in the board while trying to sort out 'issues'. But that is me I guess since I have done this crap for nearly 20 years now.

RGone...
 
RGone said:
[...]
Plus I notice that you have 32Gigs of ram in the board and I doubt for me if I were having issues with BSODs and such crap that there would still be 4x8Gigs of ram in the mobo loading down the IMC. I would already have looked in my manual and found the two primary ram slots and only have 2 of the 8 gig sticks in the board while trying to sort out 'issues'. But that is me I guess since I have done this crap for nearly 20 years now.

RGone...
Click to expand...

i already tried using only 2 of my 4 sticks in both colored slot-pairs as i didn't have access to the manual back when i tried it (was without internet for 2 days and don't know where my paper manual is). but regardless of how many (i even tried only one stick) or which slot-pairs i used, the symptoms remained the same.

i also edited my signature. sorry i haven't done that earlier.

the last change i made was setting the memory frequency to 1600. i'll try your suggestion to set the multiplier to 20 next, if my computer crashes again.
i'll also try to find the setting for cool and quiet and turn it off. but i actually set the cpu fan to go 100% duty at 50°C CPU temperature (was set to 70°C before) in the advanced tab...
sadly, i currently don't have the money to buy a new cooling system... but out of curiousity, which one would you suggest?

(btw: h*ll means hell, right? i'm no native english speaker, so if that wasn't what you meant, the right word doesn't come into my mind ;))
 
(btw: h*ll means hell, right? i'm no native english speaker, so if that wasn't what you meant, the right word doesn't come into my mind )
Click to expand...
Yes

I also agree with RGone setting the Cpu to a 20 multiplier to see if the issues go away is a good idea.

As far as cooling, looks at reviews on 2x120mm heatsinks such as Corsair 105, Coolermaster Nepton 240 etc...

Also are you posting from a mobile device, I still do not see a signature.
 
Just a quick note,
If you're still using an auto XMP profile, don't.
XMP profiles are based on Intel processors. Most of the time they don't work on AMD.
I would manually set that ram to 9-9-9-24 2T @ 1.65v @ 1600.
 
Mr.Scott said:
Just a quick note,
If you're still using an auto XMP profile, don't.
XMP profiles are based on Intel processors. Most of the time they don't work on AMD.
I would manually set that ram to 9-9-9-24 2T @ 1.65v @ 1600.
Click to expand...

i'm not familiar with most of the hardware-jargon. can you tell me what names i have to look for in bios to configure those settings?
is 1.65V the NB voltage?


Mandrake4565 said:
As far as cooling, looks at reviews on 2x120mm heatsinks such as Corsair 105, Coolermaster Nepton 240 etc...

Also are you posting from a mobile device, I still do not see a signature.
Click to expand...

thanks, will check those coolers out but probably won't be able to buy them in the near future :/

i checked the settings again... for some reason, my signature was just auto-saved bot not really saved. it should work now.



since i set my memory frequency to 1600 i haven't had a single freeze or BSOD. will test further for 1 or 2 days and then try what rgone suggested: setting back the memory frequency to 1866 and lowering the multiplier to 20. i'll post here again, then.
 
wullxz said:
i checked the settings again... for some reason, my signature was just auto-saved bot not really saved. it should work now. Great Signature.



since i set my memory frequency to 1600 i haven't had a single freeze or BSOD. will test further for 1 or 2 days and then try what rgone suggested: setting back the memory frequency to 1866 and lowering the multiplier to 20. i'll post here again, then. NO real need to try what I suggested now since you finally reported back and said after changing ram speed to DDR3-1600 that you do not (so far) have anymore BSODs. NO way in h*ll I would trade 4.0Ghz down from 4.7Ghz just to run ram at DDR3-1866. Jump from DDR3-1600 you are using now to DDR3-1866 would not give any real noticeable performance increase. Not really. I would far more greatly desire 4.7Ghz and DDR3-1600 with timings set accurately to anything at 4.0Ghz.
Click to expand...

So it actually seems now that it may well be a ram issue and as said within the quote above there is nothing really wrong with DDR3-1600 at 4.7 and quite preferable to DDR3-1866 at 4.0Ghz cpu speed.

Probably need to insert something here so we don't forget it. AMD only specs the FX processors to run one stick of ram per channel at DDR3-1866 and at the time of them "setting forth" that spec there were just about only 4Gig sticks of ram. So 2x4Gigs of ram at DDR3-1866 is the AMD official spec that has not changed.

Now the mobo makers have worked with boards and bioses so that *perhaps* 32Gig of ram will work. However will 4x8Gigs or ram at DDR3-1866 work? Maybe and maybe not! Is it possible that a particular brand and model of 4x8Gigs of DDR3-1866 ram n0t work? Sure it is possible. And just as possible that another brand/model of 4x8Gigs of DDR3-1866 ram might work.

In this mix of what-ifs about ram is also the question of whether the IMC (intergrated memory controller) inside the cpu is good enough on a per cpu basis for using 4x8Gigs of DDR3-1866 ram of any brand or model? That question can only be determined by some very detailed testing and trying various settings. Would it be possible to run DDR3-1866 ram by Corsair in 4x8Gig amounts at 4.7Ghz cpu speed? Very much unknown. AND most likely n0t worth the effort to try unless one is pretty sharp on bios setup and test for stability. If you are as you say not up to speed with computer jargon, then it maybe that a working and non-BSOD-ing 4.7Ghz with 4x8Gigs of ram at DDR3-1600 is not only good but maybe just fine without stressing about getting the ram to DDR3-1866. Somethings maybe worth the fight and effort but honestly some things just are not.

And again in all honesty, trying to type in a forum to learn to test and tune 4x8Gigs of ram to operate at DDR3-1866, takes a lot of patience and many screen captures at each level of testing since we are n0t at the keyboard of your system. Logical testing. No jumping from one suggestion to another but testing fully at one step and posting of captures before moving to the next step if necessary. I say this this way because I still am not certain of what happened when READ to READ Delay was tried or even if it has been tried.

So over the long run a working 4x8Gigs of DDR3-1600 without BSODs as you say your are now...well that might just be lovely without a lot of extra work.

RGone...
 
wullxz said:
i'm not familiar with most of the hardware-jargon. can you tell me what names i have to look for in bios to configure those settings?
is 1.65V the NB voltage?
Click to expand...
No DRAM Voltage

RGone said:
So over the long run a working 4x8Gigs of DDR3-1600 without BSODs as you say your are now...well that might just be lovely without a lot of extra work.

RGone...
Click to expand...
Agreed!
 
sadly, nothing has changed. i still get BSODs :/
i tried the following since my last post:

1. running with memory frequency 1600 and nb voltage of total 1.3V
my computer ran pretty long without causing any problems with this setting. but in the end, i received a BSOD (IRQL_NOT_LESS_OR_EQUAL)

2. trying to set the cpu multiplier to 20 instead of default 23.5, memory frequency of 1600 and nb voltage of total 1.3V
also gave me a BSOD after a while (same as in 1.)

3. set cpu multiplier back to 23.5, memory frequency to 1600, nb voltage to 1.3V, DRAM voltage to 1.65V (from 1.5V default/auto) and the timings according to the screenshots:
2015-03-23 09.45.25.jpg 2015-03-23 09.45.04.jpg 2015-03-23 09.44.54.jpg

i've been running my computer for half an hour with those settings now. no BSODs so far, but i also didn't expect them to pop up so soon.
i haven't stressed my cpu much with these settings because i wanted to report back first. i am watching some twitch streams in google chrome right now and while doing that, i have some serious lags that are more like short freezes.
i already hat those minifreezes with settings from no. 2 already but they weren't as long as they are now. the longest mini-freeze was about 5 - 8 sec seen in the second screenshot below.
you can see that 1 core has some spike in activity and the others fall down while 1 has the spike.

on a side note: during no. 3 settings i also moved my pagefile from my system drive (SSD) to my games drive because i remembered i haven't done that yet. mini-freezes happened before and after that change, so that isn't the cause. just mentioning it here for completeness' sake.

taskmgr.png taskmgr2.png

i also found the cool and quiet settings in advanced->cpu configuration and set it to "always disable".
i'll also report back as soon as the current settings produce a perma freeze and/or BSOD so you're up2date.

do you need other information than that? new screenshots from cpu-z/hwmonitor? under load (prime95?)?


edit: just had another ~7 sec freeze. task manager showed this:
taskmgr3.png

what i don't understand: there is no peak anywhere else. i had some lag problems while playing weeks ago and solved it by disabling the windows defender (it would block files for langer than it was good to block that file - disk peaked and cpu dropped).

edit 2: i saw the "system interrupts" process peaking in the details panel. i unplugged every external device (usb hub and g930 headset) except my keyboard and mouse now. still get those mini-freezes.
is there a way of finding out what's causing the system interrupts to peak so high? i guess that's not normal behavior of that process?
 
Last edited:
After reading through all this, this is what I would try. I would set the ram voltage back down to 1.55, leave it at 1600 for now and up the CPU_NB voltage. According to AMD the NB should be good for up to 1.45v and I feel this is where the issue lies. I also feel that overvolting the ram can be just as bad as undervolting when it comes to stability. When running 2400 ram my board auto sets the CPU_NB voltage to 1.4v so that the IMC can handle the added work. So, even though you're only at 1600-1866 32Gb is a lot of addressing for that poor IMC to handle. They don't even like 2x8Gb very well. I would start testing with the CPU-NB voltage and raising it a bit every time I saw instatbility. I feel the freezing is caused by the ram/IMC not communicating well. It is possible that the ram just isn't very compatible with the board but usually it can be made to work.
 
wullxz said:
sadly, nothing has changed. i still get BSODs :/
i tried the following since my last post:

1. running with memory frequency 1600 and nb voltage of total 1.3V
my computer ran pretty long without causing any problems with this setting. but in the end, i received a BSOD (IRQL_NOT_LESS_OR_EQUAL)

2. trying to set the cpu multiplier to 20 instead of default 23.5, memory frequency of 1600 and nb voltage of total 1.3V
also gave me a BSOD after a while (same as in 1.)

3. set cpu multiplier back to 23.5, memory frequency to 1600, nb voltage to 1.3V, DRAM voltage to 1.65V (from 1.5V default/auto) and the timings according to the screenshots:


i've been running my computer for half an hour with those settings now. no BSODs so far, but i also didn't expect them to pop up so soon.
i haven't stressed my cpu much with these settings because i wanted to report back first. i am watching some twitch streams in google chrome right now and while doing that, i have some serious lags that are more like short freezes.
i already hat those minifreezes with settings from no. 2 already but they weren't as long as they are now. the longest mini-freeze was about 5 - 8 sec seen in the second screenshot below.
you can see that 1 core has some spike in activity and the others fall down while 1 has the spike.

on a side note: during no. 3 settings i also moved my pagefile from my system drive (SSD) to my games drive because i remembered i haven't done that yet. mini-freezes happened before and after that change, so that isn't the cause. just mentioning it here for completeness' sake.


i also found the cool and quiet settings in advanced->cpu configuration and set it to "always disable".
i'll also report back as soon as the current settings produce a perma freeze and/or BSOD so you're up2date.

do you need other information than that? new screenshots from cpu-z/hwmonitor? under load (prime95?)?


edit: just had another ~7 sec freeze. task manager showed this:


what i don't understand: there is no peak anywhere else. i had some lag problems while playing weeks ago and solved it by disabling the windows defender (it would block files for langer than it was good to block that file - disk peaked and cpu dropped).

edit 2: i saw the "system interrupts" process peaking in the details panel. i unplugged every external device (usb hub and g930 headset) except my keyboard and mouse now. still get those mini-freezes.
is there a way of finding out what's causing the system interrupts to peak so high? i guess that's not normal behavior of that process?
Click to expand...

I would set my Ram up loose as a goose with this board and Cpu. But at least follow SPD timing set of 9-10-9-27 and roll 1.65v there. Your BSODs are from the Ram. Curious what the board posts stock. I imagine 667mhz at who knows what timings.

Another thing about the FX-9590.... It's a chip that likes to hang or freeze. The reason is unknown through most of the research I've done. So been studying my own system really hard to figure out why this happens.....

It's because this thing is made and designed to run big clocks at high voltage. It doesn't like Cool and Quiet turned off at all to be honest. It also likes low volts, 4.9ghz to be had under 1.5v. Looking for 5ghz you need a constant and stable 1.5125v and no less if not more. It's one thing to turbo 5ghz and another to run all cores 5ghz all the time. Problem becomes heat at this point.

4.9ghz around 1.488v stable, ran an hour and a half of Prime95 any more speed IE: 5ghz + 1.5+v, heat from the cores climbs really fast.

A good example of how Cooler temps prevail is simply running my bench tap to drain with water temps around 50F get me a max clock of 5.7ghz and a usable benchable clock of 5.4ghz - 5.6ghz.

But the FX-9590 does like to throttle under idle loads. Turn off power savings and be ready to actually cool the chip vs running warm.

A max stable temp, this FX-9590 is going to like temps under 50c as well. If you can achieve a max of 48c cores and socket temp, this chip will remain stable.

Only getting 100mhz OC at 1.4650v? Then your basically running stock. Keep the original settings, push the Ram to it's XMP profile and enjoy. The 5ghz turbo is plenty.
 
hey all im curious if there was any new developments in this problem i found the original thread and followed it here
i have pretty much the same setup as some have described

sabertooth 990fx r2.0
amd 9590fx black - with water cooler (came with the processor)
16 gb (2x8) ddr3 kingston hyperx fury 1866 ram

i currently get lockups quite commonly. all video and audio completely stop and the pc becomes completely unresponsive cant access task manager or even get the lights on my keyboard to turn on or off.
counting myself there are 4 users complaining of this problem with the same setup on these 2 threads is there something inherently unstable with this specific combination? would i be better just to return the sabertooth and get a crosshair V?
 
It's most likely because of the ram.
ZIf you start your own thread and give us a good list of parts as well as Shots of CPUz main , memory and spd we might be able to get things ironed out. Have you tried running the ram at 1600?
 
i havent yet but it seemed that changing the ram diddent actually help for OP but i havent seen any new posts so im just wondering if he solved and then never came back or if he gave up and went a different route
 
We don't know either. The one thing I do know is we have seen some issues with the FX and Kingston ram in general. Also some FX do have weaker IMCs and 1866 requires a bit more work than usual to get it up nd going.
 
You must log in or register to reply here.

Similar threads

I
OC advice for AMD RYZEN Threadripper 2920X - 12 Core
Replies
4
Views
1K
ivan7
I
S
CMT32GX5M2X6200C36 *4 Sticks Z790-E GAMING WIFI i9-13900K XMP Unstable
Replies
4
Views
673
Woomack
W
A
AMD FX-6300 what is the correct voltage for 3.5 GHz ?
Replies
3
Views
1K
EarthDog
EarthDog
W
  • Solved
Overclocking Ryzen 5600X using safe voltages
Replies
1
Views
5K
Johan45
Johan45
Max0r
12700k quick n' dirty undervolting/downclocking
2
Replies
24
Views
9K
EarthDog
EarthDog
Back