AMD FX 9590 underclock settings causing computer to shut down before POST screen

Wow it sounds like I have had a much better experience with my FX-9590 than most. But for the record I knew exactly what it was when I bought it and had sufficient cooling from day one. For reference I was running a similar loop to what ShrimpBrime laid out, EKWB Supremacy EVO cpu block, Monsoon bay res with a Swiftech D5 pwm pump, 360mm EK rad with 3 120mm delta fans. I also added fans to cool the VRMs (which the OP should seriously consider if going to run this cpu long term imo). All in a CoolerMaster HAF 932 case and originally I had it on a Crosshair V Formula Z later/currently on a Sabertooth R2.0. I was also able to run it at 4.8Ghz turbo disabled undervolted to 1.45vcore Prime stable for 8hrs (and a few other stress tests). Now currently it resides in a Corsair 540 Air and has an Xigmatek Aegir with Dual fans that is insufficient btw. I have it at stock clocks 2 cores disabled and undervolted to 1.425vcore with LLC at normal so with vdroop under load it hits 1.395vcore. As far as stability I have only ran Prime for 30 min so questionable at best.

Now I didn't post this to "muddy the water" but hopefully paint a picture that isn't all fire and brimstone. I think?

@ShrimpBrime I originally recommended the H80 as absolute min because I knew the 9590 was at one point sold with an similar AMD branded Asetek AIO. But after your comment I remembered reading reviews that it was insufficient so props for shooting that down. And I understand why you consider the 9590 frustrating processor I read Formula-z and FX-9590. Hangs and Freezing. back when you initially posted it and ya it helped me out. But in my case RAM stability is what killed me.

@rescuetoaster the voltage compensation your talking about is LLC or Load Line Calibration (at least on the motherboards I have used) and all the ones that have it I've used it is adjustable.

@op I agree with what these guys are telling you but you might consider disabling cores until you can upgrade your cooling to help keep temps in check with higher clock speeds for the increased "single thread performance".

Remy5460 said:

Wow it sounds like I have had a much better experience with my FX-9590 than most. But for the record I knew exactly what it was when I bought it and had sufficient cooling from day one. For reference I was running a similar loop to what ShrimpBrime laid out, EKWB Supremacy EVO cpu block, Monsoon bay res with a Swiftech D5 pwm pump, 360mm EK rad with 3 120mm delta fans. I also added fans to cool the VRMs (which the OP should seriously consider if going to run this cpu long term imo). All in a CoolerMaster HAF 932 case and originally I had it on a Crosshair V Formula Z later/currently on a Sabertooth R2.0. I was also able to run it at 4.8Ghz turbo disabled undervolted to 1.45vcore Prime stable for 8hrs (and a few other stress tests). Now currently it resides in a Corsair 540 Air and has an Xigmatek Aegir with Dual fans that is insufficient btw. I have it at stock clocks 2 cores disabled and undervolted to 1.425vcore with LLC at normal so with vdroop under load it hits 1.395vcore. As far as stability I have only ran Prime for 30 min so questionable at best.



@ShrimpBrime I originally recommended the H80 as absolute min because I knew the 9590 was at one point sold with an similar AMD branded Asetek AIO. But after your comment I remembered reading reviews that it was insufficient so props for shooting that down. And I understand why you consider the 9590 frustrating processor I read Formula-z and FX-9590. Hangs and Freezing. back when you initially posted it and ya it helped me out. But in my case RAM stability is what killed me.

Click to expand...

Memory isn't an FX-9590 issue with stability. It's the entire line up. Kingston for example isn't a strong point with these processors. Motherboard makes a big difference here too. When I discovered the way the Patriot Viper Xtreme (6GB triple channel Intel kit) posts memory speeds and timings, I knew that it would be fine with a nice 2 stick setup of Patriot memory and could run up to 1200mhz stable, which I usually stayed around 1100mhz give or take. The Crosshair always posted a Cas 11 even with lower memory frequencies. It was designed for fast capable memory.

Here's a good example of what a monster loop and top of the line hardware can do. 5.2ghz stable while benching at 5.4ghz wasn't really a big deal most days.



And for the record, I did try air cooling the FX-9590. It would run and throttle it's head off. Totally unstable in comparison to a good water loop.

I said H80 was garbage because even AMD stopped selling the cpu boxed with the cooler. It was getting returned as people just experienced major heat and stability problems.

Couldn't even run it properly on a 120.2 radiator. It would heat soak the loop within the hour. Prime95 is heavy duty. I can't believe people suggest he tests 4.5ghz while he should be closer to 4ghz with an air cooler.....

bassnut said:

Some of this will only be able to be done in the bios. There is nothing to tricky about do it there. Maybe one of your friends would be willing to help for your first time. Not many of us here use AMD Overdrive beyond extreme OCing.

Click to expand...

Unfortunately, my friend is having a family emergency and, well, he's the only one. I was hoping to be told what to do by you guys.
Without him, I can't test. I've never overclocked in my life...I never really had to before, because either I dealt with it or the hardware was good enough for me. I have another friend, but he's never owned an AMD in his life. So I have no one to turn to right now, except you guys.

What worries me about this beeping, is if this thing would cause damage to my hardware (of the computer turning itself back off during POST repeatably). Because now it's doing two beeps, or just one beep before it turns itself off. They aren't loud beeps as I would hear from a pretty old computer (like my mother's) who uses XP. I can't even record them with my phone, as they're somewhat faint.

Im not sure if you know this but the beeps have a place in troubleshooting. http://www.pcmag.com/encyclopedia/term/38515/beep The beeps are not just a general trouble message, but specific. I have never had to diagnose using the beep codes however, so that link is pretty much all I csn tell you. I have no idea if there are different beep patterns for different bios/board manufacturers or anything so be sure to double check that those beep codes are applicable to your board/bios.

We have a guide for exactly this reason. I recommend you read it, there is no reason you can't overclock this yourself. http://www.overclockers.com/forums/...ormance-Scaling-Charts-max-OCs)LN2-Results-co

There is an LED code indicator on the motherboard. When you start it up, it will display various codes that you can cross reference with the manual to see what it is reporting. I sincerely doubt anything bad is happening.
https://www.gigabyte.com/Motherboard/GA-990FX-Gaming-rev-10#support-manual

Benie said:

Unfortunately, my friend is having a family emergency and, well, he's the only one. I was hoping to be told what to do by you guys.
Without him, I can't test. I've never overclocked in my life...I never really had to before, because either I dealt with it or the hardware was good enough for me. I have another friend, but he's never owned an AMD in his life. So I have no one to turn to right now, except you guys.

What worries me about this beeping, is if this thing would cause damage to my hardware (of the computer turning itself back off during POST repeatably). Because now it's doing two beeps, or just one beep before it turns itself off. They aren't loud beeps as I would hear from a pretty old computer (like my mother's) who uses XP. I can't even record them with my phone, as they're somewhat faint.

Click to expand...

When you get into windows......And run AOD

I would start here with clocks / P-State FID 0x12 - VID 0x15 - IDD 11 (17.00x - 1.287 V)

So 200 x 17 and lower voltage to 1.300v (1.287 may be a tad low)

You work your way up from here. You can try and make your way up to / P-State FID 0x18 - VID 0x0B - IDD 14 (20.00x - 1.412 V) Again 1.4125v may be a tad low, be generous with 1.4250v. At this speed you should be getting hotter temps.

As far as the beeps go..... Do you get any error messages? Might say something like overclock failed, or loaded default settings press F1 to continue. Anything like this?

ShrimpBrime said:

As far as the beeps go..... Do you get any error messages? Might say something like overclock failed, or loaded default settings press F1 to continue. Anything like this?

Click to expand...

No, no errors. Soon as I hit the power button (the button on the top of my computer), there's a random chance either I hear the beeps and then my computer turns itself off, it turns itself off a half a second after I hit the power button, or I turn it on and there's no beeps.
Most of the time, I have to hit the power button several times for it to boot normally.
But yeah, no errors are coming up.

Does it post better when you've let it sit for a while?

Can you pull out 2 sticks of ram?

Have you ever reset cmos yet? Would you like to learn how?

ShrimpBrime said:

Does it post better when you've let it sit for a while?

Click to expand...

I've had it off for 8 hours, and it does the same thing.

ShrimpBrime said:

Have you ever reset cmos yet? Would you like to learn how?

Click to expand...

I have stated in the thread that I have reset CMOS (my friend showed me what button to push). When I did, BIOS came up asking if I wanted to load the 'optimized defaults'. I just ignored it.

ShrimpBrime said:

Can you pull out 2 sticks of ram?

Click to expand...

Stand by. I'm talking to my friend who suggests I get new pics of my BIOS screens. They'll be appearing in an edit soon.

EDIT: Here's the pics. Five of them;

Status update: I tried to remove two sticks of my RAM and put them in a sandwich bag. Booted into BIOS to verify I was at 8GB, was. Booted into Windows. Shut down, restarted.... NO BEEPS! Plus my computer didn't turn itself back off!

So it seems Mister Scott was right. The extra RAM was making my computer think it's a sheep (and has to beep). All there's left to do is run a stress test on my CPU (I guess).

To my knowledge sandwich bags are very staticy and would be a poor choice for keeping your ram. Perhaps it won't make a huge difference but you might be able to find something better? I let mine jangle around loose in an old sour cream container. Not ideal still but not staticy.

You'll need the 240 rad pretty much, pretty much what it takes if you want do 4.9 and 5 and that's lapped too get the single thread up to something at least somewhat okay and better. preferably I don't know if that mobo can run 2400mhz ram do that, it's no fanboy but this ones kinda slow to be honest on the Killer and ate threw a few lower end psu's, but this 1000w Glacier still one uninterruptible thing so far still.

So your ram is causing issues ..... I have had couple sets cause issues as have other users. I don't know Gigabyte Motherboards at all but what could have been happening is ram training, or the 4 sticks were too much for your IMC at your current settings. A little bump in your Ram volts as well as a bump in your CPU/NB volts might cure this.

If you still plan on running this CPU at stock you might want to look into better cooling.

So I just got done running the test, for 30 minutes (took a picture at the 25 minute mark). The settings I used was 21x Multiplier, and 1.4V CPU VID.
When running the test, I noticed at when nearing the 15 minute mark, that the CPU's temp starts to spike 5-10*C. With these settings, it managed to stabilize below 61*C.
The peculiar thing about these settings, is the CPU started to slowly cool itself down after the 25 minute mark (all the way to 52*C) before I stopped the test.

Glad its working ...... some things to note ......







1) 3 boxes show your speed of the cpu 4.2GHZ and 4.7GHZ ...... looks like you have some throttling going on.

2) VCore or CPU Voltage is at 1.5v ....... it will cause lots of heat.

3) Socket Temp .... 72*C ...... your right at the max we recommend staying below

4) CPU Core Temp ..... 65*C ...... Again right at the MAX we recommend staying below

Some people are happy to run at max ...... we don't recommend it. That being said Prime does heat up your chip fast, in the real usage you will probably not stress your chip this hard in everyday use.

So in a nutshell your screen shot shows what I figured you are running HOT. You should address this either better cooling down the road or do some down clocking.

bassnut said:

Glad its working ...... some things to note ......


View attachment 191206




1) 3 boxes show your speed of the cpu 4.2GHZ and 4.7GHZ ...... looks like you have some throttling going on.

2) VCore or CPU Voltage is at 1.5v ....... it will cause lots of heat.

3) Socket Temp .... 72*C ...... your right at the max we recommend staying below

4) CPU Core Temp ..... 65*C ...... Again right at the MAX we recommend staying below

Some people are happy to run at max ...... we don't recommend it. That being said Prime does heat up your chip fast, in the real usage you will probably not stress your chip this hard in everyday use.

So in a nutshell your screen shot shows what I figured you are running HOT. You should address this either better cooling down the road or do some down clocking.

Click to expand...

The test was stopped and adjusted over a period of time, that one at 4.7 with 1.5 was stock and is the constant. We adjusted for every previous test and were able to get it stable at 21x with 1.4 as the voltage. Which would be the temps you see there under value, its not being throttled those are the set values to the core multiplier and CPUVID, thats because we were testing and adjusting is what you're seeing.

Edit: And yes, we do know that we would have to resort to underclocking to get this build to work as a temporary solution til he can get watercooling, we're past that bump in the road.

Well, guys... our good friend *beep* is back for more when I turned on my computer this afternoon. Even after removing the RAM, it's still doing it.
I'm now convinced something is wrong with this motherboard (or the CPU). Either that, or these issues are actually "normal"...honestly, I don't know what to do anymore. I didn't have the beeping happen with my old board (I know that at least).

At this time, with all the hell I've gone through this (and last week).... maybe it's just not worth it? But, I'll leave it to you guys to decide. Maybe you all have another plan of making the beeping stop once and for all.
All I know is, twice, it passed the 'turn it off completely and leave it off for a minute' test yesterday. Yet it isn't 'passing' this one today.
Could it be possible the CPU suddenly goes from room temp to 80*C in a split second when I turn it on? ... Oy.

Benie said:

Could it be possible the CPU suddenly goes from room temp to 80*C in a split second when I turn it on? ... Oy.

Click to expand...

Been trying to tell you.......

It's a high leaking chip. This means the transistors leak voltage and are required a much higher voltage than say the FX-8 series chips. That's why you see the P-states like this.


P-State FID 0x22 - VID 0x02 - IDD 18 (25.00x - 1.525 V)
P-State FID 0x1F - VID 0x02 - IDD 18 (23.50x - 1.525 V)
P-State FID 0x1D - VID 0x02 - IDD 18 (22.50x - 1.525 V)
P-State FID 0x18 - VID 0x0B - IDD 14 (20.00x - 1.412 V)
P-State FID 0x12 - VID 0x15 - IDD 11 (17.00x - 1.287 V)
P-State FID 0x8 - VID 0x25 - IDD 7 (12.00x - 1.087 V)
P-State FID 0x10C - VID 0x35 - IDD 4 (7.00x - 0.887 V)

@ 4.5ghz there is not one FX-8 series that requires 1.5250v to run it.

Alright, ShrimpBrime. Let's talk. The facepalm was a bit too much for what I've gone through. I don't do research, this is why I ask people for parts (who I hope they've done theirs). I'm not tech savvy. I don't follow tech trends (as I don't care to).
All I wanna do, is buy the parts (who the person I talk to is telling me they'll work fine with everything I got), I get them, I have someone else who I can trust to install them for me, and then I play games with this upgraded hardware. That's been the way I've been used to, for years now. I've NEVER had to deal with anything like this before.

Maybe perhaps it's time that I did my own research to verify, but shouldn't that be the job of Newegg (or even PCPartPicker) having everything visible for people who don't really care for the tech specs, just the fact it's an upgrade for their old hardware?
Look, a few years ago, I attempted to use PC Part Picker to see if there was an upgrade for my 4350, and it claimed this was the best CPU out there (which is not true).

I know I ****ed up by getting this CPU. I have no business owning it. But, I also can't really blame the person that recommended me the hardware (as everyone makes mistakes)...even if their $400 mistakes.

Guess you could call this a learning experience.

How would newegg possibly know your cooling setup?
What else do you have in your case for airflow?