Crashing while in game on a new system (some advice please guys)

[switcheHHH]

Hi all.

Hoping to get some suggestions on my issue, so let me get to it.
I have a brand new setup, so naturally I have a fresh OS install (Windows 7 in this case)

Basically I am experiencing crashes only while in game, the computer locks up entirely and I cannot do anything which forces me to do a manual restart.

I was using a 520W but I changed it out for my 750W, the issue is still there but does not happen as often as it did with the 520W.

My Rig is as follows:
Mother board: Asus M5A99X PRO R2.0 AMD 990X DDR3 ATX (Socket AM3+)
CPU: AMD Piledriver FX-8 Eight Core 9590 Black Edition 4.70GHz (Socket AM3+)
Graphics: EVGA GeForce GTX 960 SSC 2048MB GDDR5
RAM:DDR3 8gig, 1333mhz
PSU: Antec Trupower Blue 750W

Because of the way the computer freezes/locks up I am unable to provide any blue screen logs, no dumps are being done when this happens, so I am only able to give you what I am experiencing first hand.

I appreciate any suggestions/solutions and any discussions on what I think the issue is.

Thanks for your time in advance!

-switch

 

[switcheHHH]

Are you overclocking anything?
It could be anything to be honest, mobo, cpu, gpu or ram

You could run a memtest to check the ram, and something like prime to check the cpu, and maybe heaven to check the gpu, you could also try the gpu in another machine if you know anyone else with a PC

Not overclocking anything.
I have already checked the hardware, I have had no issues but I can do that again tonight.

Run Asus RealBench for 15 minutes whilst using a HW Monitor program to check your temperatures.

Asus Realbench will stress the CPU and the RAM and you can see if anything is getting too hot.

It's 100% not overheating I can assure you of that.
idle temp of CPU is 4C when in a high end game its not exceeding 35C
idle temp of GFX is 20C when in a high end game its not exceeding 40C
But I can check this again when I get home.

 

[switcheHHH]

What kind of temp is your FX-8 Eight Core 9590 Black Edition 4.70GHz hitting during gaming? It might be overheating without very capable CPU cooler...


Idle temp means nothing...it's is keeping load temp under control that's important.

Just to let you know, although your CPU is at "stock" clock, it is actually the same chip as the FX8350, with a factory overclock of +700MHz...so I would imagine the temp under load would be as high as a heavily overclocked 8350, and could easily pulling over 200W-250W when under load.

For 8350, generally one would need "good" CPU cooler for around 4.50GHz, but for 4.60GHz and beyond, "very good" CPU cooler would be required.

This is the cooler I have.
http://www.overclockers.co.uk/showproduct.php?prodid=HS-018-CS&groupid=701&catid=2330&subcat=2342

Well I don't think the load is struggling, what would be the best program to monitor it with?

Also I am using MSI afterburner for my GPU fan, would that be causing an issue as I've heard that MSI overclocks stuff without you knowing. I went through all the settings in MSI, I have nothing set to overclock.

 

[switcheHHH]

Idle temp of 4°C? You must be reading the wrong value.

No CPU would idle at 4°C unless it had liquid nitrogen to cool it or something.

I could be looking wrong but that's what everything is telling me.. lol

 

[switcheHHH]

Try underclocking your FX9590 to 4.6ghz and see if that fixes it.



AMD sensors aren't really all that accurate, especially at idle.

I am fairly green to over/under clocking as I have never really needed to do it, could you possibly guide me through the process?

First of all I am going to download Prime95 and Asus realbench and post the results.

 

[switcheHHH]

Ok I'm going to go off track here...

if.. i was using a PCI-E 3.0 graphics card in a ....PCI-E 2.0 port, would that cause the problem I am having.

 

[switcheHHH]

i had the same issue, and it was because of the ram.
the mobo used to run my ram which is 1600mhz and 1.65Volt, underclocked and undervolted to 1333mhz and 1.5Volt, when i reseted the voltage, the PC became more stable.
but in the end i changed the ram altogether and got 1.5v, not a single crash for couple months now.

Have you got the latest bios?

9590 isn't officially supported on that board. And, though I would imagine it should still function, that could be why you're getting wonky temps being reported.

My RAM was underclocked at 1.5v, I changed it to the recommended voltage of 1.65v, still crashing however.
I checked bios again and it is running just under 1.70v although it is set to 1.65v.
am I right for guessing there is some kind of tolerance?

Also I think I have but I will double check that one as well.

 

[switcheHHH]

Run Asus RealBench for 15 minutes whilst using a HW Monitor program to check your temperatures.

Asus Realbench will stress the CPU and the RAM and you can see if anything is getting too hot.

Hi, I did that. I don't know if any of the below helps. (Sorry for picture being large)

Are you overclocking anything?
It could be anything to be honest, mobo, cpu, gpu or ram

You could run a memtest to check the ram, and something like prime to check the cpu, and maybe heaven to check the gpu, you could also try the gpu in another machine if you know anyone else with a PC

I ran Prime95, I did crash one time while using it but I also got an error and I have some results as below for everyone to look at. I'll be honest I am not sure what I need to be looking for but I did the tests so hopefully we can narrow this down. (Sorry for pictures being large)

Note: Each test has a fresh HW monitor open so everything listed on it is from that current test.

 

[switcheHHH]

Your cpu core temp is 71 degrees. That's way too high, re-seat your cooler and check again or check fan speed on tour cpu cooler in bios.

Keep in mind the 71C is when the stress test was run, so the CPU was at max load. that's not it's idle temp

 

[switcheHHH]

No disrespect to shady, but he doesn't seem to take into account of the CPU you are using, and thinking along the line of Intel i5 CPU.

Your 79~82C looks about right for a FX8 CPU at 4.70GHz without top-tier cooling. Looking at the prime95 result, the one where you got error where the Worker #8 stopped working would imply that one of the eight cores is not stable enough, and this could be what cause your crashes in games.

So to put simply, the CPU is not getting enough voltage for running at 4.70GHz stable, though it might be a case giving the core voltage a tiny bump might help stabilise it, but it will increase the power consumption thus increasing the temperature further. If that doesn't work, it could be the case of the motherboard not being good enough at pushing the CPU at that clock speed...

The 9590 have a TDP rating of 220W, but your motherboard spec only listed as "Supports CPU up to 140 W" and only with "6 +2 Phase Power"...this could be the problem. I got a feeling that if you lower the clock speed and the core voltage, your problem would probably go away...

I kind of want to cry now haha. What speed would you recommend lowering it to?
Also what motherboard would you recommend?
Would you also be able to recommend a better tier cooler as well?

thanks for your feedback, that was helpful.

 

[switcheHHH]

Motherboard is fine, you are overheating, you need to investigate if your h90 is working properly.

So my motherboard even though it doesn't seem to support my CPU and from
Marine's comment
''The 9590 have a TDP rating of 220W, but your motherboard spec only listed as Supports CPU up to 140 W" and only with "6 +2 Phase Power"...this could be the problem.''

My issue is 100% overheating?

EDIT: that comment seems to come across a bit rude, sorry. I just want to be sure, because I got be going back an forth like a yoyo and it will cost money every time I change something. I appreciate everyone's input, I would rather it be overheating. Could you recommend me a GOOD cooler for my CPU?

 

[switcheHHH]

Your cpu temps are dangerously high.
100% overheating, whilst the h90 isn't that great you should be seeing cpu temps of 60's and core temps of late 50's with that cooler,

Max cpu temps should be 70c and package 60c

I've pushed a fx8320 on a m5a97 evo r2.0 using 1.55v and it was fine.
Your board is similar to that board I used.

Okay, understood.
So reseat my Cooler, though I am 100% sure it was on correct... but I will redo it. (Then again how can it if I get those temps )

Would you be able to recommend me a good cooler for this CPU as I need another one anyway, so I could use my current one on the second system I have.

 

[switcheHHH]

Get yourself in here and we'll sort you out.
http://forums.overclockers.co.uk/showthread.php?t=18650181

Okay I will post there shortly!

Is you rad hot to touch when the cpu is at those temps?

From idle how quickly do your idle temps ramp up to those peak load temps we see,
is it almost instantly or gradual ?

Well those temps come from stress tests, I need to monitor these temps in game I think. But I would say gradual seeing as I can play a game for a while, it's very random. No not too hot I don't think.

 

[switcheHHH]

Actually I just noticed this:
http://www.overclockers.co.uk/showproduct.php?prodid=CP-360-AM

"Voltage: 1.4125v - 1.6500v"

This could be why the OP's 9590 is running hotter than someone's fine-tuned overclocked FX8350...

So basically if the particular 9590 was an unfortunate "duff chip", up to 1.6-1.65v would be used?

This tool will show the p-states of ops fx9590
http://www.vrworld.com/2013/04/09/amdmsrtweaker-v11-released2c-trinity-support-inside/

From hwmonitor the most I saw was 1.46v which is average.
2013 silicon is more refined than early 2012 release silicon.

In regards to the above, is it safe to say this IS an overheating issue and that my motherboard is indeed okay?

Will I need to underclock my CPU?

and I take it you want me to use that tool?

 

[switcheHHH]

I'd imagine if the Corsair AIO is seated correctly, PWM fan running correctly, pump working correctly it is more than capable as I'd imagine it will be much much better than the stock cooler that chip would be sent out with?

Yes it is, I was using this cooler on my previous setup.

You know I was looking at this again whilst at work, and like Cheesyboy he's noticed that hwmonitor isn't showing the whole information, like when a chipset doesn't support the latest hwmonitor. This seems related to when a cpu works in a motherboard but isn't supported. Whilst the fx9590 works I think it's bios string prevents the real details from being displayed,

Op is the motherboard actually an M5A99X EVO R2.0 and not a M5a99fx pro r 2.0 ?

Davedree, the exact mobo I ordered from here is as follows:
https://www.overclockers.co.uk/showproduct.php?prodid=MB-560-AS

 

[switcheHHH]

Thanks, yeah It can be seen it's the m5a99fx pro r.20 just missing the f in the advert.
This board as we've already said is only a 6+2 phase, but as long as it has the latest bios installed then it does officially support fx9370 and 9590, with good cooling.

Is the radiator and heatsink hot to touch, when stress testing ?

I am not 100% sure if I have the latest bios installed so that will be next on the list.

I just tried the first stress test on Prime95 and crashed out around 85C
the pump feels like its working, the water is warm. the rads and grill are 'warm' at best......
.....so time to reseat?

Is it possible I have applied too much paste and its causing a gap?

 

[switcheHHH]

Almost certainly a heat issue. 85C is way too hot, even for an FX. I've always aimed to keep mine below 60C ([email protected]). With my current setup it never goes above 45C under load.

Your cooler is certainly up to the task when working correctly, so the issue appears to be with transferring the heat from the CPU to the block to the radiator and into the air.

Is your rad configured as intake or exhaust?

configured? o.o

I'll be honest when this was on my other board, it was already pre configured I'm guessing?

Before that I've always had a Zalman heatsink + fan (but that was back on my DDR2 setup)

So do I configure in the bios?

This tool will show the p-states of ops fx9590
http://www.vrworld.com/2013/04/09/amdmsrtweaker-v11-released2c-trinity-support-inside/

From hwmonitor the most I saw was 1.46v which is average.
2013 silicon is more refined than early 2012 release silicon.

I ran the tool for you. Don't know if this will help at all.

 

[switcheHHH]

Configured - how it's installed. Does the fan pull air in to the case or exhaust it out?

Oh, out.

 

[switcheHHH]

The HWmonitor that I have used above is 1.23.0 driver version 136 which I have had since 2013, seems to work fine for the FX series although it is more to do with the chipset on the motherboard.

I had my FX8350 in an ASUS M5A99x R1.0 motherboard when I first had it, It was 2012 silicon, one of the first. The mobo / CPU would bench at 5GHz readily x264, cinebench etc. But would only prime at 4.5GHz I could prime on 4 workers at 4.9GHz. I could game at 4.7GHz.

Therefore I believe that the motherboard is fine and the cooler is suspect. I far prefer large air coolers (or custom water) however I am back to air cooling now. I would go for a Phanteks twin tower over any AIO cooler out there. The heligon is good, but I got it as B grade.

Literally about to upgrade my bios and chip-set now (if they are not current already)
Thanks for your feedback on the cooler also!

 

[switcheHHH]

Great thanks, I've learnt something from your cpu too, it seems the fx9590 does have the peak 5ghz p-state p0. But it doesn't have a half load 4.8ghz turbo at p-state p1, instead it's 4.7ghz, and 4.5ghz baseclock, Looks like my sources were wrong.

Ok well you're peak cpu v-core is 1.475v which if don't have LLC enabled in your bios would relate to the vdroop seen at 1.46v shown in hwmonitor.

I'm not suggesting to replace your psu, but just to be aware that in my experience of excessive temperatures many years ago. My psu was once the cause of instability and stupidly high temperatures. I replaced the psu with a better one and the temps and stability was cured. So could be something to look at if you have a spare psu or someone you can swap with to test.

Back to the PSU again haha damnit

Well put it this way, I am about to update my bios, then I'll reseat my Cooler.
If I get nothing after that I will have to cave in and get a new PSU.

Just so I am clear, my rad/grill needs to be HOT when stress testing yes?