How to solve AMD FX 9590 Freezing issue and stablize it?

[SAGARTX31]

I have AMD fx 9590 installed in ASUS TUF Sabertooth 990fx r3.0. Bios version 0212. but my system gets freeze whenever I am playing games like witcher 3, Gta 5, crysis 3,etc.
CPU- AMD FX 9590
Mobo- ASUS Sabertooth 990fx R3.0
Cooling- Coolermaster masterliquid Pro 280
PSU- Corsair RM 1000i
Ram - 16 GB (Kingston Hyperx fury 1866Mhz 8GBx2)
GPU - XFX RX480 Black edition 1328mhz

Please Help me what changes Should I do in BIOS or anything? System keeps freezing when there is a load on it about 3-4 weeks ago there was nothing happening like this but It is now happening. The Incident happens as the game is working smoothly but suddenly the screen freezes and nothing works, no keyboard, no mouse nothing just like someone placed a still picture on the screen. Please help guys Please.

 

[Paul NZ]

Test the ram with memtest www.memtest.org

 

[SAGARTX31]

Test the ram with memtest www.memtest.org

Memory is ok nothing with that but dont know where is the problem in CPU what to do?

 

[Paul NZ]

How many passes did it run for?

 

[8350rocks]

Check your temps.

If you are breaching 60-65C at the socket, your cooling system likely has a bad pump, or needs the radiator blown out with air to remove dust.

Source: FX9590 owner (see CPUZ in the bottom right hand corner)

 

[MeanMachine41]

Hi SAGARTX31

The FX-9590 has certain special requirements. It is a beast of processor and many have been challenged with getting it stable however it can be tamed.
Some changes in Bios are required for stability and I can help with that.

First make sure your Bios is up to date. (latest is 0212) available here: https://www.asus.com/Motherboards/TUF-SABERTOOTH-990FX-R3-0/HelpDesk_Download/
Best method is using your rear IO panel and a USB stick. This way you don't need to access Bios or your OS system. (page 2-11 in your MB manual)

Once done go into your Bios>AI Tweaker menu> select D.O.C.P. This will set your CPU bus and ratio in conjunction with the RAM you are using.
The Bios will set for stock frequency with most parameters set to Auto. (for now)
Conduct a stress test using AIDA64. Reason is we need to establish a base line for OCing and determine that temps,CPU frequency and RAM are reporting well.

Download AIDA64 and CPUid HWMonitor and place them on your desktop together.
In the TOOLS section of AIDA64, Run the stress test for 10mins with check boxes ticked for CPU, FPU and Cache only. Take screen shots at the ten minute mark and use your favorite file host to link the results here for analysis. You can use IMIGUR as a file host, Just upload the screen shots select your IMAGES and obtain the BBurl as a link here.

I want to see all rail voltages in HWMonitor, 100% utilization on all Cores and your temperatures under load. Once done your Bios settings for initial OC will be given.
Any issues just ask and please report back.

 

[SAGARTX31]

Hi SAGARTX31

The FX-9590 has certain special requirements. It is a beast of processor and many have been challenged with getting it stable however it can be tamed.
Some changes in Bios are required for stability and I can help with that.

First make sure your Bios is up to date. (latest is 0212) available here: https://www.asus.com/Motherboards/TUF-SABERTOOTH-990FX-R3-0/HelpDesk_Download/
Best method is using your rear IO panel and a USB stick. This way you don't need to access Bios or your OS system. (page 2-11 in your MB manual)

Once done go into your Bios>AI Tweaker menu> select D.O.C.P. This will set your CPU bus and ratio in conjunction with the RAM you are using.
The Bios will set for stock frequency with most parameters set to Auto. (for now)
Conduct a stress test using AIDA64. Reason is we need to establish a base line for OCing and determine that temps,CPU frequency and RAM are reporting well.

Download AIDA64 and CPUid HWMonitor and place them on your desktop together.
In the TOOLS section of AIDA64, Run the stress test for 10mins with check boxes ticked for CPU, FPU and Cache only. Take screen shots at the ten minute mark and use your favorite file host to link the results here for analysis. You can use IMIGUR as a file host, Just upload the screen shots select your IMAGES and obtain the BBurl as a link here.

I want to see all rail voltages in HWMonitor, 100% utilization on all Cores and your temperatures under load. Once done your Bios settings for initial OC will be given.
Any issues just ask and please report back.

I will try Thanks for help I will contact you after tests

 

[SAGARTX31]

Check your temps.

If you are breaching 60-65C at the socket, your cooling system likely has a bad pump, or needs the radiator blown out with air to remove dust.

Source: FX9590 owner (see CPUZ in the bottom right hand corner)

Temps are between 60-65 but for cpu for socket temps are ~65

 

[8350rocks]

Check your temps.

If you are breaching 60-65C at the socket, your cooling system likely has a bad pump, or needs the radiator blown out with air to remove dust.

Source: FX9590 owner (see CPUZ in the bottom right hand corner)

Temps are between 60-65 but for cpu for socket temps are ~65

That is your culprit then. Increase your cooling, or the air circulation in your case.

 

[MeanMachine41]

Hi SAGARTX31

The FX-9590 has certain special requirements. It is a beast of processor and many have been challenged with getting it stable however it can be tamed.
Some changes in Bios are required for stability and I can help with that.

First make sure your Bios is up to date. (latest is 0212) available here: https://www.asus.com/Motherboards/TUF-SABERTOOTH-990FX-R3-0/HelpDesk_Download/
Best method is using your rear IO panel and a USB stick. This way you don't need to access Bios or your OS system. (page 2-11 in your MB manual)

Once done go into your Bios>AI Tweaker menu> select D.O.C.P. This will set your CPU bus and ratio in conjunction with the RAM you are using.
The Bios will set for stock frequency with most parameters set to Auto. (for now)
Conduct a stress test using AIDA64. Reason is we need to establish a base line for OCing and determine that temps,CPU frequency and RAM are reporting well.

Download AIDA64 and CPUid HWMonitor and place them on your desktop together.
In the TOOLS section of AIDA64, Run the stress test for 10mins with check boxes ticked for CPU, FPU and Cache only. Take screen shots at the ten minute mark and use your favorite file host to link the results here for analysis. You can use IMIGUR as a file host, Just upload the screen shots select your IMAGES and obtain the BBurl as a link here.

I want to see all rail voltages in HWMonitor, 100% utilization on all Cores and your temperatures under load. Once done your Bios settings for initial OC will be given.
Any issues just ask and please report back.

I will try Thanks for help I will contact you after tests

Did you manage to conduct the stress tests as requested.?

 

[Karadjgne]

I'm going to assume from the quality parts that you do have adequate ventilation in the case.
That said, that CM mlpro 280 should be able to keep up with that cpu.
That said, you have just 2 options, first is play with the fan curve settings for both cpu and case fans, get them moving faster at lower temps, they'll need to be at maximum (or close) when you hit @50-55°C at the socket. Case fans should be moving closer to max as well when gaming like that or other high cpu use apps. Second option is overclock the cpu. Not as in raise speed to match the voltage, but lower the vcore voltage as much as possible and still maintain stability (hard with that particular cpu). You might have to adjust other settings such as LLC etc to keep the cpu stable.
Preferably, use both options together as that'll maximize airflow and minimize cpu heat output at its current speed.

Also, screen freeze can also be gpu related, specifically the vram/vrm area of the gpu, not the processor side. This is an issue where there is no definitive way of checking without the use of a heavy gpu tester like MSI kombuster. There's no temp sensor on that side of the gpu, just the processor, so gpu temps are useless, the vram can easily hit 90°C+, but gpu read 55°C. Make sure your gpu is cleaned out, fins and fans, and both fans are working correctly.
The only other possibility would maybe the gpu itself. Original release 480's had an issue with overdrawn wattage of the pcie slot, pulling in excess of 90w on a 75w rated slot. That was fixed in a later firmware revision, but if you've not had that update happen, it's possible you are just now starting to wear out that pcie slot. Could try moving the gpu down a slot.

 

[SAGARTX31]

Hi SAGARTX31

The FX-9590 has certain special requirements. It is a beast of processor and many have been challenged with getting it stable however it can be tamed.
Some changes in Bios are required for stability and I can help with that.

First make sure your Bios is up to date. (latest is 0212) available here: https://www.asus.com/Motherboards/TUF-SABERTOOTH-990FX-R3-0/HelpDesk_Download/
Best method is using your rear IO panel and a USB stick. This way you don't need to access Bios or your OS system. (page 2-11 in your MB manual)

Once done go into your Bios>AI Tweaker menu> select D.O.C.P. This will set your CPU bus and ratio in conjunction with the RAM you are using.
The Bios will set for stock frequency with most parameters set to Auto. (for now)
Conduct a stress test using AIDA64. Reason is we need to establish a base line for OCing and determine that temps,CPU frequency and RAM are reporting well.

Download AIDA64 and CPUid HWMonitor and place them on your desktop together.
In the TOOLS section of AIDA64, Run the stress test for 10mins with check boxes ticked for CPU, FPU and Cache only. Take screen shots at the ten minute mark and use your favorite file host to link the results here for analysis. You can use IMIGUR as a file host, Just upload the screen shots select your IMAGES and obtain the BBurl as a link here.

I want to see all rail voltages in HWMonitor, 100% utilization on all Cores and your temperatures under load. Once done your Bios settings for initial OC will be given.
Any issues just ask and please report back.

I will try Thanks for help I will contact you after tests

Did you manage to conduct the stress tests as requested.?

The System didn't even lasted for a minute it freezed again when just launched the program

 

[SAGARTX31]

I'm going to assume from the quality parts that you do have adequate ventilation in the case.
That said, that CM mlpro 280 should be able to keep up with that cpu.
That said, you have just 2 options, first is play with the fan curve settings for both cpu and case fans, get them moving faster at lower temps, they'll need to be at maximum (or close) when you hit @50-55°C at the socket. Case fans should be moving closer to max as well when gaming like that or other high cpu use apps. Second option is overclock the cpu. Not as in raise speed to match the voltage, but lower the vcore voltage as much as possible and still maintain stability (hard with that particular cpu). You might have to adjust other settings such as LLC etc to keep the cpu stable.
Preferably, use both options together as that'll maximize airflow and minimize cpu heat output at its current speed.

Also, screen freeze can also be gpu related, specifically the vram/vrm area of the gpu, not the processor side. This is an issue where there is no definitive way of checking without the use of a heavy gpu tester like MSI kombuster. There's no temp sensor on that side of the gpu, just the processor, so gpu temps are useless, the vram can easily hit 90°C+, but gpu read 55°C. Make sure your gpu is cleaned out, fins and fans, and both fans are working correctly.
The only other possibility would maybe the gpu itself. Original release 480's had an issue with overdrawn wattage of the pcie slot, pulling in excess of 90w on a 75w rated slot. That was fixed in a later firmware revision, but if you've not had that update happen, it's possible you are just now starting to wear out that pcie slot. Could try moving the gpu down a slot.

You might be true my gpu temps are 88C at 4000rpm fan speed and cpu temps are 60C at 2700rpm with 2x14mm fans and pump speed is 9000rpm. My room temps are about 37C.

 

[madmatt30]

I'm going to assume from the quality parts that you do have adequate ventilation in the case.
That said, that CM mlpro 280 should be able to keep up with that cpu.
That said, you have just 2 options, first is play with the fan curve settings for both cpu and case fans, get them moving faster at lower temps, they'll need to be at maximum (or close) when you hit @50-55°C at the socket. Case fans should be moving closer to max as well when gaming like that or other high cpu use apps. Second option is overclock the cpu. Not as in raise speed to match the voltage, but lower the vcore voltage as much as possible and still maintain stability (hard with that particular cpu). You might have to adjust other settings such as LLC etc to keep the cpu stable.
Preferably, use both options together as that'll maximize airflow and minimize cpu heat output at its current speed.

Also, screen freeze can also be gpu related, specifically the vram/vrm area of the gpu, not the processor side. This is an issue where there is no definitive way of checking without the use of a heavy gpu tester like MSI kombuster. There's no temp sensor on that side of the gpu, just the processor, so gpu temps are useless, the vram can easily hit 90°C+, but gpu read 55°C. Make sure your gpu is cleaned out, fins and fans, and both fans are working correctly.
The only other possibility would maybe the gpu itself. Original release 480's had an issue with overdrawn wattage of the pcie slot, pulling in excess of 90w on a 75w rated slot. That was fixed in a later firmware revision, but if you've not had that update happen, it's possible you are just now starting to wear out that pcie slot. Could try moving the gpu down a slot.

You might be true my gpu temps are 88C at 4000rpm fan speed and cpu temps are 60C at 2700rpm with 2x14mm fans and pump speed is 9000rpm. My room temps are about 37C.

with those kind of ambient temps youre going to have to downclock & undervolt.
youre 15c above what we in the western hemisphere would consider comfortable temps before you even get to component cooling.

That is a terrible terrible choice of cpu for where you live in all honesty.

 

[SAGARTX31]

I'm going to assume from the quality parts that you do have adequate ventilation in the case.
That said, that CM mlpro 280 should be able to keep up with that cpu.
That said, you have just 2 options, first is play with the fan curve settings for both cpu and case fans, get them moving faster at lower temps, they'll need to be at maximum (or close) when you hit @50-55°C at the socket. Case fans should be moving closer to max as well when gaming like that or other high cpu use apps. Second option is overclock the cpu. Not as in raise speed to match the voltage, but lower the vcore voltage as much as possible and still maintain stability (hard with that particular cpu). You might have to adjust other settings such as LLC etc to keep the cpu stable.
Preferably, use both options together as that'll maximize airflow and minimize cpu heat output at its current speed.

Also, screen freeze can also be gpu related, specifically the vram/vrm area of the gpu, not the processor side. This is an issue where there is no definitive way of checking without the use of a heavy gpu tester like MSI kombuster. There's no temp sensor on that side of the gpu, just the processor, so gpu temps are useless, the vram can easily hit 90°C+, but gpu read 55°C. Make sure your gpu is cleaned out, fins and fans, and both fans are working correctly.
The only other possibility would maybe the gpu itself. Original release 480's had an issue with overdrawn wattage of the pcie slot, pulling in excess of 90w on a 75w rated slot. That was fixed in a later firmware revision, but if you've not had that update happen, it's possible you are just now starting to wear out that pcie slot. Could try moving the gpu down a slot.

You might be true my gpu temps are 88C at 4000rpm fan speed and cpu temps are 60C at 2700rpm with 2x14mm fans and pump speed is 9000rpm. My room temps are about 37C.

with those kind of ambient temps youre going to have to downclock & undervolt.
youre 15c above what we in the western hemisphere would consider comfortable temps before you even get to component cooling.

That is a terrible terrible choice of cpu for where you live in all honesty.

I feel it from past 10-12days that you are right

 

[SAGARTX31]

I'm going to assume from the quality parts that you do have adequate ventilation in the case.
That said, that CM mlpro 280 should be able to keep up with that cpu.
That said, you have just 2 options, first is play with the fan curve settings for both cpu and case fans, get them moving faster at lower temps, they'll need to be at maximum (or close) when you hit @50-55°C at the socket. Case fans should be moving closer to max as well when gaming like that or other high cpu use apps. Second option is overclock the cpu. Not as in raise speed to match the voltage, but lower the vcore voltage as much as possible and still maintain stability (hard with that particular cpu). You might have to adjust other settings such as LLC etc to keep the cpu stable.
Preferably, use both options together as that'll maximize airflow and minimize cpu heat output at its current speed.

Also, screen freeze can also be gpu related, specifically the vram/vrm area of the gpu, not the processor side. This is an issue where there is no definitive way of checking without the use of a heavy gpu tester like MSI kombuster. There's no temp sensor on that side of the gpu, just the processor, so gpu temps are useless, the vram can easily hit 90°C+, but gpu read 55°C. Make sure your gpu is cleaned out, fins and fans, and both fans are working correctly.
The only other possibility would maybe the gpu itself. Original release 480's had an issue with overdrawn wattage of the pcie slot, pulling in excess of 90w on a 75w rated slot. That was fixed in a later firmware revision, but if you've not had that update happen, it's possible you are just now starting to wear out that pcie slot. Could try moving the gpu down a slot.

You might be true my gpu temps are 88C at 4000rpm fan speed and cpu temps are 60C at 2700rpm with 2x14mm fans and pump speed is 9000rpm. My room temps are about 37C.

with those kind of ambient temps youre going to have to downclock & undervolt.
youre 15c above what we in the western hemisphere would consider comfortable temps before you even get to component cooling.

That is a terrible terrible choice of cpu for where you live in all honesty.

Hey friend may corsair h100i v2 will get the work done for cooling as I have seen this
https://www.youtube.com/watch?v=VZxjobUE1tQ&t=347s
Am I right please tell me

 

[madmatt30]

I don't think any cooler will do the job properly with those ambient temps honestly.

The cooler you have is decent but if its set up in an exhaust configuration then the air being pushed through it from inside your case is likely 50c +

Your CPU & GPU fans are already maxed out & those temps are still high.

What's your case & the rest of your case fan setup ??

You need to have a proper rethink about airflow with your room temps being so high.

I think you're going to have to limit both CPU & maybe GPU clocks for stability (88c on the GPU is close to the danger zone , are you running unlimited fps or vsynced ?? Because id think about locking it down to 60fps if you're not already)

For starters go in bios , disable turbo & set a 22 multiplier for 4.4ghz , it'll drop 10c off your CPU load temps instantly.

 

[MeanMachine41]

I agree with the others. With an ambient room temperatures of 37C you not going to achieve stability with the FX-9590 unless you Air-condition the room or drop your Core voltage and frequency.
Certainly not the best climate conditions for the FX-9590.

 

[Karadjgne]

Yeah. 37°C is brutal for ambients. Part of the issue is that the radiator for the cpu is pulling heated air from the case once the gpu gets cranking
This lowers the ability of the rad to effectively remove heat from the coolant. And you can't reverse the fans or you'll dump the cpu heat on the gpus intake, doing the same to that. What's needed is separation of the airflow or a way to eliminate both sources of heat at the same time. If you really really want to keep this system cool running under stressful situations like gaming, a full custom loop with gpu and cpu both on blocks and an overkill amount of radiator space. Not cheap. Either that or invest in a window mount AC unit and drop the ambients back into the lower 20's when you want to game heavily.

 

[SAGARTX31]

I don't think any cooler will do the job properly with those ambient temps honestly.

The cooler you have is decent but if its set up in an exhaust configuration then the air being pushed through it from inside your case is likely 50c +

Your CPU & GPU fans are already maxed out & those temps are still high.

What's your case & the rest of your case fan setup ??

You need to have a proper rethink about airflow with your room temps being so high.

I think you're going to have to limit both CPU & maybe GPU clocks for stability (88c on the GPU is close to the danger zone , are you running unlimited fps or vsynced ?? Because id think about locking it down to 60fps if you're not already)

For starters go in bios , disable turbo & set a 22 multiplier for 4.4ghz , it'll drop 10c off your CPU load temps instantly.

You are right I am at unlimited fps and vsynced mode I will try to limit the fps count and th 88c is only my top gpu temps which is not good I know and one more thing my gpuis running at 1328mhz both ones but only top one goes to that high lower one is only under 60C I think I should clean my pc as like as it was new.

 

[madmatt30]

You're running 2 x 480s in crossfire too ?????

Wow mate , if correct I don't know what to say , you're running a high powered & very hot running system in ambient temps that are close to double what most of us would be happy with.

 

[Karadjgne]

Kinda missed that too about the CF 480's. Musta gotten lost in translation somewhere.
Anyways, speed of the gpus is almost irrelevant, they'll work in tandem at the slowest cards speed, so if top card has highest, both will run at lower cards speed. Reason bottom card is always colder than top card is two-fold. First is that the top card is doing most of the work, using resources from the bottom card. It's not an even 50/50 split of the workload. Second is that the top card suffers from heat intake from the lower cards exhaust, so top cards have lower efficiency to start with.

There is an absolute must to get ambient/case temps lowered, mechanical cooling (cpu and gpu) cannot physically go below ambient temps. You are starting with 37ish°C, then start adding in the rising case temps as that CF starts dumping heat into the case. By the time air is drawn through the radiator it's closer to 45-50°C. That's what is killing the cpu, minimum temps at max airflow will be @10-12°C higher on the cpu. That puts the cpu at @55-62°C at minimums. Even a decent Intel cpu will be 6-8°C higher, BUT, Intel cpu's can go upto @100°C before stalling, so running Intel would be a viable option, temps would be acceptable at @60°C gaming.
The only other options I can see are a window unit AC to drop ambients by @10° or better, or a full custom loop liquid cooling, which would start out by eliminating most of the gpu's physical heat exhaust into the case.

 

[SAGARTX31]

I think There might be an issue with the rx480's as few days ago when my room temps were at 27c for hours the gpu temps were
Top Rx480- 48c Idle; 70-75c load 100%
Bottom- 39c( lowest temps); 59c Load

these are the results when I was playing DISHONORED 2 .
I dont know who is the culprit of the system freezing the cpu (fx 9590) or The Gpu's( XFX RX 480 Black Edition)
These are my GPU's- http://xfxforce.com/en-us/products/amd-radeon-rx-400-series/rx480be-rx-480m8bba6
model No.- 480M8BA
Part No.- RX-480M8BBA6

 

[Karadjgne]

Nope that looks about right for xfire temps. Your case temps will run @10-12°C above ambient, its just the heat from the gpus, VRM's heatsinks, hdds etc heating up the inside of the case. At idle the bottom gpu is the one to read, as it's unaffected by the other cards heated exhaust. Top cards always run hotter, both barrel-fan and axis-fan designs. The top card is the one to read at load, and 100% at 70-75°C is actually pretty good. So the gpus are fine.

What you didn't say was the cpu temps when the ambient was 27°

 

[SAGARTX31]

Nope that looks about right for xfire temps. Your case temps will run @10-12°C above ambient, its just the heat from the gpus, VRM's heatsinks, hdds etc heating up the inside of the case. At idle the bottom gpu is the one to read, as it's unaffected by the other cards heated exhaust. Top cards always run hotter, both barrel-fan and axis-fan designs. The top card is the one to read at load, and 100% at 70-75°C is actually pretty good. So the gpus are fine.

What you didn't say was the cpu temps when the ambient was 27°

I think the CPU temps are 42-47C at Idle and 60-77 at 90-99% load
Ambient room temps are 33-37C
1 more thing to notice is that the VCORE-1 temps are 71-82C at load and 50-60C at idle
and VCORE-2 temps are about 37-40C at idle and 60-77 at load
Actually I installed Noctua NF-A4x10 near VCORE-2
The PCIEx16 top temps are 60-67C when playing heavy games like Dishonored-2 and 45-56C at idle or normal tasking
and the bottom gpu pcie temps are about 40-46C at idle and 50-59/60C while playing games