locked CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update


Amos Sobel 4X4MF
 

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

  1. I did change the CPU power plan “power Saver” CPU frequency became variable at 1000-4000MHz. Decode rate did drop by half. JTAlert did crash again at 22:42and again at 23:30 and again at 23:52.  Max temp was 74 C.
  2. The CPU diagram was this. CPU load does not touch 100%.

 

So, it isn’t CPU temperature and it isn’t CPU load. What is it then? And why is it that only JTAlet fail again and again?

 

Amos 4X4MF

 

=================================================

Bill

 

Thank you for the elaborate discussion of the problem.

 

My computer has I7 4790K 4.0GHz CPU which is 4th generation 4 core 8 threads. It does fit your description. It is 6 years old and has no additional upgrade path.

The fact is that it does manage the CPU load presented by SmartSDR, HRD logbook, Outlook  plus the 4 instances of WSJT-X and JTAlert without throttling only that JTAlert does crash from time to time when computational load does increase to 100%..

Following your hint, I will check today if CPU temperature comes to the limit of throttling.  In that case upgrade of the CPU fan might help.

To prove the case, I will try 3 instances. I have done that in the past but there was no definite  conclusion.

 

Can you come up with an additional productive idea?

 

Many thanks

 

Amos  4X4MF

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Bill Somerville
Sent: Friday, September 4, 2020 11:31 PM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash

 

On 04/09/2020 20:17, Amos Sobel 4X4MF wrote:

I am using 4 instances of WSJT-X v2.2.2 and JTAlert fed by  Flex Radio 6600 .Often, especially in times of FT8 intense activity, one or more of my JTAlert instances does crash. Looking at the Resource Monitor images I have added here makes me conclude that the crashes are caused by the peaks of the CPU usage which come to 100%. The CPU load of our software did increase with the last updates. Consequently I did update my computer to it's maximum. It does currently use Intel 4 core I7 processor and 24 Gbt RAM.
1. Is there a way to flatten the peaks of the CPU usage by WSJT-X and JTAlert?
2. Is there a way to configure the computer such as to prevent it from reaching 100% CPU usage?
3. Is there a way to make JTAlert insensitive to 100% CPU peaks?
4. Does anybody use my configuration of 4 instances? With which computer configuration?

Amos 4X4MF

Amos,

WSJT-X, when using slow modes, typically uses one CPU thread for decoding. Because decoding is almost 100% computational it will use that CPU thread at 100% until it has finished. If other processes are consuming all CPU threads then decoding threads may be delayed, there's nothing particularly problematic about that, decoding will just take a little longer. Note that a Core i7 quad core CPU can run up to 8 independent CPU threads each at 100%. Only when all 8 CPU threads are at 100% utilization will the aggregate CPU usage indicate 100%.

An SDR is largely I/O bound although it also has non-trivial computational load, and graphic processing. JTAlert should be OK waiting for CPU resource, it will just slow down if it can't get enough CPU resource.

Running an SDR console and 4 instances of WSJT-X will likely require up to 100% aggregate CPU resource for short periods, that's fine but make sure your CPU has adequate cooling because the CPU management hardware will throttle the maximum CPU speed severely if it detects overheating. The same applies to graphics cards which often do a lot of processing.

In summary, short peaks of high CPU load are absolutely fine when you are running compute intensive tasks like DSP. Continuous excessive CPU usage, particularly if the maximum clock rate has been throttled down by temperature management.

Your resource monitor summary does show some throttling of CPU performance (the blue CPU line dropping below 100%), but that may be due to idle periods when there is no demand for CPU. Note the orange CPU line is the net CPU usage of WSJT-X and jt9, and I'll bet is is mostly jt9 doing the decoding. The CPU usage of those two together is considerably less that one CPU thread capacity. You need to check the select boxes on every instance of  WSJT-X and jt9 to show their net usage of resources with the orange trace. Here's my Core i7 laptop running two instances of WSJT-X in FT8 mode:

73
Bill
G4WJS.


M0PWX
 

do you you have anti virus on the PC?

try disabling it for a while,

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

i see this on servers at work

peter

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 





This email has been scanned by iomartcloud.
http://www.iomartcloud.com


Amos Sobel 4X4MF
 

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com



Amos Sobel 4X4MF
 

Bill, Mike, Peter and Neil

 

  1. I did update the BIOS of my computer. JTAlert does continue to crash on heavy loads.
  2. I did reduce the number of instances from 4 to 3. JTAlert does continue to crash on heavy loads.
  3. I did reduce the number of instances from 3  to 2. JTAlert does stop  crashing  on heavy loads.
  4. Does anybody use 4 instances successfully? Using which computer?

 

Amos 4X4MF

 

================================

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com



neil_zampella
 

Amos,

did you reduce the decode from DEEP  and turn off AP?

Neil, KN3ILZ

On 9/6/2020 7:22 AM, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

  1. I did update the BIOS of my computer. JTAlert does continue to crash on heavy loads.
  2. I did reduce the number of instances from 4 to 3. JTAlert does continue to crash on heavy loads.
  3. I did reduce the number of instances from 3  to 2. JTAlert does stop  crashing  on heavy loads.
  4. Does anybody use 4 instances successfully? Using which computer?

 

Amos 4X4MF

 

================================

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com




    


Bruce N7XGR
 

Amos,  You just answered your own problem.  Get another PC with equal or greater
processing power and have only 2 X of WSJT-X and 2 X of JT Alert running.
It is madness to have that much running at the same time, a recipe for disaster.

Bruce  N7XGR


On Sun, Sep 6, 2020 at 8:22 AM Amos Sobel 4X4MF <asobel@...> wrote:

Bill, Mike, Peter and Neil

 

  1. I did update the BIOS of my computer. JTAlert does continue to crash on heavy loads.
  2. I did reduce the number of instances from 4 to 3. JTAlert does continue to crash on heavy loads.
  3. I did reduce the number of instances from 3  to 2. JTAlert does stop  crashing  on heavy loads.
  4. Does anybody use 4 instances successfully? Using which computer?

 

Amos 4X4MF

 

================================

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com




Amos Sobel 4X4MF
 

Neil

 

Using decode=fast, JTAlert does crash too!

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of neil_zampella
Sent: Sunday, September 6, 2020 4:30 PM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

Amos,

did you reduce the decode from DEEP  and turn off AP?

Neil, KN3ILZ

On 9/6/2020 7:22 AM, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

  1. I did update the BIOS of my computer. JTAlert does continue to crash on heavy loads.
  2. I did reduce the number of instances from 4 to 3. JTAlert does continue to crash on heavy loads.
  3. I did reduce the number of instances from 3  to 2. JTAlert does stop  crashing  on heavy loads.
  4. Does anybody use 4 instances successfully? Using which computer?

 

Amos 4X4MF

 

================================

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com




 


m_scot_alex
 

Bruce -

I know this is a little difficult to follow as Amos seems to be starting multiple threads on this topic, but I run eight instances of WSJT-X and JT Alert with Flex Radio SmartSDR and a couple of utilities running with around 55% peak CPU utilization. He has more than enough system resources to do this as well, as a lot of people (including me) do this with less 'powerful' CPUs, so there's something unique to his system that's preventing this from working. A couple of us have asked questions and posted ideas, but haven't go to the root of the problem - yet.

But simply running multiple instances on a 4GHz i7 isn't the problem.

73,

Mike - N8MSA


On Sun, Sep 6, 2020 at 10:00 AM, Bruce N7XGR wrote:
Amos,  You just answered your own problem.  Get another PC with equal or greater
processing power and have only 2 X of WSJT-X and 2 X of JT Alert running.
It is madness to have that much running at the same time, a recipe for disaster.
 
Bruce  N7XGR

On Sun, Sep 6, 2020 at 8:22 AM Amos Sobel 4X4MF <asobel@...> wrote:

Bill, Mike, Peter and Neil

 

  1. I did update the BIOS of my computer. JTAlert does continue to crash on heavy loads.
  2. I did reduce the number of instances from 4 to 3. JTAlert does continue to crash on heavy loads.
  3. I did reduce the number of instances from 3  to 2. JTAlert does stop  crashing  on heavy loads.
  4. Does anybody use 4 instances successfully? Using which computer?

 

Amos 4X4MF

 

================================

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com



Amos Sobel 4X4MF
 

Mike

 

Thanks for encouraging me no to go on and  buy even more expansive computer which may have the same problem..

Today I had a little surprise:

I did forget to turn on my HRD logbook and the system did survive for a long time. Later on I did force it to the busy 4x 14.074MHz and it did crash after an hour.

Returned it to the normal 7, 10, 14, 18MHz without HRD and it did not crash since then.

Whenever I need to log a DX, there are few this days, I turn HRD on and off.

This is an early joy but maybe there is a hint in it.

 

Bruce

 

I am not sure it is possible to run SmartSDR on 2 computers connected to 1 Flex 6600.

 

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of m_scot_alex
Sent: Monday, September 7, 2020 1:16 PM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

Bruce -

I know this is a little difficult to follow as Amos seems to be starting multiple threads on this topic, but I run eight instances of WSJT-X and JT Alert with Flex Radio SmartSDR and a couple of utilities running with around 55% peak CPU utilization. He has more than enough system resources to do this as well, as a lot of people (including me) do this with less 'powerful' CPUs, so there's something unique to his system that's preventing this from working. A couple of us have asked questions and posted ideas, but haven't go to the root of the problem - yet.

But simply running multiple instances on a 4GHz i7 isn't the problem.

73,

Mike - N8MSA

On Sun, Sep 6, 2020 at 10:00 AM, Bruce N7XGR wrote:

Amos,  You just answered your own problem.  Get another PC with equal or greater

processing power and have only 2 X of WSJT-X and 2 X of JT Alert running.

It is madness to have that much running at the same time, a recipe for disaster.

 

Bruce  N7XGR

 

On Sun, Sep 6, 2020 at 8:22 AM Amos Sobel 4X4MF <asobel@...> wrote:

Bill, Mike, Peter and Neil

 

  1. I did update the BIOS of my computer. JTAlert does continue to crash on heavy loads.
  2. I did reduce the number of instances from 4 to 3. JTAlert does continue to crash on heavy loads.
  3. I did reduce the number of instances from 3  to 2. JTAlert does stop  crashing  on heavy loads.
  4. Does anybody use 4 instances successfully? Using which computer?

 

Amos 4X4MF

 

================================

Peter

 

I am using the Norton360 anyi-virus.

Today I did exclude all WSJT-X and JTAlert related folders.

JTAlert does continue to fail on heavy loads.

 

Amos 4X4MF

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of Peter SWL2370
Sent: Sunday, September 6, 2020 1:08 AM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

do you you have anti virus on the PC?

 

try disabling it for a while,

 

it may need the WSJT-X folders excluding from scanning, it may be locking the files while scanning

 

i see this on servers at work

 

peter

 

Sat, 05/09/2020 22:29, Amos Sobel 4X4MF wrote:

Bill, Mike, Peter and Neil

 

Today, I did the following:

  1. Installed a CPU thermometer : https://www.techpowerup.com/realtemp/ . It is very elaborate but I still do not know how to use all it’s features
  2. Increase the computational capability of my CPU to the top by: Norton Utilities Premium – Ultra Performance. Maximizes system performance by engaging all CPU cores at full clock speed.
  3. Set all the 4 instances of WSJT-X to 14.074Mhz for maximum FT8 input rate. Input rate was greater than 36 decodes per cycle as measured by JTAlert on each instance.
  4. All 4 JTAlert instances did crash after about 2 hours. Maximum temperature did reach 100 Celsius.
  5. This evening at 19:50 I did restart the  4x JTAlert and put all WSJT-X to 7.074MHz. Decode rate is 33-36 per 15 Sec cycle. CPU frequency is 4240Mhz. Max temp is 89 Celcius. One JTALert instance did crash at 20:10 and I did restart it. 20:13 another instant did crash at 91 Celsius. At 22:20 another 2 instances did crash an 91 C.
  6. Here is how the thermometer did look like. I was using the left column. Can anybody explain the other coulmns?

 

 

 

 


This email has been scanned by iomartcloud.
http://www.iomartcloud.com



joe gay
 

Hi Amos,
Open your event viewer (type 'event viewer' in the lower left search box) & see what errors, etc have been reported - the logs are time stamped so you can see what was reported when you experienced a crash or hang. It will give the info on what is going on. I had a driver issue on a new Win10 PC that generated 1000's of errors per minute & prevented the PC from working well - turned out to be a microsoft reporting error involving the 'Cortana' & skype apps which MS loads with a new PC whether you want them or not. You can do a search on the error code reported to find help from others on fixing it but beware - scammers & virus peddlers pose as friends to 'Help you fix the error' don't click on a offer to download a fix.
Good luck


Amos Sobel 4X4MF
 

Joe

 

Thank you for trying to help.

I am not familiar with the Event Viewer but I did try:

I did force my system to JTAlert crash and it did happen on 10:26 PM. I did then go over all the items in event viewer but could not find anything for that hour.

How do I find the culprit?

 

Amos 4X4MF

 

 

 

From: main@WSJTX.groups.io <main@WSJTX.groups.io> On Behalf Of joe gay via groups.io
Sent: Monday, September 7, 2020 5:41 PM
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] CPU overload by 4 Instance WSJT-X and JTAlert causes JTAlert to crash / Update

 

Hi Amos,
Open your event viewer (type 'event viewer' in the lower left search box) & see what errors, etc have been reported - the logs are time stamped so you can see what was reported when you experienced a crash or hang. It will give the info on what is going on. I had a driver issue on a new Win10 PC that generated 1000's of errors per minute & prevented the PC from working well - turned out to be a microsoft reporting error involving the 'Cortana' & skype apps which MS loads with a new PC whether you want them or not. You can do a search on the error code reported to find help from others on fixing it but beware - scammers & virus peddlers pose as friends to 'Help you fix the error' don't click on a offer to download a fix.
Good luck


Jim Bacher - WB8VSU
 

Amos, unless you are using HRD to control your radio, just run the HRD logbook. The logbook will run stand alone. 

Some of this likely has already been mentioned, but this is the order I would follow to solve an issue. 

When I have an issue with Windows (or Linux). First thing I check is how much unused disk space I have. I want to see the unused disk space to be over 10% on all disks.

Next thing I look at is how much memory is in use by using the Task Manager on Windows. If there is a lot of memory in use, then things start swapping to disk and that hurts performance and can cause issues. My system has 8 gigs and 60% of memory is being used. 

When looking at all processes with the Task Manager, you can sort on columns. So next I would sort on cpu % to see what is consuming cpu cycles. Make sure you are looking at all, not just user. 

On my computer WSJT-X has hit as much as 10% cpu cycles (I only have one instance running). Jtalert shows 0% cpu usage. The system itself is mostly under 25%. My system will step the processor speed down if nothing is going on and I suspect that is why the 10% reading on WSJT-X, as most of the time WSJT-X is under 5%.

Then what I do is shut down the computer and then power it off for a few minutes. Then turn it back on. Once it's back up, but before running any apps, I bring up Task Manager, and switch to its graphical display, (the tab called Performance). Then I watch what it's doing idle for a bit. If nothing seems out of order, I then start the apps one at a time and watch what happens. 


Jim Bacher, WB8VSU 
wb8vsu@... 
https://trc.guru