AnsweredAssumed Answered

Eagleware/Empower 2004.07 - physical memory issue

Question asked by mfaric on Feb 16, 2005
Latest reply on Mar 3, 2005 by ROB_LEFEBVRE

Like • Show 0 Likes0
Comment • 7

Eagleware/Empower crashes and complains that there is not enough physical memory available. The windows system monitor utility, however, indicates that there is more physical memory still unused than the Eagleware/Empower simulation status window or the listing file indicate that it was trying to use. The Windows system performance monitor utility also suggests that Eagleware/Empower did not try to acquire and use the free physical memory that was there - that it aborted itself without trying to acquire and use the memory.
Is there something I am missing in my understanding of how Eagleware/Empower and Windows interact regarding physical memory requirements/availability?
Thank you,
Michael Farich
RF Engineer
Quintech Electronics

Pentium 3.2 GHz, 2GB physical RAM
Windows XP Pro SP2 with all windowsupdate patches.
Administrator privileges on my account.
Eagleware/Empower 2004.07

Attachments

genesys 2004.07 won't use all available physical memory.jpg660.9 KB

No one else has this question

Outcomes

mfaric Feb 21, 2005 3:44 AM

Here's another smaller simulation. The Windows process list shows it using 555,440K bytes of RAM after it aborted with the complaint of not enough memory. The simulation ran in Discontinuity Analysis mode for several minutes before it crashed and the process list showed it using over 600,000K of RAM K. This machine has 1,000,000K of RAM installed, though. I tried the simulation on another machine that has 2,000,000 of RAM installed (2GB) and it did the same thing.
Mike

Attachments

image2.jpg622.1 KB

Actions

mfehr Feb 21, 2005 6:11 AM

:?: can I get this file. I'm interested to test it on my computer.
Matthias Fehr

Actions

abomp Feb 21, 2005 7:44 AM

Some time ago ( maybe a couple months) there was an issue with leaky memory. Maybe it is still an issue.

Actions

mfaric Feb 21, 2005 8:07 AM

The .zip file contains the .wsp and the defaults .tbl and the footprint library .LIB files. The last two files are found in the c:\\Program Files\\Genesys...\\Lib folder and may be needed to view and simulate the contents of the .wsp file.
Mike

[file]s.zip[/file]

Attachments

s.zip45.9 KB

Actions

ROB_LEFEBVRE Feb 28, 2005 9:53 AM

This file has not been set-up optimally. Make the following changes:

1) Grid size can be 5 mils. This is a predominantly lumped circuit, so the metal is a secondary effect, so a coarse grid is no problem.
2) Turn on thinning out.

This problem only requires a few megabytes of memory, not half a gigabyte.

We will look into the crash as a separate issue. There were never memory leaks related to this, but there was an issue that EMPOWER would take more memory than expected. It is possible that that is the cause of the crash, but I'm not sure.

Rob Lefebvre

Actions

mfaric Mar 3, 2005 1:58 AM

Perhaps you are correct about the actual memory requirement being larger than that indicated in the Empower simulation status window. I often observe Windows Task Manager to report a memory amount for the process that is maybe 25% larger than that indicated in the Eagleware simulation status window. The one pictured below is an extreme case percentagewise. Windows reports twice the memory reported by Eagleware.
Mike

Attachments

image1.jpg374.0 KB

Actions

ROB_LEFEBVRE Mar 3, 2005 3:14 AM

Actually, this is not surprising. EMPOWER reports the amount of memory necessary for the EMPOWER simulation, not the entire footprint of GENESYS in memory. This is normally appropriate, because the rest of GENESYS will actually swap out to disk using virtual memory. However, if the total footprint of GENESYS is in the gigabyte+ range, you will run into operating system memory limitations, and virtual memory won't help.

Again, though, in this case, looking at your circuit, I'd be extremely surprised if it really needed 98 megs to run.

Rob Lefebvre

Actions

7 Replies

mfaric Feb 21, 2005 3:44 AM
Mark Correct
Correct Answer
Here's another smaller simulation. The Windows process list shows it using 555,440K bytes of RAM after it aborted with the complaint of not enough memory. The simulation ran in Discontinuity Analysis mode for several minutes before it crashed and the process list showed it using over 600,000K of RAM K. This machine has 1,000,000K of RAM installed, though. I tried the simulation on another machine that has 2,000,000 of RAM installed (2GB) and it did the same thing.
Mike
Attachments
- image2.jpg622.1 KB
Like • Show 0 Likes0
Actions
mfehr Feb 21, 2005 6:11 AM
Mark Correct
Correct Answer
:?: can I get this file. I'm interested to test it on my computer.
Matthias Fehr
Like • Show 0 Likes0
Actions
abomp Feb 21, 2005 7:44 AM
Mark Correct
Correct Answer
Some time ago ( maybe a couple months) there was an issue with leaky memory. Maybe it is still an issue.
Like • Show 0 Likes0
Actions
mfaric Feb 21, 2005 8:07 AM
Mark Correct
Correct Answer
The .zip file contains the .wsp and the defaults .tbl and the footprint library .LIB files. The last two files are found in the c:\\Program Files\\Genesys...\\Lib folder and may be needed to view and simulate the contents of the .wsp file.
Mike

[file]s.zip[/file]
Attachments
- s.zip45.9 KB
Like • Show 0 Likes0
Actions
ROB_LEFEBVRE Feb 28, 2005 9:53 AM
Mark Correct
Correct Answer
This file has not been set-up optimally. Make the following changes:

1) Grid size can be 5 mils. This is a predominantly lumped circuit, so the metal is a secondary effect, so a coarse grid is no problem.
2) Turn on thinning out.

This problem only requires a few megabytes of memory, not half a gigabyte.

We will look into the crash as a separate issue. There were never memory leaks related to this, but there was an issue that EMPOWER would take more memory than expected. It is possible that that is the cause of the crash, but I'm not sure.

Rob Lefebvre
Like • Show 0 Likes0
Actions
mfaric Mar 3, 2005 1:58 AM
Mark Correct
Correct Answer
Perhaps you are correct about the actual memory requirement being larger than that indicated in the Empower simulation status window. I often observe Windows Task Manager to report a memory amount for the process that is maybe 25% larger than that indicated in the Eagleware simulation status window. The one pictured below is an extreme case percentagewise. Windows reports twice the memory reported by Eagleware.
Mike
Attachments
- image1.jpg374.0 KB
Like • Show 0 Likes0
Actions
ROB_LEFEBVRE Mar 3, 2005 3:14 AM
Mark Correct
Correct Answer
Actually, this is not surprising. EMPOWER reports the amount of memory necessary for the EMPOWER simulation, not the entire footprint of GENESYS in memory. This is normally appropriate, because the rest of GENESYS will actually swap out to disk using virtual memory. However, if the total footprint of GENESYS is in the gigabyte+ range, you will run into operating system memory limitations, and virtual memory won't help.

Again, though, in this case, looking at your circuit, I'd be extremely surprised if it really needed 98 megs to run.

Rob Lefebvre
Like • Show 0 Likes0
Actions

Eagleware/Empower 2004.07 - physical memory issue

Attachments

Outcomes

Attachments

Attachments

Attachments

Related Content

Recommended Content