DisplayFusion RSS: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Mon, 20 Nov 2017 18:54:56 GMT

No worries, glad to hear it. Thanks for the kind words!

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sat, 18 Nov 2017 07:45:05 GMT

After days of testing the beta version, it has been stable at every time. I installed the final last release a few days ago and is stable too.

Thank you Keith for your support.

I'm happy to see so good support from DF team =D

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Mon, 13 Nov 2017 18:50:23 GMT

Excellent, glad to hear it. Thanks for following up!

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sun, 12 Nov 2017 00:10:49 GMT

Also can confirm I haven't seen the problem since v9.1 Beta 1. Great work!

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sat, 11 Nov 2017 14:29:34 GMT

Testing the Beta since it's release, my memory usage is now at a stable 50-100 MB at all times. Seems to be working great, thanks!

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Thu, 09 Nov 2017 09:11:45 GMT

Quote:

@Maliquem: It looks like you're still running one of the older test builds. Could you try 9.1 Beta 1 that we released on Friday?

After 72 hours on testing, this build is, by now, stable in RAM usage for me (50-100 Mb).

I'll keep on testing it.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Mon, 06 Nov 2017 16:33:00 GMT

@Maliquem: It looks like you're still running one of the older test builds. Could you try 9.1 Beta 1 that we released on Friday?

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sat, 04 Nov 2017 15:22:12 GMT

The new build was working properly yesterday. But today the same problem show up again... And freeze my explorer.exe too.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sat, 04 Nov 2017 11:00:10 GMT

@Knurr: We updated our code signing certificate with 9.1 B1, so that might have thrown it off. If you want to be safe, you can run the installer through virustotal.com to verify that it's clean.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sat, 04 Nov 2017 09:42:28 GMT

Quote:

It actually had to do with some polling we added for the new Process Created trigger event. We've just posted 9.1 Beta 1, so if everyone can update to that and let us know how it works out, that would be great!

Installing now and checking.

A few days ago this version (https://www.binaryfortress.com/Files/DisplayFusion/DisplayFusionSetup-9.0-20171030-1145.exe) worked for me for several days until DF began to consuming more memory (aprox. 600-800 Mb). The RAM consume increased very slowly anyway.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Sat, 04 Nov 2017 00:02:38 GMT

Quote:

I've tried installing it but my antivirus thought it's a virus and put it in quarantine....didn't have the problem with the 9.0 1545 fix posted yesterday. Is this normal for a beta version? I've never installed one for DF before...

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Fri, 03 Nov 2017 19:10:50 GMT

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Fri, 03 Nov 2017 17:51:54 GMT

So what is the cause of this, and why do some get it and others don't?

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Fri, 03 Nov 2017 17:37:35 GMT

Excellent, glad to hear it! We'll have a more proper fix coming in the next public beta version

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Fri, 03 Nov 2017 14:44:57 GMT

Quote:

For the folks that are still running into issues with the previous test build, could you try this new one? Please note that the Monitor Fading feature is disabled in this build.
http://www.binaryfortress.com/Files/DisplayFusion/DisplayFusionSetup-9.0-20171102-1545.exe

This build appears stable to me after 8 hours running.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Fri, 03 Nov 2017 01:52:15 GMT

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 22:26:08 GMT

Quote:

Ok guys, we have a new test build to try that may fix up this memory leak. Could you download and install it over top of your existing install, and let us know how it works out?
https://www.binaryfortress.com/Files/DisplayFusion/DisplayFusionSetup-9.0-20171030-1145.exe

It seem to have worked Still no crashes

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 18:52:57 GMT

@Marcin: Yep, you're definitely right. I've merged that topic with the big one

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 18:47:54 GMT

@ral1sh: We had found those leaking PNG images as well, and they should be fixed up in the test build (https://www.binaryfortress.com/Files/DisplayFusion/DisplayFusionSetup-9.0-20171030-1145.exe). The test build will still create a massive number of threads under certain circumstances so we're still trying to track that down. I'll pass your info on to our devs and if they need to grab some info from you, they'll definitely give you a shout.

Thanks!

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 08:10:38 GMT

I have the same issue, but I think is related to memory issue. It stops working after memory peak, and all icons from right side of taskbar are gone.

It looks this memory issue can be reltaed to taskbar icons.
I'm using the latest build from memory issue thread.

This post was merged on Nov 1, 2017.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 07:35:38 GMT

Quote:

Hello everybody.

Until now, 36 hours after installing this build, the memory usage dropped dramatically, at least in my computer. No memory leak had happened. It looks like this build works for me.

Thanks about the help.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 04:53:09 GMT

Right, this time it happened again and I caught it happening rather than when memory usage had peaked. A huge number of threads are being created which are almost all waiting on unique event handles to be signalled. My earlier dump had a huge 4886 threads and 32,586 event handles. I've attached a dump of unique stacks which might provide some insight. If any DF devs want me to provide other info let me know. I could potentially turn on handle tracing or break on thread creation to try and capture the stack trace corresponding to all these threads being created.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 04:18:16 GMT

I'm seeing the same issue on DisplayFusion v9.0. I've seen it a few times, but just now managed to capture a VMMap and process dump (a whopping ~11GB!). A few observations which may be of assistance in tracking the problem down (very quick analysis):

There's a huge number of allocation on the managed and unmanaged heap (see attached).

Looking at the managed heap statistics the two major offenders are (I'm assuming the first one has a "malformed" type name due to some sort of code obfuscation):

Code
Copy
Select All


00007ffd800ed7f8  7616032    182784768 .+
00007ffddb28bd38   209966    952711704 System.Byte[]
Total 9884624 objects

For the "malformed" type all of the allocations are 24 bytes each:

Code
Copy
Select All


0000000007ac6128 00007ffd800ed7f8       24     
0000000007ac6140 00007ffd800ed7f8       24     
0000000007ac6158 00007ffd800ed7f8       24     
0000000007ac6170 00007ffd800ed7f8       24     
0000000007ac6188 00007ffd800ed7f8       24

Dumping one of the objects gives:

Code
Copy
Select All

Name:        .+
MethodTable: 00007ffd800ed7f8
EEClass:     00007ffd801382b8
Size:        24(0x18) bytes
File:        C:\Program Files (x86)\Binary Fortress\DisplayFusion\DisplayFusion.exe
Fields:
              MT    Field   Offset                 Type VT     Attr            Value Name
00007ffd800c1038  400216d        8                  .  0 instance 0000000000000000 
00007ffd800a9938  400216e     1920 ...legates.GetString  0   static 0000000007581b28

The Byte array is more interesting. Dumping one of them:

Code
Copy
Select All

Name:        System.Byte[]
MethodTable: 00007ffddb28bd38
EEClass:     00007ffddaca26c8
Size:        41088(0xa080) bytes
Array:       Rank 1, Number of elements 41064, Type Byte (Print Array)
Content:     .PNG........IHDR..............>a.....sRGB.........gAMA......a.....tEXtSoftware.Adobe ImageReadyq.e<....IDATx^...X...=.sNz1E.{...
Fields:
None

Let's write that out to disk:

Code
Copy
Select All

.writemem C:\Users\SDL\Desktop\test.png ba438b8 La068
Writing a068 bytes.....................

Sure enough, it's a PNG, and yes I can view it:

Code
Copy
Select All

test.png: PNG image data, 128 x 128, 8-bit/color RGBA, non-interlaced

I went through a whole bunch of them, and nearly all of them seemed to be PNG icons. There's also a lot of duplicates, but they're nonetheless unique entries in the array. Some sort of icon/image cache maybe?

Hope this helps narrow it down ...

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Wed, 01 Nov 2017 00:02:17 GMT

I have the same problem.
I've seen the process take as much as 19GB ram. After that DF is unresponsive and the process has to be killed.
I had the problem with 9.0 beta 5 and with the 9.0 stable release. It was not happening with 9.0 beta 4 if i recall correctly.
I believe the problem occurs faster if i watch fullscreen streams in firefox, but i have no proof.
OS: Windows 10 Prof. x64 v1709
Will hand in a log with enabled debugging the next time the problem occurs.

RE: DisplaFusion V9.0 Memory Bug

Binary Fortress Software — Tue, 31 Oct 2017 21:41:51 GMT

Here

This post was merged on Nov 1, 2017.

DisplayFusion RSS: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

CodeCopySelect All

CodeCopySelect All

CodeCopySelect All

CodeCopySelect All

CodeCopySelect All

CodeCopySelect All

RE: DisplaFusion V9.0 Memory Bug

RE: DisplaFusion V9.0 Memory Bug

Code
Copy
Select All

Code
Copy
Select All

Code
Copy
Select All

Code
Copy
Select All

Code
Copy
Select All

Code
Copy
Select All