Poor Performance with controller input - Page 2

Toni Wilen · 29 May 2017, 19:20

It is not the first time WinUAE triggers some driver issue because it does input reading quite differently than most other programs..

At least it is very rare issue because no one has ever reported something like this.

If you want to do boring testing:

Use latest beta (see sticky thread in this thread), run with -log -inputlog 7 command line parameters in windowed mode ("winuae.exe -log -inputlog 7"). Do you get log messages every time you for example press pad button or is there huge continuos output of log messages?

Sinphaltimus · 29 May 2017, 19:37

I will give that a go in a minute - and get back to you ASAP.
In the meantime...
Here's the video of AF running the same hdf.

[ Show youtube player ]

Sinphaltimus · 29 May 2017, 19:44

It's probably a false positive but AVG detects a virus in that ZIP -
Here's the details of what it thinks its infected with.

http://www.avgthreatlabs.com/us-en/v...x3aW51YWUuZXhl

EDIT:
Ah, I see this is a known issue - I'll run the 64 bit version just so I don't have to jump through AVG hoops.

http://eab.abime.net/showthread.php?p=1093050

Toni Wilen · 29 May 2017, 19:47

Quote:

Originally Posted by Sinphaltimus

It's probably a false positive but AVG detects a virus in that ZIP -
Here's the details of what it thinks its infected with.

http://www.avgthreatlabs.com/us-en/v...x3aW51YWUuZXhl

Yes, it is. AVG always does that. It is crap. It is the only one that does that after every update. No one else does it (according to Virus Total)

DamienD · 29 May 2017, 19:50

Quote:

Originally Posted by Sinphaltimus

In the meantime...
Here's the video of AF running the same hdf.

[ Show youtube player ]

...in case you don't know Sinphaltimus, Toni never watches videos

Sinphaltimus · 29 May 2017, 20:07

OK, so some new info. And looking at the cmd window confirms.

I only get lag when the controller status changes. in other words, I press UP and the lag happens but stops if I continue to hold UP or release. So the jitter is occurring when input status changes but not if unchanged. So every button press causes the jitter but not beyond the initial press, so press and hold causes a stutter and that's it as long as I'm not moving the stick at all. That's probably why it's worse when using the mouse because every mouse movement is a different value with each move unless it is sitting still.

There's the log as an attachment.

EDIT "Do you get log messages every time you for example press pad button or is there huge continuous output of log messages?" - I get a small stream of msgs with each movement change and/or button press.

Sinphaltimus · 29 May 2017, 20:09

Quote:

Originally Posted by DamienD

...in case you don't know Sinphaltimus, Toni never watches videos

LOL! Good to know. with my 1990s internet speeds it sure saves me a lot of headache.

Toni Wilen · 29 May 2017, 20:37

Videos practically never show anything useful when compared to log files. (It only shows "something is wrong"

)

So there is not any kind of event flood either that would explain it. This is getting far too mysterious..

If you move x360 pad analog stick, do the movement messages show instantly or is there small delay? (move the stick, then quickly release it, does log still show messages after releasing it?)

If you also have windows task manager open, does it show sudden heavy CPU usage at the same time?

Sinphaltimus · 29 May 2017, 21:12

OK, interesting, yes, there are CPU spikes.

I tested this at the title screen and the options screen because pressing the fire button brought me there.

The command prompt window shows instant changes.

Mouse produces a constant stream which we determined but I'm using that stream to clear the screen essentially. What I'm seeing is a burst of streams in the log output.

When I quickly press and release the fire button I see 3 bursts:

1 15 00D6828D8BB47BC886008001000000 H=0000000000020049
9/1 ON
DEV 0 AXIS 0: 0 32767 (0 - 65535) 33494 727
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31668 -1099
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34504 1737
DEV 0 AXIS 0: 0 32767 (0 - 65535) 33494 727
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31668 -1099
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34504 1737
1 15 00D6828D8B467BC886008000000000 H=0000000000020049
9/1 OFF
DEV 0 AXIS 0: 0 32767 (0 - 65535) 33494 727
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31558 -1209
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34504 1737
1 15 0046838D8BB47B5586008000000000 H=0000000000020049
DEV 0 AXIS 0: 0 32767 (0 - 65535) 33606 839
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31668 -1099
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34389 1622

then I move the mouse a little for readability on the next test..

0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 1 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 0 1 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 0 1 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 -1 0 00000000 M=4 F=2
0001003b 0000 0000 0000 00000000 0 1 00000000 M=4 F=2

Then I move the stick as fast as possible (note, this is super sensitive, i started seeing bursts as soon as I touched the stick getting ready to move it but I won't paste that part below). so hitting the stick really fast in one direction as best i could produced 4 bursts.

1 15 00F7818D8BB47BC886008000000000 H=0000000000020049
DEV 0 AXIS 0: 0 32767 (0 - 65535) 33271 504
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31668 -1099
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34504 1737
1 15 00C97F8D8BB47B5586008000000000 H=0000000000020049
DEV 0 AXIS 0: 0 32767 (0 - 65535) 32713 -54
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31668 -1099
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34389 1622
1 15 004C7C8D8BB47BC886008000000000 H=0000000000020049
DEV 0 AXIS 0: 0 32767 (0 - 65535) 31820 -947
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35725 2958
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31668 -1099
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34504 1737
1 15 00FD7AFE8B467BC886008000000000 H=0000000000020049
DEV 0 AXIS 0: 0 32767 (0 - 65535) 31485 -1282
DEV 0 AXIS 1: 0 32767 (0 - 65535) 35838 3071
DEV 0 AXIS 2: 0 32767 (0 - 65535) 32768 1
DEV 0 AXIS 3: 0 32767 (0 - 65535) 31558 -1209
DEV 0 AXIS 4: 0 32767 (0 - 65535) 34504 1737

NOTE - The bursts are instantaneous. no changes in output screen once controller is released (no lag in the output).

Sinphaltimus · 30 May 2017, 00:08

This guy might be experiencing the same thing -

http://eab.abime.net/showpost.php?p=1161304&postcount=1

Sinphaltimus · 30 May 2017, 00:19

Quote:

Originally Posted by DamienD

I don't use Amiga Forever but isn't this just a glorified front-end that actually uses WinUAE behind the scenes?!?

...if so, it's an older version than v3.4.0 so could try that version directly.

Interesting you posted this. I was surfing elsewhere at the time but....

I do believe you're correct about AF and UAE or at least it was that way previously. I read the other post about the guy having issues with 3.4 in OctaMed but not in 3.3. I just downloaded 3.3 from Sourceforge and it's even worse for me in the sense that it's jittery all the time, just barely gets worse when gamepad or mouse are invoked.

Sinphaltimus · 30 May 2017, 00:28

OH! News. When I launch WinUAE via Amiga Forever, it launches v 3.2.2.

So I guess that's what AF is using. so I launched my same config using that and yep, works flawlessly.

So my problem start in v3.3 with improvements in 3.4 but no issues in 3.2.2

DamienD · 30 May 2017, 00:31

Quote:

Originally Posted by Sinphaltimus

OH! News. When I launch WinUAE via Amiga Forever, it launches v 3.2.2.

That is good news

...but I have no issues like you've mentioned at all with 3.0 / 3.1 / 3.2 / 3.2.2 / 3.3 / 3.4

I also use Xbox 360 controllers etc.

Mind you, I'm using Windows 7 and refuse to go to Windows 10 at the moment for various reasons...

Anyway, would be good to get to the bottom of this and I'm sure Toni will; he always does.

Toni Wilen · 30 May 2017, 15:18

Obviously it is not 360 pad nor Windows 10 problem. It is also not above linked problem (if it is CPU idle related because it only affects fastest possible modes)

So if 3.3.0 works, try 3.4.0 betas, one by one until you find where the issue was introduced. (Links in 3.4.0 beta thread)

Sinphaltimus · 30 May 2017, 15:37

3.2.2 works. 3.3 is poorer performance issues than 3.4. 3.4 is the release I used when I originally reported the issue. I'll stick with 3.2.2. Something terrible happened between 3.2.2 and 3.3 in my little niche if the world.

Weekend over, working all week. I don't know how much time I will have during the week to trouble shoot this further but I'm willing to help where and when I can.

mark_k · 30 May 2017, 18:23

Check the 3.3.0 beta series thread. If you can find the specific beta version in which the problem first appeared that could help. You don't need to test every single beta. E.g. start with beta 8, if the problem isn't present try beta 12 next. If it is present in beta 8, try beta 4 to see whether it's there. Etc.

It probably isn't the root cause of your problem, but have you tried disabling all Windows and BIOS power-saving features?

Toni Wilen · 30 May 2017, 20:02

One detail I found from your logs:

Quote:

LCD enabled

Do you have Logitech keyboard with LCD? Or remains of old installation? Or perhaps it is "emulated" by some other manufacturer. This could cause side-effects if it isn't properly working.

3.3.0 switched to newer Logitech LCD API.

Sinphaltimus · 30 May 2017, 20:20

I don't have an lcd anything besides display. Never have. I'll search the logs for more clues on that and see what my system turns up. I've only ever had generic Logitech minimalistic keyboards. I can't think if anything I've ever connected to one of my pcs with an lcd display that wasn't an lcd monitor.

Sinphaltimus · 31 May 2017, 00:46

Quote:

Originally Posted by mark_k

Check the 3.3.0 beta series thread. If you can find the specific beta version in which the problem first appeared that could help. You don't need to test every single beta. E.g. start with beta 8, if the problem isn't present try beta 12 next. If it is present in beta 8, try beta 4 to see whether it's there. Etc.

It probably isn't the root cause of your problem, but have you tried disabling all Windows and BIOS power-saving features?

Unfortunately, They are no longer available anyplace I can find (the old beta versions) - Links to the downloads from eab topic are no longer working.

Yes, disabling power saving features is a must on all my builds. Bios and Windows.

Sinphaltimus · 31 May 2017, 01:15

@Toni - Taking a closer look at these issues I ran process lasso to look at the CPU Utilization of the PC. Strange thing I see is over-all utilization of less active cpu cores rise up slightly when moving mouse KB and controller while UAW is running but the total percentage of the WinUAE process itself drops from about 12% to 8% when this occurs.

I do have some very old logitech drivers installed for my wheel and NASCAR racing season 2002 and new racing games (Yes my old wheel works with the likes of Microsoft's Dirt Rally). I am able to close and quit all processes related to it and it results in no performance changes in Winuae. The only LED driver I have found was related to a Samsung TV once plugged in here. I have removed that wit no change in performance either.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
D-Pad input (Xbox360 controller) partially broken in 2.6.1	Dr.Venom	support.WinUAE	6	25 June 2013 12:15
Why does controller only work on port 2? How to change input targets?	jackhulk	New to Emulation or Amiga scene	12	18 December 2012 16:37
Workench 3.9 screen poor	dezmond	support.Apps	2	05 November 2009 17:26
Poor WinUAE performance and flashing screen	dekarAF	support.WinUAE	68	25 October 2009 10:09
Help this poor idiot	Grey Area	New to Emulation or Amiga scene	7	07 April 2005 15:39

29 May 2017, 19:20	#21
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,517	It is not the first time WinUAE triggers some driver issue because it does input reading quite differently than most other programs.. At least it is very rare issue because no one has ever reported something like this. If you want to do boring testing: Use latest beta (see sticky thread in this thread), run with -log -inputlog 7 command line parameters in windowed mode ("winuae.exe -log -inputlog 7"). Do you get log messages every time you for example press pad button or is there huge continuos output of log messages?

29 May 2017, 19:37	#22
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	I will give that a go in a minute - and get back to you ASAP. In the meantime... Here's the video of AF running the same hdf. [ Show youtube player ]

29 May 2017, 19:44	#23
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	It's probably a false positive but AVG detects a virus in that ZIP - Here's the details of what it thinks its infected with. http://www.avgthreatlabs.com/us-en/v...x3aW51YWUuZXhl EDIT: Ah, I see this is a known issue - I'll run the 64 bit version just so I don't have to jump through AVG hoops. http://eab.abime.net/showthread.php?p=1093050

29 May 2017, 20:37	#28
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,517	Videos practically never show anything useful when compared to log files. (It only shows "something is wrong" ) So there is not any kind of event flood either that would explain it. This is getting far too mysterious.. If you move x360 pad analog stick, do the movement messages show instantly or is there small delay? (move the stick, then quickly release it, does log still show messages after releasing it?) If you also have windows task manager open, does it show sudden heavy CPU usage at the same time?

30 May 2017, 00:08	#30
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	This guy might be experiencing the same thing - http://eab.abime.net/showpost.php?p=1161304&postcount=1

30 May 2017, 00:28	#32
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	OH! News. When I launch WinUAE via Amiga Forever, it launches v 3.2.2. So I guess that's what AF is using. so I launched my same config using that and yep, works flawlessly. So my problem start in v3.3 with improvements in 3.4 but no issues in 3.2.2

30 May 2017, 15:18	#34
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,517	Obviously it is not 360 pad nor Windows 10 problem. It is also not above linked problem (if it is CPU idle related because it only affects fastest possible modes) So if 3.3.0 works, try 3.4.0 betas, one by one until you find where the issue was introduced. (Links in 3.4.0 beta thread)

30 May 2017, 15:37	#35
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	3.2.2 works. 3.3 is poorer performance issues than 3.4. 3.4 is the release I used when I originally reported the issue. I'll stick with 3.2.2. Something terrible happened between 3.2.2 and 3.3 in my little niche if the world. Weekend over, working all week. I don't know how much time I will have during the week to trouble shoot this further but I'm willing to help where and when I can.

30 May 2017, 18:23	#36
mark_k Registered User Join Date: Aug 2004 Location: Posts: 3,343	Check the 3.3.0 beta series thread. If you can find the specific beta version in which the problem first appeared that could help. You don't need to test every single beta. E.g. start with beta 8, if the problem isn't present try beta 12 next. If it is present in beta 8, try beta 4 to see whether it's there. Etc. It probably isn't the root cause of your problem, but have you tried disabling all Windows and BIOS power-saving features?

30 May 2017, 20:20	#38
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	I don't have an lcd anything besides display. Never have. I'll search the logs for more clues on that and see what my system turns up. I've only ever had generic Logitech minimalistic keyboards. I can't think if anything I've ever connected to one of my pcs with an lcd display that wasn't an lcd monitor.

31 May 2017, 01:15	#40
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	@Toni - Taking a closer look at these issues I ran process lasso to look at the CPU Utilization of the PC. Strange thing I see is over-all utilization of less active cpu cores rise up slightly when moving mouse KB and controller while UAW is running but the total percentage of the WinUAE process itself drops from about 12% to 8% when this occurs. I do have some very old logitech drivers installed for my wheel and NASCAR racing season 2002 and new racing games (Yes my old wheel works with the likes of Microsoft's Dirt Rally). I am able to close and quit all processes related to it and it results in no performance changes in Winuae. The only LED driver I have found was related to a Samsung TV once plugged in here. I have removed that wit no change in performance either.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)