VRAM Object List blank in Windows 11
VRAM Object List blank in Windows 11
Hey all,
I'm trying to use no$psx to do some PS1 debugging, but the object list in the vram viewer goes blank a few seconds after starting a game. Unfortunately, neither duckstation nor pcsx-r have this functionality. It provides a lot of useful information when working, and I'm kind of up a creek without it.
Is there a solution that doesn't involve downgrading to Win 10 or running a VM? I've tried playing with the application's scaling and compatibility settings via its Windows properties to no success.
At boot time: A few seconds later:
I'm trying to use no$psx to do some PS1 debugging, but the object list in the vram viewer goes blank a few seconds after starting a game. Unfortunately, neither duckstation nor pcsx-r have this functionality. It provides a lot of useful information when working, and I'm kind of up a creek without it.
Is there a solution that doesn't involve downgrading to Win 10 or running a VM? I've tried playing with the application's scaling and compatibility settings via its Windows properties to no success.
At boot time: A few seconds later:
You do not have the required permissions to view the files attached to this post.
I have never heard about that problem.
Does it occur when the scrollbar gets displayed? The scrollbar does also look kinda distorted, unless that's an intended "minimalistic" appearance. The dots and vertical lines look as if it's only displaying the leftmost pixel of the text strings, but I have no idea why. Does win11 have new features for reconfiguring the texg width?
Does it happen with all psx games? Does the tree view in no$psx filesystem viewer window have similar problems? If you have any NDS games, does the 3d/vram viewer in no$gba have that problem, too?
Does it occur when the scrollbar gets displayed? The scrollbar does also look kinda distorted, unless that's an intended "minimalistic" appearance. The dots and vertical lines look as if it's only displaying the leftmost pixel of the text strings, but I have no idea why. Does win11 have new features for reconfiguring the texg width?
Does it happen with all psx games? Does the tree view in no$psx filesystem viewer window have similar problems? If you have any NDS games, does the 3d/vram viewer in no$gba have that problem, too?
It does seem to coincide with the scroll bar appearing, and yes, the tree view in the filesystem viewer has the same issue. I don't have many other PS1 ISOs on-hand to test with but I'm working with another developer who's experienced the same issue, and there's a post on RHDN's forum talking about it as well: https://www.romhacking.net/forum/index. ... ic=39098.0 So I'd imagine it's an issue with multiple games.
I had some... really interesting experiences trying to resize the debugger window today. It kept trying to snap way beyond the left-hand side of my main display. It's really hard to describe but it'd grow vertically as it grew horizontally, despite my only trying to stretch it wider, not taller.
I tried no$gba with Dementium - The Ward, and the 3D/RE pane seemed to work fine.
For what it's worth, no$gba 3.05 and no$psx 2.2.
I had some... really interesting experiences trying to resize the debugger window today. It kept trying to snap way beyond the left-hand side of my main display. It's really hard to describe but it'd grow vertically as it grew horizontally, despite my only trying to stretch it wider, not taller.
I tried no$gba with Dementium - The Ward, and the 3D/RE pane seemed to work fine.
For what it's worth, no$gba 3.05 and no$psx 2.2.
Hmmm, I don't why no$gba works better than no$psx, I thought they were both using the same GUI code.
Next no$psx update will use a "list" instead of "tree" in the vram viewer, maybe that fixes the problem (but the filesystem viewer would probably still have the problem).
Or maybe it's happening only if the window is on the 1st or 2nd monitor?
I do know pretty much nothing about multiple monitors, I have once seem somebody dragging a window from one monitor to another, but that's about all I know about it. And no$psx doesn't specifically contain any "support" for such things - windows does somehow automatically apply the working (or not so well working) dual monitor support to no$psx.
The interesting experience might be because you had moved the mouse to a negative screen coordinate (which isn't possible when having only one monitor).
Anyways, I have a bunch questions about that stuff...
Next no$psx update will use a "list" instead of "tree" in the vram viewer, maybe that fixes the problem (but the filesystem viewer would probably still have the problem).
Could that be related? Maybe the issue is dual screen related, not windows 11 related?
Or maybe it's happening only if the window is on the 1st or 2nd monitor?
I do know pretty much nothing about multiple monitors, I have once seem somebody dragging a window from one monitor to another, but that's about all I know about it. And no$psx doesn't specifically contain any "support" for such things - windows does somehow automatically apply the working (or not so well working) dual monitor support to no$psx.
The interesting experience might be because you had moved the mouse to a negative screen coordinate (which isn't possible when having only one monitor).
Anyways, I have a bunch questions about that stuff...
- What is that feature called officially, Dual-Screnn, Multi-Monitor, or so?
- Is there some limit on the number of monitors, only two, or do some people have three or monitors?
- Which windows version did invent support for multiple monitors?
- Are there some docs with advice how to support multiple monitors?
- GetSystemMetrics can be used to obtain the size of the SCREEN, how would that work with multiple screens?
- No$psx supports "fullscreen" mode (via right mouse button), does that work with multiple screens?
- Is it possible to encounter screens with different resolutions, like simultaneously using 640x480 and 1280x1024?
- Each window does have a screen position with X,Y coordinates, how does that work with multiple monitors?
- Win9x has "EnumMonitors" as part of "Printing and Print Spooler" functions. But that seems to be unrelated? (the Printer API describes it as "A monitor is a Windows DLL that passes the raw device commands over the network, through a parallel port, or through a serial port to the device.")
Last edited by nocash on January 3rd, 2025, 1:25 pm, edited 2 times in total.
I just tested it on my side, no issues in Windows 11.
Thanks! Good to know that it isn't completely incompatible with win11.
Maybe it was a bug in win11, and microsoft has fixed in newer updates, or maybe it happens only for people with more than one monitor.
Please let me know if or when the bug does still occur for somebody with no$psx, or if it should also appear in next no$gba update. I am not exactly sure what's causing it. Maybe things like...
Maybe it was a bug in win11, and microsoft has fixed in newer updates, or maybe it happens only for people with more than one monitor.
Please let me know if or when the bug does still occur for somebody with no$psx, or if it should also appear in next no$gba update. I am not exactly sure what's causing it. Maybe things like...
- Only happens when the tree contains enough items to display the scrollbar.
- Maybe depending on win11 and/or having more than one monitor.
- Maybe depending on resizing the window, and on how the resizing is applied (that is a bit different in different trees):
- No$psx filesys viewer has the preview underneath of the tree
- No$gba filesys viewer currently doesn't have that preview (but will have it in next update)
- No$psx vram viewer does adjust the tree width when resizing the window
- No$psx vram viewer always keeps the tree width unchanged
-
- What is PSXDEV?
- Posts: 2
- Joined: Jan 05, 2025
Unfortunately I'm seeing this issue as well on my machine. I am using Windows 11 with a 2 monitor setup and the tree does include a scrollbar. Example attached, please let me know if there's anything else I can provide that would help!
- MasterLink
- Active PSXDEV User
- Posts: 46
- Joined: Jul 20, 2024
- I am a: Electronics, graphics, music.
- PlayStation Model: DTL-H1001H
- Location: USA
Thought I'd chime in on my observations and with some answers as I do know Windows' multi-monitor support rather well.
Windows 98 was the first to support multiple monitors, which was 9 at the time. These days on Windows 11 this was reduced to 6 (I have no idea why it went down not up), however "how" multiple monitors are dealt with depends squarely on the GPU configuration (only speaking for today with Windows 10 and 11, but primarily Windows 11 since that's the subject here). Also, nocash, to answer your question, yes multiple monitors can have different resolutions at the same time (Windows just uses a virtual larger framebuffer to encompass them all with black voids in it where a window simply can't be drawn), but worse, both monitors can have different vertical sync as well. (I've seen this cause so much microstuttering in games or even simply scrolling on a web page.) For this instance though that virtual framebuffer may be causing the snapping behavior they are talking about? I can't replicate it right now in lieu of a second display at the moment.
Now for the complex part: How Windows 11 handles multiple monitors depends so much on your GPU configuration and how these monitors are attached. If you have multiple monitors but they are plugged into different GPU's, then the emulator is likely running on one GPU, but copying the applications framebuffer to another GPU where the second monitor is attached. In general this doesn't actually cause issues. It's highly rare, however this list dynamically refreshes and I wonder if the DXGI swap chain is failing to transfer the updates (hence a blank spot).
Can you confirm both monitors are plugged into the same *physical* GPU and not say, primary monitor on GPU, second monitor attached to onboard video? Also, what GPU make and model are you both running that are running into issues? (Driver versions may help too, in case they are severely out of date, or are too up to date with a newer driver that maybe has known bugs?)
Later today I'll grab a second monitor and just try with two displays, but I did coax NO$PSX to run on both of my GPU's in my computer (Radeon in the Ryzen, and a discrete 1080 Ti) off the same screen, and neither instance corrupted itself. But without a second monitor I can't introduce that variable quite yet.
Windows 98 was the first to support multiple monitors, which was 9 at the time. These days on Windows 11 this was reduced to 6 (I have no idea why it went down not up), however "how" multiple monitors are dealt with depends squarely on the GPU configuration (only speaking for today with Windows 10 and 11, but primarily Windows 11 since that's the subject here). Also, nocash, to answer your question, yes multiple monitors can have different resolutions at the same time (Windows just uses a virtual larger framebuffer to encompass them all with black voids in it where a window simply can't be drawn), but worse, both monitors can have different vertical sync as well. (I've seen this cause so much microstuttering in games or even simply scrolling on a web page.) For this instance though that virtual framebuffer may be causing the snapping behavior they are talking about? I can't replicate it right now in lieu of a second display at the moment.
Now for the complex part: How Windows 11 handles multiple monitors depends so much on your GPU configuration and how these monitors are attached. If you have multiple monitors but they are plugged into different GPU's, then the emulator is likely running on one GPU, but copying the applications framebuffer to another GPU where the second monitor is attached. In general this doesn't actually cause issues. It's highly rare, however this list dynamically refreshes and I wonder if the DXGI swap chain is failing to transfer the updates (hence a blank spot).
Can you confirm both monitors are plugged into the same *physical* GPU and not say, primary monitor on GPU, second monitor attached to onboard video? Also, what GPU make and model are you both running that are running into issues? (Driver versions may help too, in case they are severely out of date, or are too up to date with a newer driver that maybe has known bugs?)
Later today I'll grab a second monitor and just try with two displays, but I did coax NO$PSX to run on both of my GPU's in my computer (Radeon in the Ryzen, and a discrete 1080 Ti) off the same screen, and neither instance corrupted itself. But without a second monitor I can't introduce that variable quite yet.
Retail Consoles: SCPH-1001 (Faulty CD Decoder Currently), SCPH-5501
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
- MasterLink
- Active PSXDEV User
- Posts: 46
- Joined: Jul 20, 2024
- I am a: Electronics, graphics, music.
- PlayStation Model: DTL-H1001H
- Location: USA
Wait I just had a thought, since the second persons screenshot is in a higher DPI than 100% (looks like 200%? or you screenshot is for some reason 2x enlarged)...
Do you both who have multiple monitors, happen to have monitors with mixed DPI's? As in, one of the monitors is NOT on 100%? If so, set both to 100%, reboot (seriously, purge the RAM of this mixed DPI nonsense) and try NO$PSX again, if the issue goes away, it was the DPI mixing causing the problem. The user with a high DPI, and the OP with the window snapping, it's starting to feel like mixed DPI issues are happening here perhaps.
(I know the OP says they adjusted the applications DPI settings, but to rule this out, perhaps set the monitor instead of per-app. Windows 10 and 11's DPI handling is a travesty in my opinion and it only got worse with Windows 11, with simple applications that aren't UWP based.)
Do you both who have multiple monitors, happen to have monitors with mixed DPI's? As in, one of the monitors is NOT on 100%? If so, set both to 100%, reboot (seriously, purge the RAM of this mixed DPI nonsense) and try NO$PSX again, if the issue goes away, it was the DPI mixing causing the problem. The user with a high DPI, and the OP with the window snapping, it's starting to feel like mixed DPI issues are happening here perhaps.
(I know the OP says they adjusted the applications DPI settings, but to rule this out, perhaps set the monitor instead of per-app. Windows 10 and 11's DPI handling is a travesty in my opinion and it only got worse with Windows 11, with simple applications that aren't UWP based.)
Retail Consoles: SCPH-1001 (Faulty CD Decoder Currently), SCPH-5501
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
-
- What is PSXDEV?
- Posts: 2
- Joined: Jan 05, 2025
It definitely seems related to DPI settings. I have both monitors at 4k with 200% DPI, and I see the issue regardless of the application DPI settings I use. However, if I set both monitors to 100%, I can see the text again. So there's something to do with 200% that's causing an issue.
FYI: I'm using a RTX 4090 with the latest driver, both monitors attached to the same GPU.
FYI: I'm using a RTX 4090 with the latest driver, both monitors attached to the same GPU.
- MasterLink
- Active PSXDEV User
- Posts: 46
- Joined: Jul 20, 2024
- I am a: Electronics, graphics, music.
- PlayStation Model: DTL-H1001H
- Location: USA
Good to know, so Windows' DPI settings are the root cause. I'm glad I caught that in the screenshot (and comments) or I'd have missed that DPI issues were at play.
I'm gonna have to point the finger at Windows 11 for this one, if it were doing this at 100% I'd think GPU and driver, but since noticing you both had DPI settings set, I'm starting to lean towards this being a downright messy Windows 11 scaling issue. For what it's worth, at 125% on my display, I couldn't replicate this, and my Windows build is 26100.2605. (But just in case, my driver for my GTX 1080 Ti, which I think shares the same driver as the RTX series, is 566.03 from October.)
I'm gonna have to point the finger at Windows 11 for this one, if it were doing this at 100% I'd think GPU and driver, but since noticing you both had DPI settings set, I'm starting to lean towards this being a downright messy Windows 11 scaling issue. For what it's worth, at 125% on my display, I couldn't replicate this, and my Windows build is 26100.2605. (But just in case, my driver for my GTX 1080 Ti, which I think shares the same driver as the RTX series, is 566.03 from October.)
Retail Consoles: SCPH-1001 (Faulty CD Decoder Currently), SCPH-5501
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
The feature is called "Multiple Display Monitors". Some docs can be found here...
https://learn.microsoft.com/en-us/windo ... -functions
above is only for "functions" (some web browsers are able to display a table of contents on the left of the browser window, with more multiple display monitor related chapters, not only the functions).
From what I can see, the feature does rely on a "primary screen", and additional screens that can be arranged at left or right or above or perhaps below of the "primary screen". As shown here:
https://learn.microsoft.com/en-us/windo ... ual-screen
Of those, the monitors left/above of the primary screen can have negative coordinates, which may cause problems for software that expects unsigned coordinates (although, even with a single monitor, it's possible to move a window to parts offscreen locations on left/top sides (the mouse arrow can't be move to negative coordinates on a single monitor though)).
For the fullscreen feature, the standard system metrics function reports only the size of the "primary screen". So I assume that no$xxx in fullscreen mode will properly display the game screen on that monitor at correct size.
DPI... I wasn't aware of such features. I've found something similar in win98, under Display Properties, Settings, Advanced, General, Font Size (Normal or Large, with 96 or 120 dpi).
That seems to be a driver specific setting, and it may be absent or different for other display drivers. The effect seems to be that some things like Dialog box units are getting zoomed (affecting captions, meno bars, tab controls, scrollbars - but not affecting any pixel sizes of owner drawn bitmaps). I am not sure if that is the same DPI feature as on windows 11 (?)
Some of those settings like menu bar and tab control height are important for positioning controls and child windows within another window. And no$xxx does assume that those settings are same for all windows. But even if the menu bar height varies by +/-10 pixels, that should only cause minor glitches.
---
Some more questions:
Is there still any need to fix the issue, or is it accepable simply not to use different DPI settings?
Does the problem occur on the "primary display"?
Does the problem occur on "non-primary displays"?
Or both on primary and non-primary displays?
And are the non-primary ones at negative coordinates (left/above of the primary)?
Does it make a difference to close and re-open the window after dragging it to another monitor?
Is it even possible to drag no$xxx windows to different monitors? Theoretically, my code should prevent to move windows to completely offscreen locations (whereas "offscreen" would mean outside of the primary display area) (or at least so when storing "offscreen" coordinates in the .ini file).
Speaking of the .ini file. It might be worth to have a look in there. If the display problem occurs, go to Options, Save Options in no$xxx, and then check in .ini file in a text editor. Especially these lines:
fsys_xsiz, fsys_ysiz --> Filesystem Viewer size
Vram_xsiz,Vram_ysiz --> VRAM Viewer size
Vram_xloc,Vram_yloc --> VRAM Viewer position
The size values should be always positve numbers, simething like 00000xxx.
The position should be usually positive, or negative if the window is parts offscreen, or on another display left/above of the primary display, so that might be 0000xxxx or FFFFxxxx.
Another problem would be if there any values like 0000Fxxx (negative 16bit).
https://learn.microsoft.com/en-us/windo ... -functions
above is only for "functions" (some web browsers are able to display a table of contents on the left of the browser window, with more multiple display monitor related chapters, not only the functions).
From what I can see, the feature does rely on a "primary screen", and additional screens that can be arranged at left or right or above or perhaps below of the "primary screen". As shown here:
https://learn.microsoft.com/en-us/windo ... ual-screen
Of those, the monitors left/above of the primary screen can have negative coordinates, which may cause problems for software that expects unsigned coordinates (although, even with a single monitor, it's possible to move a window to parts offscreen locations on left/top sides (the mouse arrow can't be move to negative coordinates on a single monitor though)).
For the fullscreen feature, the standard system metrics function reports only the size of the "primary screen". So I assume that no$xxx in fullscreen mode will properly display the game screen on that monitor at correct size.
DPI... I wasn't aware of such features. I've found something similar in win98, under Display Properties, Settings, Advanced, General, Font Size (Normal or Large, with 96 or 120 dpi).
That seems to be a driver specific setting, and it may be absent or different for other display drivers. The effect seems to be that some things like Dialog box units are getting zoomed (affecting captions, meno bars, tab controls, scrollbars - but not affecting any pixel sizes of owner drawn bitmaps). I am not sure if that is the same DPI feature as on windows 11 (?)
Some of those settings like menu bar and tab control height are important for positioning controls and child windows within another window. And no$xxx does assume that those settings are same for all windows. But even if the menu bar height varies by +/-10 pixels, that should only cause minor glitches.
---
Some more questions:
Is there still any need to fix the issue, or is it accepable simply not to use different DPI settings?
Does the problem occur on the "primary display"?
Does the problem occur on "non-primary displays"?
Or both on primary and non-primary displays?
And are the non-primary ones at negative coordinates (left/above of the primary)?
Does it make a difference to close and re-open the window after dragging it to another monitor?
Is it even possible to drag no$xxx windows to different monitors? Theoretically, my code should prevent to move windows to completely offscreen locations (whereas "offscreen" would mean outside of the primary display area) (or at least so when storing "offscreen" coordinates in the .ini file).
Speaking of the .ini file. It might be worth to have a look in there. If the display problem occurs, go to Options, Save Options in no$xxx, and then check in .ini file in a text editor. Especially these lines:
fsys_xsiz, fsys_ysiz --> Filesystem Viewer size
Vram_xsiz,Vram_ysiz --> VRAM Viewer size
Vram_xloc,Vram_yloc --> VRAM Viewer position
The size values should be always positve numbers, simething like 00000xxx.
The position should be usually positive, or negative if the window is parts offscreen, or on another display left/above of the primary display, so that might be 0000xxxx or FFFFxxxx.
Another problem would be if there any values like 0000Fxxx (negative 16bit).
- MasterLink
- Active PSXDEV User
- Posts: 46
- Joined: Jul 20, 2024
- I am a: Electronics, graphics, music.
- PlayStation Model: DTL-H1001H
- Location: USA
The Windows 9x DPI setting basically just scaled the fonts, forcing UI elements to adjust. Windows 11 however does adjust bitmaps too. There is the ability to have custom DPI, but it seems to do a hybrid of the old style and new style and oh it's just awful, nothing ever looks right.
I'm now on driver version 566.36 and still can't replicate this issue with my Nvidia card, so I'm torn as to where the issue lies. Since I feel this may be a Windows 11 specific issue, I'm on Windows 11 24H2, build 26100.2605. Maybe the build itself is the issue? Also the OP hasn't stated what GPU they are using either, or driver version.
(Please, don't use "latest" driver as a reply, specify the version, drivers change so often that it's best to "trust but verify" here, as if it's a specific driver, I can test that specific version, as perhaps I missed it and skipped over it when updating this month.)
I'm now on driver version 566.36 and still can't replicate this issue with my Nvidia card, so I'm torn as to where the issue lies. Since I feel this may be a Windows 11 specific issue, I'm on Windows 11 24H2, build 26100.2605. Maybe the build itself is the issue? Also the OP hasn't stated what GPU they are using either, or driver version.
(Please, don't use "latest" driver as a reply, specify the version, drivers change so often that it's best to "trust but verify" here, as if it's a specific driver, I can test that specific version, as perhaps I missed it and skipped over it when updating this month.)
Retail Consoles: SCPH-1001 (Faulty CD Decoder Currently), SCPH-5501
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
Debugging Consoles: DTL-H1001H, SCPH-5501 (Via Unirom)
Who is online
Users browsing this forum: No registered users and 0 guests