Could not find any compatible Direct3D Devices
 

The following message is displayed in the display properties- screen saver : Could not find any compatible Direct3D Devices.
This happens in the following circumstances: Open the display properties- screen saver. MA is running in the little screen. Click preview, enter testfish. MA is running. Put the computer to sleep while MA is running (in my case, close the laptop lid). Wake up the computer. MA is no longer running.The display properties-screen saver no longer shows MA in the little screen but displays the message mentioned above.
This is only the case when MA is running while putting the computer to sleep.
I run XP and beta 8k.
This might be an XP issue and is not hampering functionality, but I mention it anyway in case this is interesting to you.

i had this one too , and i had spyware.

Try to run Malwarebytes that solved the problem with me :)

Succes and let us know

No spyware...
 

Just ran Malwarebytes, no spyware found...

Thanks Highlander , I've been looking for a good anti malware program for a while now .

Does anybody else have this problem?

Latest beta on Windows 7 and I get the "Could not find any compatible Direct3D devices." error.

I just got it on Windows 7.

ur welcome

Could you explain the steps to reproduce this? Someone reported this in another thread, but in that case it only happened after the computer had gone into sleep mode, and then tried to go back into screensaving mode.

What is being tested, to determine when to display the "Could not find any compatible Direct3D devices" dialog box?

I have a computer that, with dxdiag, in the "Test Direct3D" section, passes all of the Direct3D 7, 8, and 9 tests. All tests successful.

However, MA3Beta8k insists (when trying to select MarineAquarium3 as a screensaver) on giving me the "Could not find any compatible Direct3D devices" dialog box?

Unless I can reproduce the bug, there's little chance of my being able to fix it.

Assuming that was a response to my "what is being tested" question - I can certainly ship the computer to you, if that's needed.

I can tell you it's running Intel 82810 graphics (yes, it's old) and dxdiag says it passes all tests (and yes, I can see the dxdiag rotating cubes).

The problem also shows up some things that I think are "bugs" (or suggested improvements), but I thought I would ask the simple question first.

Perhaps Prolific wrote the piece of code that tests for Direct3D device compatibility.

Actually, that's Microsoft boilerplate code. Don't go shipping any computers - I barely have room to walk around here now.

Did you say that it only happens with Beta8k, and not previous versions?

Oh, sorry - I misread your comment. Here are the steps.

On the system in question, ONE of the ways to reproduce it is:
Conditions: MarineAquarium3 is *NOT* the current default screensaver.
Riught-click on desktop, go to Properties, go to Screensaver tab, scroll through the list and click on MarineAquarium3. [It's not necessary to click on Apply].

The mini-preview shows as completely black, and the previously-mentioned dialog box appears.

Clicking OK clears out the dialog box, of course.

Hope that helps. As I recall, I ran into one or two other repeatable things that will make the dialog box appear. I can try to re-discover them, if that would be useful.

Oh, I think that dialog box also appeared at the very end of the installation process, when the installation asks "do you want to make this the default screensaver" (or whatever the question is). That's not easy to recreate (although possible).

It's the Installer version I just bought. I thought I would first install it on a "spare" computer that had never been touched by MA3Beta-anything. It's WXP-Pro, and had (has) MA2.6 installed.

Gotta run - more later.

Dale,
Are you able to run the screensaver(MA3) on that machine without errors?
Does it always get the error?
What is machine's spec?

The right-click-installation feature (it's a Windows thing, not ours) has an annoying characteristic. Even though you have right-clicked on the copy that you wish to deal with, Windows looks through its list of known screensavers (usually in the System32 directory) for a local copy of the same program. The name that it uses is the internal name, not the filename. If it finds a match, it runs that copy, NOT the one you clicked on.

So, when testing a new version, it's important remove all old versions in the system32 drawer (or wherever your OS likes to keep screensavers). I'm not sure if this is what's happening here, but it is a feature of Windows which makes it very difficult to tell if something has been fixed because you never know exactly which copy is running. This only applies to right-clicking the file -- if you double-click the file, then that's the copy that will run.

OK, let me clear that up first (again).

I didn't do a "right click-installation". I downloaded the .exe from SereneScreen, and ran it.

I chose this machine to test first, because it had NEVER had MA3 on it before. Just MA2.6. And it had been sitting on a shelf for about 2 months, so I had to (chose to) do a bunch of Windows and Office updates to it first, and also ran the DirectX9c updater to make sure that was up to date.

Also, before I ran the MA3 installer, I ran dxdiag (which is, of course, a Microsoft test program) to check for Direct3D function and verify that I truly had version 9c.

As I said, the computer passed all tests for Direct3D functionality (7, 8, and 9) - visually displaying the rotating cube reasonably well.

The system is a Compaq Presario 5BW250, latest BIOS and video chipset drivers from Compaq, Celeron 700 processor (yeah, I know the spec is Pentium 800 minimum). i810E (82810E) chipset, 512 MB memory.

I understand that this system is wimpy and I didn't expect great performance. Jerky movements, etc., would be expected.

But in addition to testing the installer, I thought I would test the current MA3 on a machine that was at essentially the lower limit of the specified requirements.

Another issue would be, how would some "plain user" determine if his/her machine would run MA3. [Yes, I recognize that the SereneScreen site "encourages" downloading the trial version, etc. But it's common to just ignore that when somebody says "Hey, this is great, you should buy it". Or you get it as a gift.]

That last bit would be a different topic, for another day.

Sorry, I read "Right-click on the desktop" as "Right-click on the icon". It's been a long 72 hours - getting these bugs chased down, trying to implement the clock/calendar/logo system, and creating the new website.

That system might be off the bottom end of the chart, but I've got a weak old Sony Vaio that still runs MA3 at about 20 FPS. How much video RAM does your card have?

Please note, for future reference, that I did what I could as an experienced professional, and tested, to determine that the machine "closely" met the minimum specifications. And that I then ran the installer.

No, it doesn't run in any mode without errors. When started (by any method) it produces a black screen. That includes clicking on the desktop icon, running "preview", or letting it start as a screensaver. It "does something with a black screen" for about 60 seconds, and then produces the "...has experienced an error and must quit - would you like me to tell Microsoft" box.

HOWEVER - the "could not find any compatible 3D devices" dialog box ONLY appears (as far as I can tell) when it's trying to do the "mini-preview" in the Screensaver tab of Display Properties. And acknowledging (OK) the dialog box allows me to go ahead and set MA3 as the default screensaver.

Relative to the simulation program (Jim's stuff), I suggest the following:

A. If MA3 tries to run, it shouldn't "crash" even if there's no compatible Direct3D device (easy for me to say - perhaps not easy to do). It could, of course, exit with an informative message.

B. If the test for "compatible 3D devices" fails, MA3 should not try to run. [I suspect this implies that the test should be done every time MA3 executes]

C. If the test for "compatible 3D devices" fails at the "screensaver selection" phase, then "apply as default screensaver" should not be allowed (perhaps hard to do?).

Sorry, I recognize that this isn't a fun set of comments.

Oh, 32MB - which is only half of the "minimum spec". In my experience, that isn't a killer issue - it just makes performance really "inhale briskly".

However, tomorrow I'll try to dig up some old video card(s) to try in the machine to see what happens.

Note 1: My suggestions are really aimed at "if it isn't going to work, then don't run".

Note 2: This is of course NOT a FIRST-priority issue. But I think it needs to be sorted out before a "production version" is declared.

Note 3: There's also a related installer issue - which I will tackle in the Installer thread.

That old Vaio I mentioned only has a 32-meg video card (chip), but it is still able to use the HAL and not fall back to software rendering. Speed is poor - you can tell that the system is paging the textures - and there are occasional video hiccups. But it does run MA3 without errors, so it seems like your 32-meg card should, too. I'll be interested to see if swapping cards has any effect.

I think 82810 does not support pixel or vertex shader.
It will pass DirectX and MA2.6 specs but will fail MA3.

Jim, which Intel graphics card do you have?

Edit: http://www.intel.com/support/graphics/sb/cs-014257.htm

Intel 82915G.

MA3 uses Vertex Shaders, but not Pixel Shaders.

Good catch, Edgar. That's probably the problem, but it would seem that the proper drivers would step in and take over for the shader. Some people have reported that this type of problem is fixed by new drivers.

It looks like you have the lowest Intel card with Shader support.

I checked to make sure I have the latest driver version from Compaq. However, that turns out to be a higher version than the latest driver version from Intel. I'll try that one later today or tomorrow.

Of course, how do you tell "Joe User" what to look for?

I read somewhere but wasn't sure, that 82810 does NOT support multi-textures also.

I don't see how the Aquarium could deal with this situation, short of having a table of all supported hardware.

So many variations on the same video card hardware are released with just minor changes that this table would be massive and impossible to maintain.

Fortunately, there's always MA2.6.


P.S. You may not be, but I'm really impressed with the support you're getting in this thread. The research into various video cards and the functionality that they offer.

I have a kludge built into the program to deal with cards that have no multi-texture capability. That's how I'm able to get it to work on my old Vaio and certain ATI cards. The fish don't look quite as realistic (the shading, especially), but the casual observer probably wouldn't notice the difference unless they saw a side-by-side comparison.

Jim,
But if the driver and card does not support shaders, then the card is out of luck. Do you still plan to try to support it?

Yes, I'm impressed with this thread.

With respect to "dealing with" this issue, there are (in my view) two separate aspects: (a) how the simulation software deals with the issue, and (b) how Prolific deals with the issue. I think both parts have reasonable solutions. This thread is about part (a). And we're still doing diagnostics, but I believe that's just from intellectual curiosity.

One aspect not already mentioned, is what to do if the user changes video cards to a new one that does not support MA3 requirements (after MA3 has already been installed). [Or, off-topic, uninstalls DirectX9.0c]

I've made a suggestion about how to handle the part (a) issues - "simply" detect if MA3 isn't going to run properly, and don't run. [Of course, not really simple - but the alternative is to do something unpredictable like "black screen" or "blue screen"]

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

I agree that a "massive out-of-date table" isn't appropriate as some of part (b). However, this thread may perhaps allow Prolific to more-correctly describe the video requirements. Clearly "Direct3D" isn't sufficient. I don't really expect a lot of people to read and analyze the requirements, but I am confident that Prolific wants them to be clear and comprehensive. For several obvious reasons.

Edgar - no.

OK, after about 2 hours of testing, I've nailed down what is happening with the builtin chipset (haven't tried another card yet, and maybe won't).

This chipset allows two screen resolutions (1024x768 and 800x600) with a standard monitor, and two color qualities (16-bit and 24-bit). [And several different refresh rates, which I didn't test - yet]. And I have 3 driver versions on hand - two Compaq and one Intel.

I finally "got a clue" that indicated that I should test all of the permutations. Note that's 12 combinations to test, with reboots, etc. But I'll simplify the results.

It's not directly the chipset. Either resolution (1024x768 or 800x600), at 24-bit quality, fails. Either resolution at 16-bit quality WORKS. (Well, "works" is sorta judgmental, but we expected that).

The good news, for this particular issue, is that the test which produces the "Could not find any compatible Direct2D devices" message seems to be an exact discrimination. If the message appears when selecting MA3 as screensaver, MA3 will not run in any full-screen mode. (I have not yet tested it in windowed mode, so perhaps it's different with "start in windowed mode"? -- maybe back to the test bench?)

Bottom line: if the test fails, don't let it run, and tell the user why.

Footnote: it's obviously not exactly correct/sufficient to say it couldn't find a compatible Direct3D device - because the device never changed.

Thanks for all the testing.

The term "Direct3D Device" has an exact meaning in the Microsoft lexicon. People tend to think of it as meaning hardware, but it's really the hardware/software/resolution/bit-depth combination that allows 3D to be displayed. If any part of the software or driver is missing or damaged, the Direct3D Device fails.

I've tried to account for just about anything that could go wrong within the program, and instead of just locking up or crashing, I display Mircosoft's recommended bouncing box with text listing the error. If one of these tests fails, it returns an error code, which I run through a Microsoft program to convert it to human-readable text. So, if the message seems cryptic, blame them, not me. Maybe some day I'll have time to go through them all and replace some of these messages with my own wording, but that's about 30 years down on a very long list of things-to-do.

I think what you are asking for is that the program not run AT ALL, or immediately shut down if it encounters an error. I'm not sure this is possible. Microsoft is pretty picky about these screenblankers being able to cover up sensitive information on a monitor. If someone leaves his desk at the DMV assuming that the screen blanker is about to come on, it's better if there's at least a black screen with a bouncing box than someone noticing that he'd been playing Solitaire.

Sorry to again make this "not fun", but I believe I have been unsuccessful in communicating, once again. Simply put, MA3 crashes in each of the three startup modes.

It looks to me like it crashes. "MarineAquarium3 Beta8 has encountered a problem and needs to close."

FIRST, re: screensaver mode (other modes later):

Yes, that's what I'm suggesting. But I'm never seeing a bouncing box with text. As you say, switching to some other screensaver would be an acceptable alternative.

What I am seeing (in the previously described instances), when the screensaver starts up, is:

A. Black screen - no logo. Followed in a couple of seconds with an hourglass.

B. About 15 seconds later, the "standard" Microsoft dialog box that says "MarineAquarium3 Beta8 has encountered a problem and needs to close.", followed by the usual invitation to notify Microsoft. In my experience, that dialog box is never deliberately generated by program authors.

===========
SECOND: the other 2 startup modes (click on the icon, and preview) - both also operate the same way. Black screen followed by "MarineAquarium3 Beta8 has encountered a problem and needs to close."

So, in those two startup modes, you're dealing with the simulation, not the screensaver.

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

From an external (program behavior) viewpoint, it looks to me just like the program crashes. Any way it's started.

You clearly have some sort of test that is (at least in the environment I'm testing) absolutely diagnostic. That test runs every time MarineAquarium3 is selected in the Screen Saver tab of the Display Properties dialog box is clicked. I suspect (hard to tell without seeing the code) that the test actually runs each time the MINI-Preview (in the Screen Saver pane) tries to start. That test currently just pops up a dialog box warning about the (claimed) absence of Direct3D. Clicking "OK" allows everything else to just go ahead and happen (for instance, clicking on "Apply" to select this as the default screensaver, regardless of the warning). [And the MINI-preview doesn't run and doesn't crash]

But, as far as I can tell, that test is never run when starting MA3 by clicking on the desktop icon, OR by clicking on the Preview button on the Screen Saver panel, OR when the screensaver starts automatically.

My suggestion is that perhaps you consider running that test for each of those three starting methods. (Perhaps it's really one place - I can't see the code, of course). If the test fails, do something intelligent at that point - either just pop up the "no Direct3D" dialog box, or (where appropriate) start some other screensaver.

That will prevent MA3 from looking like it crashed (whether or not it did).

However, there's at least one other issue that I'll detail later - this reply is already long enough, and perhaps it's helpful to make sure one issue is understood before we go on.

Note: I'm just trying to make sure the issue is clearly understood. How/if it is fixed, is up to the software author.

That's quite a post. I'll need to wait 'til morning to wade into it.

Just a wild-hair thought. My Compaq is showing the error when the "color quality" is 24-bit. That's a rather unusual setting (but as high as this wimpy machine goes). Does your old Vaio have that available? [Or can somebody else test?] Is there a chance that MA3 croaks on 24-bit color?

Probably not what's happening, but it's worth asking.

No, the old Vaio only has 16 and 32-bit available. The Aquarium runs at 9 FPS in 32-bit and 14 FPS in 16-bit. Like all my other machines, dithering makes the two modes virtually indistinguishable visually, so I always run everything in 16-bit.

OK, I'll see if I can dredge up some other system that has 24-bit.

I presume you've digested my reply #35. As others have indicated, simulation "features" are more important at the moment.

I have one other set of problems that is related to this topic. Would you prefer that I describe it in this topic, or in a new topic?

[Of course, I could save it for later posting - or never post it - but I prefer to get Beta-testing observations on the record, and move on]

Unless they are trivial things to fix, hold off for now. I'm totally swamped.