View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000012 | Falcon BMS Known Bugs | General | public | 2020-12-30 18:21 | 2021-05-07 05:18 |
| Reporter | oz | Assigned To | I-Hawk | ||
| Priority | low | Severity | trivial | Reproducibility | always |
| Status | assigned | Resolution | open | ||
| Platform | PC | OS | Windows | OS Version | Windows 10 |
| Product Version | BMS 4.35 | ||||
| Summary | 0000012: MSI AFterburner don't display datas | ||||
| Description | MSi display data on 4.34 but not in 4.35. | ||||
| Tags | No tags attached. | ||||
| Theatre of Operations | KTO | ||||
| Confirmed by I-Hawk here: https://www.benchmarksims.org/forum/showthread.php?40929-4-35-Cannot-Get-MSI-Afterburner-Overlay-to-Display&p=560108&viewfull=1#post560108 | |
|
I spent sometime on tries to get it to work, but obviously something is wrong/missing. Seems like the OSD showing while in UI but get lost as you go 3D (In splash screen it's already gone). Also let me note that OSD shows fine on EVERY RTT display, if you use them. |
|
| I don't use rtt | |
|
Partial workaround -- the overlay from NVidia FrameView shows just fine .. on both 2D and 3D views. Doesn't show everything RivaTuner does, but covers the most important bits (cpu/gpu temps, avg and 1%-low framerates, etc) |
|
| oui tout comme la superposition du benchmark de falcon. | |
|
Something interesting I found. If you'll open the MSI Afterburner information page, it will show you the relevant running apps and how they run. I found that for some unexplained reasons BMS is considered to be a D3D11 app when entering and in UI, but then it changes to a D3D10 app (close the info window and open again after entering 3D) when BMS is in 3D world, and it stays like that even when going back to the UI. Now... just to make sure no bullshit rumors are coming out of this find: BMS is a D3D11 app, that is a fact, but apparently there is something which cause it to be reported as a D3D10 app. https://i.imgur.com/p6DF6Rl.png https://i.imgur.com/sdWVWJu.png |
|
| Could it be a bug in MSI Afterburner? | |
|
I dunno about DX10 but I do see d3dx9_43.dll loaded, at process-startup (before d3d11). No idea if that's expected/necessary or a sign of some old/dead code still being hot. I'm honestly not sure how these screen-overlay FPS counter tools work.. I suspect they hook into the DirectX swapchain-creation APIs and inject a small stub, to track the timing of calls to Present(), and inject draw calls to render the overlay. (?) If there's code still consuming old DX9 objects/APIs, maybe there is code also referencing DX10? and that might be enough to confuse RivaTuner. All just speculation, mind you. TL/DR - I don't see any D3D10 modules loaded so I agree it's probably a bug in RivaTuner / Afterburner, but maybe worth scanning the codebase to understand why d3dx9.dll is being loaded, if that's not expected, and audit if any DX10 era APIs are being called. |
|
|
JFYI - I see the same also with Nvidia's Nsight tool which also thinks BMS is going DX10 as it loads into 3D and it even say something like: App is running in DX10 so Nsight will not be able to debug (Not exact words but that was the meaning), but then Nsight is functioning perfectly and does all the debug it can do without problems. Conclusion from that behavior: BMS is being though of as running DX10 (reason unknown) but obviously it runs DX11 as Nsight does work perfectly with it and Nsight doesn't support anything below DX11. Regarding D3DX9 being used - Yes, some of the Vectors/Matrices/Functions of GFX are using D3DX types and so we still depend on that. However all that is "old" code, at some point I assume we will need to replace that anyway. |
|
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2020-12-30 18:21 | oz | New Issue | |
| 2020-12-30 21:05 | administrator | Severity | minor => trivial |
| 2020-12-31 05:08 | administrator | Status | new => confirmed |
| 2020-12-31 05:08 | administrator | Note Added: 0000022 | |
| 2020-12-31 05:53 | I-Hawk | Assigned To | => I-Hawk |
| 2020-12-31 05:53 | I-Hawk | Status | confirmed => assigned |
| 2020-12-31 05:54 | I-Hawk | Note Added: 0000024 | |
| 2021-01-21 11:04 | oz | Note Added: 0000160 | |
| 2021-02-11 23:57 | airtex2019 | Note Added: 0000250 | |
| 2021-02-12 03:14 | oz | Note Added: 0000252 | |
| 2021-02-12 03:41 | I-Hawk | Note Added: 0000254 | |
| 2021-05-06 23:13 | Blu3wolf | Note Added: 0000559 | |
| 2021-05-07 00:53 | airtex2019 | Note Added: 0000561 | |
| 2021-05-07 05:18 | I-Hawk | Note Added: 0000562 |