-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
MacOS support? #2
Comments
finally someone with a mac 😀 |
@krauthaufen how can I help? (except stack trace from the bootstapping) |
not really but all other libs use semantic version, so we're quite used to writing ~> X.Y.Z
I think it might be hard to solve these super-low-level problems as a first issue. Our whole native dependency mechanism apparently doesn't work on MacOS. A friend of mine has a Mac and I'll see if I can borrow it for some days. But if you don't mind dealing with ugly native interop stuff you could sort out how to build our native-libs on macOS. These are:
and potentially some more... |
As an alternative to macOS laptop I can propose you to try Azure DevOps ... You are able to setup CI for you projects and run it for free even on macOS |
hey cool, that will help with the basic stuff but for rendering tests we sadly need a real GPU... Does apple still prevent users from using MacOS in a VM? |
sorry, I do not know |
just for the record (not really sure how important it is for you)
Does it means that virtualized GPU is not good enough? |
It basically means that we never cared about super-bad GPUs and the virtualized GPU is the worst GPU possible 😀 We could certainly make our OpenGL backend run on this kind of driver but it would require lots of special code which we simply didn't have the time/resources for. |
Hi try to compile it on macOs and I'm getting these errors: Details: Date/Time: 2020-02-12 18:31:54.987 +0100 Sleep/Wake UUID: F59F1518-8C21-4076-BFE4-FC4E08C92B43 Time Awake Since Boot: 5400 seconds System Integrity Protection: enabled Crashed Thread: 0 Dispatch queue: com.apple.main-thread Exception Type: EXC_CRASH (SIGABRT) Application Specific Information: Thread 0 Crashed:: Dispatch queue: com.apple.main-thread Thread 1: Thread 2: Thread 3: Thread 4: Thread 5: Thread 6: Thread 7: Thread 0 crashed with X86 Thread State (64-bit): Logical CPU: 0 Binary Images: External Modification Summary: VM Region Summary:
REGION TYPE SIZE COUNT (non-coalesced) Model: iMac12,1, BootROM 87.0.0.0.0, 4 processors, Intel Core i5, 2,5 GHz, 20 GB, SMC 1.71f21 |
The previous error generated by VisualStudio 0: Loader exceptions are 0: ReflectionTypeLoadException error in assembly Aardvark.UI Unhandled Exception: System.IO.FileLoadException: Could not load file or assembly 'Aardvark.SceneGraph, Version=4.12.3.0, Culture=neutral, PublicKeyToken=null'. The located assembly's manifest definition does not match the assembly reference. (Exception from HRESULT: 0x80131040) |
Hey, thanks for the report. Our company finally ordered a macbook for us 😁🤘, so we'll start to fix these problems once it's shipped... |
Great! I would love to help! |
Hey, good news: we got our macbook today and will start fixing things soon. Actually we already managed to save some images using DevIL 🤘 |
Good news everyone, Rendering/Media things should be working on MacOS now. The only example ported yet is https://github.com/aardvark-platform/aardvark.docs/tree/v5/src/SSAO ( |
It seems both Rendering and Media projects work on MacOS now with OpenGL backend! So I think we can close it.
|
Cool! It is builds and run on macOS! Application logs:
|
thanks for the info. really strange! especially since we have mac versions in production. what aardvark.rendering version was resolved? |
Update after the last concurrent post: ok - this is most likely a bug we fixed (framebuffer download). newest asrdvark.rendering should not have this problem |
updated my comments with |
aardvark-platform/aardvark.rendering@d704198 this one is the fix and should be in 5.1.14. interestingly the tag is missing in rendering - little hard to see on my mobile... edit: according to release notes in rendering 5.1.12 fixed it, subsequenced versions fixed assimp model loading on mac |
@sergey-tihon Could you please run |
@krauthaufen I have no idea ( no code changes to template)
@gsomix no it does not help I've created repo with code for both projects. (with paket.lock files)
update: the Link :) https://github.com/sergey-tihon/aadvark-macos |
@sergey-tihon Thanks! @krauthaufen If I understood correctly it's Media window ( |
i have two test machines and can check your repro on monday. thanks! @gsomix did you see this as well in your testing session @krauthaufen native works, ui not, pretty sure it is download although i fixed it. maybe it is multisampling related. need to check with "data-samples" "1" on the rendercontrol |
@haraldsteinlechner Nope, both Rendering and Media projects work just fine with 5.1.11 on my machine. |
@gsomix could you summarize your system or show a log which prints opengl stuff - if not much effort. maybe we can spot something... |
@haraldsteinlechner Sure. Link. |
superb, thanks. |
I did further investigation and tries the same code on my newer work MacBook and it run without issues. So it is definitely not general macOS issue, but rather something hardware/driver specific. Then I rebooted my personal laptop, unplug all displays and turned off all application - no success. Then I downloaded gfxCardStatus and have played with manual switching between video cards. I find out that is WORKS when I run on integrated video card
|
Okay, I suspect we're dealing with a AMD specific driver issue here. The bad thing is we don't have a GPU like this one. To do so you need to pass a With a little luck that might give us at least a hint where to search for the problem. Cheers |
I do not see much difference but here is a log https://gist.github.com/sergey-tihon/c3d27b707baeee7922c5f3eb9bdc8528 with |
Okay, so debug output either doesn't work or there are no OpenGL spec violations 😜 I suspect our texture download causes the problem, since OpenGL implementations are rather creative when it comes to making this unintuitive and the spec is rather vague there. Furthermore your image looks a bit like something with a wrong pixel-stride... I'm currently on vacation but @haraldsteinlechner I would try to set Btw are we using Maybe what we see there is the internal storage representation of a multisampled image??? Braindump end, back to vacation for now 😎 |
Okay, one other suspicion: Assumption: What we see here is every 4th (or so) pixel of the rendered cube in different poses (assuming you tried to move the camera) That (or something else related to blit) would potentially also explain what we see here. |
yes it looks like multisampling problem. actually ms counts are also not
checked properly in rendering. setting ```attribute "data-samples" "1"```
on render control or controlledcontrol could rule this one out. could you
try this? would be really cool to find it out :) (will prepare a gist)
also @krauthaufen dsa stuff is suspicious if this is reported as supported.
Georg Haaser ***@***.***> schrieb am So., 29. Aug. 2021,
22:26:
… Okay, one other suspicion:
Assumption: What we see here is every 4th (or so) pixel of the rendered
cube in different poses (assuming you tried to move the camera)
we're rendering to a 4x multisampled image, blit that to a single sampled
one (that is never cleared since we expect it to be overwritten every
frame) and maybe for some weird reason the blit does not "resolve" the
samples but instead just copies every fourth sample to every fourth target
pixel (no idea why someone would think that's a good idea)
That (or something else related to blit) would potentially also explain
what we see here.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAD5KAKE4CPZ4PL3WOSKSODT7KJYDANCNFSM4IZRON2A>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
that actually FIX the bug! With useMapping=false it runs without artifacts! Do you have idea why? Update: with this fix it works even without paket dependencies update (Aardvark.Rendering (5.1.11)) |
wow. this is really a great finding. just to wondering - did you by chance
try to resize the window when testing?
i have another track, maybe this is a sync issue @luithefirst?
Am 30.08.2021 07:39 schrieb "Sergey Tihon" ***@***.***>:
attribute "data-samples" "1" does not fix it... but searching through your
tests
<https://github.com/aardvark-platform/aardvark.media/blob/8ceadad2f3d810311790fa2d9916cff6780598ba/src/Scratch/08%20-%20ColorTest/App.fs>
to find out the place where I should specify this attribute I found another
one
attribute "useMapping" "false"
that actually FIX the bug! With useMapping=false
<sergey-tihon/aadvark-macos@d064385>
it runs without artifacts!
Do you have idea why?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAD5KAIIW4DHLPXRQBV5P7TT7MKPPANCNFSM4IZRON2A>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
@haraldsteinlechner I don't believe a broken memory-mapping could cause that sort of broken image... |
blocked by gatekeeper could do it maybe also yet unlikely. yes i was on to comparing download codes. @hyazinthh could you check and spot on differences we did not make consistent to all downloads? |
@haraldsteinlechner just tried resize and full screen - both does not fix image. |
What OS versions are supported?
failed with following error on
macOS
The text was updated successfully, but these errors were encountered: