I'm extending an AudioUnit to generate multi-channel output, and trying to write a unit test using AVAudioEngine. My test installs a tap on the AVAudioNode's output bus and ensures the output is not silence. This works for stereo. I've currently got: auto avEngine = [[AVAudioEngine alloc] init]; [avEngine attachNode:avAudioUnit]; auto format = [[AVAudioFormat alloc] initStandardFormatWithSampleRate:44100. channels:channelCount]; [avEngine connect:avAudioUnit to:avEngine.mainMixerNode format:format]; where avAudioUnit is my AU. So it seems I need to do more than simply setting the channel count for the format when connecting, because after this code, [avAudioUnit outputFormatForBus:0].channelCount is still 2. Printing the graph yields: AVAudioEngineGraph 0x600001e0a200: initialized = 1, running = 1, number of nodes = 3 ******** output chain ******** node 0x600000c09a80 {'auou' 'ahal' 'appl'}, 'I' inputs = 1 (bus0, en1) <- (bus0) 0x600000c09e00, {'aumx' 'mcmx' 'appl'}, [ 2 ch, 44100 Hz, 'lpcm' (0x00000029) 32-bit little-endian float, deinterleaved] node 0x600000c09e00 {'aumx' 'mcmx' 'appl'}, 'I' inputs = 1 (bus0, en1) <- (bus0) 0x600000c14300, {'augn' 'brnz' 'brnz'}, [ 2 ch, 44100 Hz, 'lpcm' (0x00000029) 32-bit little-endian float, deinterleaved] outputs = 1 (bus0, en1) -> (bus0) 0x600000c09a80, {'auou' 'ahal' 'appl'}, [ 2 ch, 44100 Hz, 'lpcm' (0x00000029) 32-bit little-endian float, deinterleaved] node 0x600000c14300 {'augn' 'brnz' 'brnz'}, 'I' outputs = 1 (bus0, en1) -> (bus0) 0x600000c09e00, {'aumx' 'mcmx' 'appl'}, [ 2 ch, 44100 Hz, 'lpcm' (0x00000029) 32-bit little-endian float, deinterleaved] So AVAudioEngine just silently ignores whatever channel counts I pass to it. If I do: auto numHardwareOutputChannels = [avEngine.outputNode outputFormatForBus:0].channelCount; NSLog(@"hardware output channels %d\n", numHardwareOutputChannels); I get 30, because I have an audio interface connected. So I would think AVAudioEngine would support this. I've also tried setting the format explicitly on the connection between the mainMixerNode and the outputNode to no avail.

I'm getting the following error on Intel Iris integrated graphics. Code works well on newer Mac GPUs as well as Apple GPUs. Execution of the command buffer was aborted due to an error during execution. Invalid Resource (00000009:kIOAccelCommandBufferCallbackErrorInvalidResource) The error is for a compute command, not a draw command. The constant isn't in the documentation. All buffers and textures seem to be created successfully. I've also checked that the GPU supports the required threadgroup size for the compute pipeline. thanks!

How should an App Extension (in this case an Audio Unit Extension) determine if an IAP has been purchased in the containing app? (and related: can an IAP be purchased from within the extension?) On macOS, I suppose I could share the receipt file with the extension? and on iOS, suppose I could write some data to shared UserDefaults in an app group. Is there any official guidance on this? thanks!

I have on the order of 50k small meshes (~64 vertices), all different connectivity, some subset of which change each frame (generated by a compute kernel). Can I render those in a performant way with Metal? I'm assuming 50k separate draw calls would be too slow. I have a few ideas: encode those draw calls on the GPU or lay out the meshes linearly in blocks, with some maximum size, and use a single draw call, but wasting vertex shader threads on the blocks that aren't full or use another kernel to combine the little meshes into a big mesh thanks!

I'm modifying <1mb of a 256mb managed buffer (calling didModifyRange), but according to Metal System Trace, the GPU copies the whole buffer (SDMA0 channel, "Page On 268435456 bytes"), taking 13ms. I'm making lots of small modifications (~4k) per frame. I also tried coalescing into a single call to didModifyRange (~66mb) and still the entire buffer is copied. I also tried calling didModifyRange for the first byte, and then the copied data is small. So I'm wondering why didModifyRange doesn't seem to be efficient for many small updates to a big buffer?

Why Auto Layout instead of the horizontal and vertical boxes used by other systems like Qt? Or a layout system like CSS?What is missing in the other ways of doing this stuff?Over a couple years, Auto Layout has given me mostly headaches, and I now generally try to avoid it, so I want to know if there's a good reason for my suffering.thanks!