hi everyone, 我们发现了一个和Metal相关崩溃。应用中使用了Metal相关的接口，在进行性能测试时，打开了设置-开发者-显示HUD图形。运行应用后，正常展示HUD，但应用很快发生了崩溃，日志主要信息如下： Incident Identifier: 1F093635-2DB8-4B29-9DA5-488A6609277B CrashReporter Key: 233e54398e2a0266d95265cfb96c5a89eb3403fd Hardware Model: iPhone14,3 Process: waimai [16584] Path: /private/var/containers/Bundle/Application/CCCFC0AE-EFB8-4BD8-B674-ED089B776221/waimai.app/waimai Identifier: Version: 61488 (8.53.0) Code Type: ARM-64 Parent Process: ? [1] Date/Time: 2025-06-12 14:41:45.296 +0800 OS Version: iOS 18.0 (22A3354) Report Version: 104 Monitor Type: Mach Exception Exception Type: EXC_BAD_ACCESS (SIGBUS) Exception Codes: KERN_PROTECTION_FAILURE at 0x000000014fffae00 Crashed Thread: 57 Thread 57 Crashed: 0 libMTLHud.dylib esfm_GenerateTriangesForString + 408 1 libMTLHud.dylib esfm_GenerateTriangesForString + 92 2 libMTLHud.dylib Renderer::DrawText(char const*, int, unsigned int) + 204 3 libMTLHud.dylib Overlay::onPresent(id<CAMetalDrawable>) + 1656 4 libMTLHud.dylib CAMetalDrawable_present(void (*)(), objc_object*, objc_selector*) + 72 5 libMTLHud.dylib invocation function for block in void replaceMethod<void>(objc_class*, objc_selector*, void (*)(void (*)(), objc_object*, objc_selector*)) + 56 6 Metal __45-[_MTLCommandBuffer presentDrawable:options:]_block_invoke + 104 7 Metal MTLDispatchListApply + 52 8 Metal -[_MTLCommandBuffer didScheduleWithStartTime:endTime:error:] + 312 9 IOGPU IOGPUNotificationQueueDispatchAvailableCompletionNotifications + 136 10 IOGPU __IOGPUNotificationQueueSetDispatchQueue_block_invoke + 64 11 libdispatch.dylib _dispatch_client_callout4 + 20 12 libdispatch.dylib _dispatch_mach_msg_invoke + 464 13 libdispatch.dylib _dispatch_lane_serial_drain + 368 14 libdispatch.dylib _dispatch_mach_invoke + 456 15 libdispatch.dylib _dispatch_lane_serial_drain + 368 16 libdispatch.dylib _dispatch_lane_invoke + 432 17 libdispatch.dylib _dispatch_lane_serial_drain + 368 18 libdispatch.dylib _dispatch_lane_invoke + 380 19 libdispatch.dylib _dispatch_root_queue_drain_deferred_wlh + 288 20 libdispatch.dylib _dispatch_workloop_worker_thread + 540 21 libsystem_pthread.dylib _pthread_wqthread + 288 我们测试了几个不同的机型，只有iPhone 13 Pro Max会发生崩溃。 Q1：为什么会发生这个崩溃？ Q2：相同的逻辑，为什么仅在iPhone 13 Pro Max机型上出现崩溃？ 期待您的解答。

Hello experts, I'm trying to implement a material with custom shader code, but I saw that visionOS doesn't allow you to inject custom Metal functions or use CustomMaterial like iOS/macOS, nor can you directly write Metal Shading Language (.metal) and use it through ShaderGraphMaterial. So my question is, if i want to implement your own shader code, how should i do it?

Hi all, I'm running into an issue with an app that previously worked fine on device using visionOS 2.0. After updating to visionOS 26, the same code runs fine in the simulator but crashes on the device with the following error: -[MTLDebugComputeCommandEncoder _validateThreadsPerThreadgroup:]:1330: failed assertion `(threadsPerThreadgroup.width(32) * threadsPerThreadgroup.height(32) * threadsPerThreadgroup.depth(1))(1024) must be <= 832. (kernel threadgroup size limit)` Is there any documented way to check or increase the allowed threadsPerThreadgroup size on Apple Vision Pro? Or any recommended workaround for this regression? Thanks in advance!

Hi there, I'm wondering if it's possible under iOS 28 developer beta to enable MetalFX scaling info with '{"MTL_HUD_ENABLED": "1" for my App. This information has been added to Mac, but looks to be absent on iPhone / iPad

I'm using a class with tangents to render on RealityKit for VisionOS but in Vision26 it cause a crash on App and there not documentation how implement cp_drawable_compute_projection I have tried a few options but without success. Could you help me to implement it ? The part of code is: return drawable.views.map { view in let userViewpointMatrix = (simdDeviceAnchor * view.transform).inverse let projectionMatrix = ProjectiveTransform3D( leftTangent: Double(view.tangents[0]), rightTangent: Double(view.tangents[1]), topTangent: Double(view.tangents[2]), bottomTangent: Double(view.tangents[3]), nearZ: Double(drawable.depthRange.y), farZ: Double(drawable.depthRange.x), reverseZ: true ) let screenSize = SIMD2(x: Int(view.textureMap.viewport.width), y: Int(view.textureMap.viewport.height)) return ModelRendererViewportDescriptor(viewport: view.textureMap.viewport, projectionMatrix: .init(projectionMatrix), viewMatrix: userViewpointMatrix * translationMatrix * rotationMatrix * scalingMatrix * commonUpCalibration, screenSize: screenSize) }

I maintain a couple of CoreImage libraries that provide custom Metal kernel backed CIFilters. In iOS/iPadOS 26, the CIColorKernel.apply() method invoked in the CIFilter subclass fails to add the coreimage::destination parameter to the Metal function call: -[CIColorKernel applyWithExtent:arguments:options:] argument count mismatch for kernel 'FractalNoise3D', expected 13 but saw 12. I've compiled the code with Xcode 26 and deployed to iOS 18 devices without any breakage, so this is definitely an iOS problem, not an Xcode problem. Library here: https://github.com/JoshuaSullivan/SimplexNoiseFilter Feedback ID: FB17874311

I'm building a weather map that shows the rain on the map. I'm able to retrieve PNG images that are used as tiles to put onto the map. I then reload all the tiles on the map with each timeframe (tile set for every 10 minutes). I'm able to get the map loaded up and I'm able to place the tiles and reload the data for each time slot. I preload all the PNG data needed for the tiles and store that NSData for them in memory so that they are quick for loading and showing on the map. I have timer's set to reload the overlay with the next set of tiles for each time slot. Giving the view of a moving precipitation map over time (just like you'd see on any weather map.) I have 12 time slots (timestamps) showing every 10 minutes for the past 2 hours. I have it showing each in sequence and then repeating. Over time I get a crash with this error as a Thread 1: signal SIGABRT. Failed to acquire drawable, rendering to temporary texture validateRenderPassDescriptor:782: failed assertion `RenderPass Descriptor Validation MTLRenderPassAttachmentDescriptor MTLStoreActionMultisampleResolve store action at attachment 0 requires resolve texture ' validateRenderPassDescriptor:782: failed assertion `RenderPass Descriptor Validation MTLRenderPassAttachmentDescriptor MTLStoreActionMultisampleResolve store action at attachment 0 requires resolve texture ' Through some searching I've discovered that this seems to be console output from Metal. I assume Metal is used for MapKit to render the overlay tiles? I'm using the same custom overlay where I set the timestamp on it and then tell it to reload. I also reuse the same MKOverlayRenderer as shown here... - (MKOverlayRenderer*)mapView:(MKMapView*)mapView rendererForOverlay:(id<MKOverlay>)overlay { if ([overlay isKindOfClass:[MKTileOverlay class]]) { if (!self.rainRenderer) { self.rainRenderer = [[MKTileOverlayRenderer alloc] initWithTileOverlay:overlay]; self.rainRenderer.alpha = 0.5; } return self.rainRenderer; } return nil; } And here's the function that reloads the overlay... - (void) updateRainFrame { self.currentFrameIndex = (self.currentFrameIndex + 1) % self.timestamps.count; if ((self.currentFrameIndex >= 0) && (self.timestamps.count > self.currentFrameIndex)) { NSLog (@"self.currentFrameIndex = %lu", self.currentFrameIndex); NSString *timestamp = self.timestamps[self.currentFrameIndex]; [self.overlay setTimestamp:timestamp]; [self.rainRenderer reloadData]; } } The time it takes to crash seems arbitrary. Sometimes it's very quick. Less than a minute. But usually it's several minutes. 10 or 20 minutes or more. Feels like some sort of race condition that's occurring. Perhaps ARC is not able to release the images for the tiles quick enough for each overlay reload? That's a wild guess but I think it's something more deeper in Metal as I feel I would see other errors related to memory availability. Some of my searches point to something about MSAA needing to be turned off in Metal to resolve this. However I have no idea how I would do that through MapKit. Any suggestions? Let me know if there is somehow a way to capture more from the crash to give more insight.

I am develop visionOS app. I am now very interested in Metal and Compositor Services, but I have not explored them in depth. I know that Metal has a higher degree of control freedom. I am wondering if using Compositor Services will have fewer functions than RealityKit in AR technology (such as scene reconstruction and understanding, hover effect, etc.).

Just downloaded XCode 26 and I see build fails despite Metal toolchain 26.0 downloaded. What am I missing? cannot execute tool 'metal' due to missing Metal Toolchain; use: xcodebuild -downloadComponent MetalToolchain

Hi, I'm developing a SwiftUI app using RealityKit and ARKit for an AR measuring feature. I’ve noticed that after navigating away from my AR view and performing extensive cleanup (including removing all anchors/entities, pausing the ARSession, and nil-ing out all references), memory usage remains elevated and sometimes grows with repeated AR sessions. Each time I enter and exit the AR view, memory increases The memory does not return to the baseline after cleanup, even though all custom objects are deallocated. Are there best practices beyond what I’ve described to ensure all ARKit/RealityKit resources are released after an AR session?

Hey, I need to know how to use texture mapping for rendering a spectrogram in metal. As I need smoothens the spectrogram. In my current project I am using vertex based approach which results in blocky behaviour between each quad. I need to smooth across each qaud so that It will smoothly gradient over.

Hi everyone, I'm creating an educational App that allows doing computational design in an immersive environment with the Vision Pro. The App is free and can be found here: https://apps.apple.com/us/app/arcade-topology/id6742103633 The problem I have is that the mesh of voxels I currently create use ModelEntity and I recently read that this is horrible for scalability. I already start to see issues when I try to use thousands of voxels. I also read somewhere that I should then take advantage of GPUs and use metal to that end. I was wondering if someone could point me to a tutorial or article that discusses this. In essence, I need to create a 3D voxel mesh, and those voxels have to update their opacity within an iterative loop. Thanks! —Alejandro

Hey, I've been struggling with this for some days now. I am trying to write to a sparse texture in a compute shader. I'm performing the following steps: Set up a sparse heap and create a texture from it Map the whole area of the sparse texture using updateTextureMapping(..) Overwrite every value with the value "4" in a compute shader Blit the texture to a shared buffer Assert that the values in the buffer are "4". I have a minimal example (which is still pretty long unfortunately). It works perfectly when removing the line heapDesc.type = .sparse. What am I missing? I could not find any information that writes to sparse textures are unsupported. Any help would be greatly appreciated. import Metal func sparseTexture64x64Demo() throws { // ── Metal objects guard let device = MTLCreateSystemDefaultDevice() else { throw NSError(domain: "SparseNotSupported", code: -1) } let queue = device.makeCommandQueue()! let lib = device.makeDefaultLibrary()! let pipeline = try device.makeComputePipelineState(function: lib.makeFunction(name: "addOne")!) // ── Texture descriptor let width = 64, height = 64 let format: MTLPixelFormat = .r32Uint // 4 B per texel let desc = MTLTextureDescriptor() desc.textureType = .type2D desc.pixelFormat = format desc.width = width desc.height = height desc.storageMode = .private desc.usage = [.shaderWrite, .shaderRead] // ── Sparse heap let bytesPerTile = device.sparseTileSizeInBytes let meta = device.heapTextureSizeAndAlign(descriptor: desc) let heapBytes = ((bytesPerTile + meta.size + bytesPerTile - 1) / bytesPerTile) * bytesPerTile let heapDesc = MTLHeapDescriptor() heapDesc.type = .sparse heapDesc.storageMode = .private heapDesc.size = heapBytes let heap = device.makeHeap(descriptor: heapDesc)! let tex = heap.makeTexture(descriptor: desc)! // ── CPU buffers let bytesPerPixel = MemoryLayout<UInt32>.stride let rowStride = width * bytesPerPixel let totalBytes = rowStride * height let dstBuf = device.makeBuffer(length: totalBytes, options: .storageModeShared)! let cb = queue.makeCommandBuffer()! let fence = device.makeFence()! // 2. Map the sparse tile, then signal the fence let rse = cb.makeResourceStateCommandEncoder()! rse.updateTextureMapping( tex, mode: .map, region: MTLRegionMake2D(0, 0, width, height), mipLevel: 0, slice: 0) rse.update(fence) // ← capture all work so far rse.endEncoding() let ce = cb.makeComputeCommandEncoder()! ce.waitForFence(fence) ce.setComputePipelineState(pipeline) ce.setTexture(tex, index: 0) let threadsPerTG = MTLSize(width: 8, height: 8, depth: 1) let tgCount = MTLSize(width: (width + 7) / 8, height: (height + 7) / 8, depth: 1) ce.dispatchThreadgroups(tgCount, threadsPerThreadgroup: threadsPerTG) ce.updateFence(fence) ce.endEncoding() // Blit texture into shared buffer let blit = cb.makeBlitCommandEncoder()! blit.waitForFence(fence) blit.copy( from: tex, sourceSlice: 0, sourceLevel: 0, sourceOrigin: MTLOrigin(x: 0, y: 0, z: 0), sourceSize: MTLSize(width: width, height: height, depth: 1), to: dstBuf, destinationOffset: 0, destinationBytesPerRow: rowStride, destinationBytesPerImage: totalBytes) blit.endEncoding() cb.commit() cb.waitUntilCompleted() assert(cb.error == nil, "GPU error: \(String(describing: cb.error))") // ── Verify a few texels let out = dstBuf.contents().bindMemory(to: UInt32.self, capacity: width * height) print("first three texels:", out[0], out[1], out[width]) // 0 1 64 assert(out[0] == 4 && out[1] == 4 && out[width] == 4) } Metal shader: #include <metal_stdlib> using namespace metal; kernel void addOne(texture2d<uint, access::write> tex [[texture(0)]], uint2 gid [[thread_position_in_grid]]) { tex.write(4, gid); }

I'm trying to build an MDLMesh then add normals let mdlMesh = MDLMesh.newBox(withDimensions: SIMD3<Float>(1, 1, 1), segments: SIMD3<UInt32>(2, 2, 2), geometryType: MDLGeometryType.triangles, inwardNormals:false, allocator: allocator) mdlMesh.addNormals(withAttributeNamed: MDLVertexAttributeNormal, creaseThreshold: 0) When I render the mesh, some normals are (0,0,0). I don't know if the problem is in the mesh, or in the conversion to MTKMesh. Is there a way to examine an MDLMesh with the geometry viewer? When I look at the variable values for my mdlMesh I get this: Not too useful. I don't know how to track down the normals. What's the best way to find out where the normals getting broken?

I am building a MacOS desktop app (https://anukari.com) that is using Metal compute to do real-time audio/DSP processing, as I have a problem that is highly parallelizable and too computationally expensive for the CPU. However it seems that the way in which I am using the GPU, even when my app is fully compute-limited, the OS never increases the power/performance state. Because this is a real-time audio synthesis application, it's a huge problem to not be able to take advantage of the full clock speeds that the GPU is capable of, because the app can't keep up with real-time. I discovered this issue while profiling the app using Instrument's Metal tracing (and Game tracing) modes. In the profiling configuration under "Metal Application" there is a drop-down to select the "Performance State." If I run the application under Instruments with Performance State set to Maximum, it runs amazingly well, and all my problems go away. For comparison, when I run the app on its own, outside of Instruments, the expensive GPU computation it's doing takes around 2x as long to complete, meaning that the app performs half as well. I've done a ton of work to micro-optimize my Metal compute code, based on every scrap of information from the WWDC videos, etc. A problem I'm running into is that I think that the more efficient I make my code, the less it signals to the OS that I want high GPU clock speeds! I think part of why the OS is confused is that in most use cases, my computation can be done using only a small number of Metal threadgroups. I'm guessing that the OS heuristics see that only a small fraction of the GPU is saturated and fail to scale up the power/clock state. I'm not sure what to do here; I'm in a bit of a bind. One possibility is that I intentionally schedule busy work -- spin threadgroups just to waste energy and signal to the OS that I need higher clock speeds. This is obviously a really bad idea, but it might work. Is there any other (better) way for my app to signal to the OS that it is doing real-time latency-sensitive computation on the GPU and needs the clock speeds to be scaled up? Note that game mode is not really an option, as my app also runs as an AU plugin inside hosts like Garageband, so it can't be made fullscreen, etc.

I made a box with MDLMesh.newBox(). I added normals. let mdlMesh = MDLMesh.newBox(withDimensions: SIMD3<Float>(1, 1, 1), segments: SIMD3<UInt32>(2, 2, 2), geometryType: MDLGeometryType.triangles, inwardNormals:false, allocator: allocator) mdlMesh.addNormals(withAttributeNamed: MDLVertexAttributeNormal, creaseThreshold: 0.25) After I convert to MTKMesh the normals are (0,0,0) for a group of vertices. I can only inspect the geometry after I convert to MTKMesh. Is there a way you can use Geometry Viewer on a MDLMesh?

hi When analyzing our game using Instruments, I've always been confused about the two items "Drawable Present" and "Drawable Presented" in the GPU column. The timing of Drawable Present seems to be when the CPU layer calls commandbuffer:present, rather than when the actual encoding is completed on the GPU. Also, what does drawable presented specifically mean? In our case, when a CPU stall occurs, it appears that the vsync interval changes in the next frame, and a surface that has already been calculated is not displayed. Why is this happening?

I am new to Xcode and trying to learn how to use Metal for my internship. I am trying to link the binaries of Foundation.framework, Metal.framework, and Quartcore.framework. But whenever I try to build it always fails to find any of them. I have my Header Search Path as $(PROJECT_DIR)/metal-cpp, I tried adding some for the Frameworks but that did not work either. I do have the binaries linked in the Build Phases, so I don't know what else I could be missing.

In the Creating A 3D Application With Hydra Rendering tutorial on the Apple Developer website, on the last step where I execute this command: cmake -S ~/Users/macuser/CreatingA3DApplicationWithHydraRendering/ -B ~/Users/macuser/CreatingA3DApplicationWithHydraRendering/ I keep getting an error: CMake Error at CMakeLists.txt:5 (include): include could not find requested file: /Users/macuser/USDInstall/bin/pxrConfig.cmake I've tried to follow the instructions as mentioned in the README.md file included in the project files at least 5 times as well as moving the pxrConfig.cmake file around and copying it in different folders, then executed the command and was still unsuccessful into generating the proper file expected to compile and render the HydraPlayer renderer. How do I get cmake to generate the Xcode file to create the HydraPlayer renderer?

I’m trying to follow Apple’s “WWDC24: Bring your machine learning and AI models to Apple Silicon” session to convert the Mistral-7B-Instruct-v0.2 model into a Core ML package, but I’ve run into a roadblock that I can’t seem to overcome. I’ve uploaded my full conversion script here for reference: https://pastebin.com/T7Zchzfc When I run the script, it progresses through tracing and MIL conversion but then fails at the backend_mlprogram stage with this error: https://pastebin.com/fUdEzzKM The core of the error is: ValueError: Op "keyCache_tmp" (op_type: identity) Input x="keyCache" expects list, tensor, or scalar but got state[tensor[1,32,8,2048,128,fp16]] I’ve registered my KV-cache buffers in a StatefulMistralWrapper subclass of nn.Module, matching the keyCache and valueCache state names in my ct.StateType definitions, but Core ML’s backend pass reports the state tensor as an invalid input. I’m using Core ML Tools 8.3.0 on Python 3.9.6, targeting iOS18, and forcing CPU conversion (MPS wasn’t available). Any pointers on how to satisfy the handle_unused_inputs pass or properly declare/cache state for GQA models in Core ML would be greatly appreciated! Thanks in advance for your help, Usman Khan