If I compile a compute kernel with a call to texture.read(), it fails with the following error: "Error Domain=AGXMetalG13X Code=3 "Encountered unlowered function call to air.get_read_sampler" UserInfo={NSLocalizedDescription=Encountered unlowered function call to air.get_read_sampler}." This error occurs on both macOS and iOS 26 Beta 5, but not when running on a simulator or in a playground. It does not occur on a macOS Sequoia VM. It occurs whether I use the old metal 3 or new metal 4 compilation method. A workaround would be to use a sampler, but according to the feature tables, all platforms support reading from textures of all formats. Below is a minimal example which produces the error: let device = MTLCreateSystemDefaultDevice()! let library = device.makeDefaultLibrary()! let computeFunction = library.makeFunction(name: "compute_test")! do { let pipeline = try device.makeComputePipelineState(function: computeFunction) debugPrint(pipeline) } catch { debugPrint("Metal 3 failed with error:\n\(error)") } #import <metal_stdlib> using namespace metal; kernel void compute_test(uint2 gid [[thread_position_in_grid]], texture2d<float, access::read> in [[texture(0)]], texture2d<float, access::write> out [[texture(1)]]) { out.write(in.read(gid), gid); } I filed feedback FB19530049.

Hi everyone, This project uses PyTorch on an Apple Silicon Mac (M1/M2/etc.), and the goal is to use the MPS backend for GPU acceleration, notes Apple Developer. However, the workflow depends on Float64 (double-precision) floating-point numbers for certain computations, notes PyTorch Forums. The error "Cannot convert a MPS Tensor to float64 dtype as the MPS framework doesn't support float64. Please use float32 instead" has been encountered, notes GitHub. It seems that the MPS backend doesn't currently support Float64 for direct GPU computation. Questions for the community: Are there any known workarounds or best practices for handling Float64-dependent operations when using the MPS backend with PyTorch? For those working with high-precision tasks on Apple Silicon, what strategies are being used to balance performance with the need for Float64? Offloading to the CPU is an option, and it's of interest to know if there are any specific techniques or libraries within the Apple ecosystem that could streamline this process while aiming for optimal performance. Any insights, tips, or experiences would be appreciated. Thanks in advance, Jonaid MacBook Pro M3 Max

We set the CVDisplayLink on macOS to 0 or 120, and get the following. This then clamps maximum refresh to 60Hz on the 120Hz ProMotion display on a MBP M2 Max laptop. How is this not fixed in 4 macOS releases? CoreVideo: currentVBLDelta returned 200000 for display 1 -- ignoring unreasonable value CoreVideo: [0x7fe2fb816020] Bad CurrentVBLDelta for display 1 is zero. defaulting to 60Hz.

Can't seem to get the Metal HUD to display value range's (pre 26 Tahoe). The documented environment variable MTL_HUD_SHOW_VALUE_RANGE doesn't seem to work. https://developer.apple.com/documentation/xcode/monitoring-your-metal-apps-graphics-performance#Display-the-value-range-of-metrics Anyone having any luck?

Hello, Shaders in our application is written using HLSL and we rely on Metal Shader Converter to convert DXIL to Metal IR. We ran into an issue that causes metal pipeline state creation to fail when vertex stage-in function is used on AMD GPUs. Here's the error reported by Metal in Xcode output: Compiler failed with XPC_ERROR_CONNECTION_INTERRUPTED XPC_ERROR_CONNECTION_INTERRUPTED MTLCompiler: Compilation failed with XPC_ERROR_CONNECTION_INTERRUPTED on 4 try. This error suggests an unexpected interruption in the connection. Possible reasons: a crash in the compiler service, termination by the OS due to resource constraints (e.g., jetsam), a timeout in the service, or an issue with IPC. Verify system stability and check the logs for more details. Compiler failed with XPC_ERROR_CONNECTION_INVALID XPC_ERROR_CONNECTION_INVALID MTLCompiler: Compiler encountered XPC_ERROR_CONNECTION_INVALID: failed to check-in, peer may have been unloaded: mach_error=10000003 (is the OS shutting down or process jetsammed?) Compilation failed due to an interrupted connection: XPC_ERROR_CONNECTION_INTERRUPTED. This error occurred after multiple retries. which seems to indicate a internal compiler error. I have a minimal repro here: https://github.com/kcloudy0717/metal_pso_fail/tree/main, simply follow the instructions in README.

I rewrote my graphics pipeline to use Load/Store better for clearing and don't care cases. All my tests pass, and in the Metal debugger, all the draw calls succeed. But when I present drawables (before [commandBuffer commit]) I only get a pink screen. I've tried everything I can think of: making sure the pixel formats are the same for the back buffer as my render targets, etc. But it's still pink. Could you point me in the right direction so I can fix this, or help describe why it's pink. That would be really helpful. Thank you, Brian Hapgood

Hi, I’m using the latest iPad Pro (13-inch) and I can see that Metal offers an rgb10a2unorm texture for rendering, but when I render a grey ramp and measure the actual luminance, I get a pattern that I would expect from an 8-bit texture (see below). Before I start ripping apart all my code, is there anything else I need to do to convince iOS to render my texture in 10-bit? I already tried setting the PixelFormat in my CMetalLayer to rgb10a2unorm, but that didn’t change anything.

Hello, This exact question was already asked in this forum (8 years ago) but I can't find a definitive answer: Does Metal allow using the same color texture as both an input and output (color attachment) of a fragment shader? Is the behavior defined somewhere? I believe this results in undefined behavior under both DirectX and OpenGL, so I'd assume the same for Metal, but then why doesn't Metal warn me about this as it does on some many other "misconfigurations"? It also seems to work correctly in my case, as I found out by accident. Would love to get a clarification! Thanks ahead!

It’s great that we’ll be able to use Metal custom renderers in passthrough mode on visionOS. https://developer.apple.com/wwdc24/10092 This is a lot of complicated set-up, however. It’s also unclear how occlusion and custom algorithms / raytracing will work in tandem with scene understanding. May we have a project template and/or sample? Preferably with the C api and not just swift. This would be much-appreciated and helpful to everyone who wants this set-up. I’d like to see the whole process. Thank you for introducing this feature!

The title is self-exploratory. I wasn't able to find the CAMetalDisplayLink on the most recent metal-cpp release (metal-cpp_macOS15_iOS18-beta). Are there any plans to include it in the next release?

When generating large arrays of random numbers, NaNs show up. They also show up at the same indices when using the same seed, leading me to believe that this is a bug with MPSMatrixRandom's normally distributed Float32 random number distribution. Happens with both Philox and MTGP32. Is this intentional and how do I work around this? See the original post for a MWE in Swift and Julia: https://github.com/JuliaGPU/Metal.jl/issues/474

I'm trying to use MTLBinaryArchive. I collected a BinaryArchive from one device and used metal-tt to translate it for all supported iPhone devices, ranging from iPhone 7 Plus to iPhone 16. However, this BinaryArchive is quite large, around 1.5GB uncompressed, and about 500MB compressed in the IPA. I'm wondering how to address the size issue. I watched the WWDC 2022 video, which mentioned that the operating system or app installation process would handle compatibility. Does this compatibility support different GPU chips? I tried installing an IPA with a BinaryArchive collected only from an iPhone 12 on an iPhone 13, but the BinaryArchive didn't take effect. I also saw that Apple supports App Thinning. However, it seems that resources in the Asset Catalog cannot be accessed via URL, and creating an MTLBinaryArchive requires a URL. Is it possible for MTLBinaryArchive to be distributed through App Thinning? The WWDC 2022 video also mentioned using the -Os optimization flag to reduce size. Can this give an estimate of how much compression it would achieve? Are there any methods to solve the BinaryArchive size issue without impacting performance?

Hi, I am working with a large project. We are compiling each material to its own .metallib. They all include many common files full of inline functions. Finally we link it all together at the end with a single big pathtrace kernel. Everything works as expected, however the compile times have gotten completely out of hand and it takes multiple minutes to compile at runtime (to native code). I have gathered that I can do this offline by using metal-tt however if I am wondering if there is a way to reduce the compile times in such a scenario, and how to investigate what the root cause of the problem is. I suspect it could have to do with the fact that every materials metallib contains duplications of all the inline functions. Any ideas on how to profile and debug this? Thanks, Rasmus

Just wondering if anyone knows what it will take to hit greater than 60hz when targeting iPhone. If I set the preferredFramesPerSecond of an MTKView to 120, it works on the iPad, but on iPhone it never goes over 60hz, even with a simple hello triangle sample app... is this a limitation of targeting iPhone?

I am trying to learn the new Metal Peformance Primitives APIs. I have added the MetalPeformancePrimitives framework and included the header in my shader code as per documentation #include <MetalPeformancePrimitives/MetalPeformancePrimitives.h> Unfortunately, Xcode complains that the header cannot be found. How do I include it properly? I am using Xcode 26 on Tahoe. The MetalPeformancePrimitives framework is present on my machine and I can inspect the headers in the filesystem.

I'm updating our app to support metal 4, but the metal 4 types don't seem to get recognized when targeting simulator. Is it known if metal 4 will be supported in the near future, or am I setting up the app wrong?

Hi, when I run my app, in console say: NSBundle file:///System/Library/PrivateFrameworks/MetalTools.framework/ principal class is nil because all fallbacks have failed

Ever since the release of iOS18, we've been seeing a new crash related to calling jpegData(compressionQuality:). From reports, this isn't crashing during foreground usage of the app, but instead will prompt the user about a background app crash upon foregrounding. The stacks from crash reports show this crash is happening from a variety of callers, but all go through jpegData(compressionQuality:), down through [HDRImageConverter_Metal init] and end up in apthread_mutex_local call when it crashes. Attached is a sample crash report from 18.2(22C5125e), but we've been seeing this since the first iOS18 release. Did something change with around these calls in iOS18 that prohibits their use in the background? crash.txt

I have a very basic usdz file from this repo I call loadTextures() after loading the usdz via MDLAsset. Inspecting the MDLTexture object I can tell it is assigning a colorspace of linear rgb instead of srgb although the image file in the usdz is srgb. This causes the textures to ultimately render as over saturated. In the code I later convert the MDLTexture to MTLTexture via MTKTextureLoader but if I set the srgb option it seems to ignore it. This significantly impacts the usefulness of Model I/O if it can't load a simple usdz texture correctly. Am I missing something? Thanks!

So I've been trying out GPTK with Elite Dangerous Horizons game and it looks like from what I can tell. The VRAM keeps going up until it goes over the limit where it drops the FPS to 1-3 FPS and then crashes the game. From the Performance HUD I can see that it looks like when using GPTK, the VRAM usage just keeps climbing and I never saw it drop down at all. I did some limited testing, and from that I think I can conclude that it is probably not a VRAM leak, but it might be caching it. The reason for this is because I noticed that if I went back to the area that I've been before. It won't increase the VRAM usage. So either there is something wrong with the freeing VRAM memory part, or it could be that GPTK might not be reporting the right amount of VRAM available to use? So maybe that's why it keeps allocating VRAM until it went out of memory and crashed the game. Just to test, I did try running the game with DXVK+MoltenVK combo, and I can see that it works just fine. VRAM is being freed up when it's no longer used. Is this a known issue in some games?