Always use MTLResourceStorageModeShared for Metal argument buffers
to ensure writes to argument buffers propagate on discrete GPUs.
diff --git a/MoltenVK/MoltenVK/GPUObjects/MVKDescriptorSet.mm b/MoltenVK/MoltenVK/GPUObjects/MVKDescriptorSet.mm
index a2e9d51..632c2b5 100644
--- a/MoltenVK/MoltenVK/GPUObjects/MVKDescriptorSet.mm
+++ b/MoltenVK/MoltenVK/GPUObjects/MVKDescriptorSet.mm
@@ -816,15 +816,7 @@
setConfigurationResult(reportError(VK_ERROR_FRAGMENTATION_EXT, "vkCreateDescriptorPool(): The requested descriptor storage of %d MB is larger than the maximum descriptor storage of %d MB per VkDescriptorPool.", (uint32_t)(metalArgBuffSize / MEBI), (uint32_t)(maxMTLBuffSize / MEBI)));
metalArgBuffSize = maxMTLBuffSize;
}
-
- // The MTLBuffer can have Managed storage if possible because we don't set constant inline data directly in the buffer.
-#if MVK_MACOS
- MTLResourceOptions rezOpts = MTLResourceStorageModeManaged;
-#endif
-#if MVK_IOS_OR_TVOS
- MTLResourceOptions rezOpts = MTLResourceStorageModeShared;
-#endif
- _metalArgumentBuffer = [getMTLDevice() newBufferWithLength: metalArgBuffSize options: rezOpts]; // retained
+ _metalArgumentBuffer = [getMTLDevice() newBufferWithLength: metalArgBuffSize options: MTLResourceStorageModeShared]; // retained
_metalArgumentBuffer.label = @"Argument buffer";
}
}
