Fix barrier setting in async compute sample

[yknishidate]

Description

In the async compute sample, when "Enable async queue" is checked in the GUI, the following validation error is output.

Validation Error: [ UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout ]
Object 0: handle = 0x18b310c3570, type = VK_OBJECT_TYPE_COMMAND_BUFFER;
Object 1: handle = 0x1ff0c40000000a81, type = > VK_OBJECT_TYPE_IMAGE; |
MessageID = 0x4dae5635 |
vkQueueSubmit(): pSubmits[0].pCommandBuffers[0] command buffer VkCommandBuffer 0x18b310c3570[] expects VkImage 0x1ff0c40000000a81[] (subresource: aspectMask 0x1 array layer 0, mip level 0) to be in layout VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL--instead, current layout is VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL.

Note

If you get output different from the one below when you check the checkbox, the error may not be reproducible in your environment.

[info] Device has 2 or more graphics queues.
[info] Device has async compute queue.

Since the sample contains multiple images and command buffers, I added set_debug_name() to make the error more understandable, as shown below.

Validation Error: [ UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout ]
Object 0: handle = 0x24dda4ae730, name = compute_post, type = VK_OBJECT_TYPE_COMMAND_BUFFER;
Object 1: handle = 0xdd5f6f0000000a82, name = color_targets[0], type = VK_OBJECT_TYPE_IMAGE; |
MessageID = 0x4dae5635 |
vkQueueSubmit(): pSubmits[0].pCommandBuffers[0] command buffer VkCommandBuffer 0x24dda4ae730[compute_post] expects VkImage 0xdd5f6f0000000a82[color_targets[0]] (subresource: aspectMask 0x1 array layer 0, mip level 0) to be in layout VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL--instead, current layout is VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL.

This shows that the barrier setting for color_targets[0] in compute_post is incorrect. The code is as follows:

Vulkan-Samples/samples/performance/async_compute/async_compute.cpp

Lines 526 to 527 in f7e97a1

	memory_barrier.old_layout = VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL;
	memory_barrier.new_layout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL;

In the previous forward_offscreen_pass, the layout already transitioned to SHADER_READ_ONLY_OPTIMAL after drawing. This means that the current code is performing the same image layout transition twice (However, both barriers are required for queue transitions.).

Vulkan-Samples/samples/performance/async_compute/async_compute.cpp

Lines 409 to 410 in f7e97a1

	memory_barrier.old_layout = VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL;
	memory_barrier.new_layout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL;

So I fixed the old_layout setting on the compute_post barrier, and the validation errors disappeared. Also, because this sample is complex and hard to debug, I'm leaving set_debug_name() for future.

• Tested on Windows
• No related issues

General Checklist:

Please ensure the following points are checked:

• My code follows the coding style
• I have reviewed file licenses
• I have commented any added functions (in line with Doxygen)
• I have commented any code that could be hard to understand
• My changes do not add any new compiler warnings
• My changes do not add any new validation layer errors or warnings
• I have used existing framework/helper functions where possible
• My changes do not add any regressions
• I have tested every sample to ensure everything runs correctly
• This PR describes the scope and expected impact of the changes I am making

Note: The Samples CI runs a number of checks including:

• I have updated the header Copyright to reflect the current year (CI build will fail if Copyright is out of date)
• My changes build on Windows, Linux, macOS and Android. Otherwise I have documented any exceptions

[asuessenbach]

Great catch!

And even better fix description!