Vanilla Vulkan Compute Shader 未写入 output 缓冲区

Question

EDIT: fix double unmapping (but doesn't fix the issue)
EDIT2: fix API version and remove validation layer from code. Instead, run with VK_INSTANCE_LAYERS=VK_LAYER_KHRONOS_validation env. Issue is still present
EDIT3: forgot descriptors set, which allow binding buffers to shader input. But still doesn't fix the issue:'(

<TL;DR> I've written a simple Vulkan compute-only sample code with a basic compute shader. No Vulkan nor shader errors but output buffer is not written by compute shader:(

Trying to learn Vulkan API, I've started writing a simple compute-only sample with a basic compute shader. It uploads a buffer of int to GPU, run a compute shader that increment each int and write results in a second buffer.

My problem is that everything runs fine but I don't get expected results in my output buffer and I can't figure out why. It looks like the compute shader is dispatched but output buffer is never written.

To observe that, I first upload random numbers to my input buffer and fill my output buffer with value of 2. Then dispatch compute shader which is supposed to read each value X from input, and write X+1 into the output buffer.
After waiting for completion, I map my output buffer and display its data. I only got 2's:'(

NB: memory bound to buffers is created with VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BITVK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT .

So there's definitely a concept in Vulkan that I got wrong, or a subtlety in the flags/setup that I can't see...

The compute shader code:

#version 450 core

layout (set = 0, binding = 0) buffer InputBuffer {
    uvec4 inputData[25];
};

layout (set = 0, binding = 1) buffer OutputBuffer {
    uvec4 outputData[25];
};

layout (local_size_x = 8, local_size_y = 1, local_size_z = 1) in;
void main()
{
    uint gid = gl_GlobalInvocationID.x;
    if(gid < 25)
        outputData[gid] = inputData[gid] + uvec4(1,1,1,1);
}

And the whole sample code (as I don't know where I could be wrong, I've pasted the whole stuff, sorry):

#include <vulkan/vulkan.h>
#include <iostream>
#include <vector>
#include <assert.h>
#include <fstream>

// Some helper functions
typedef uint32_t            u32;
typedef uint64_t            u64;

// Vulkan two steps enumeration function
#define COUNT_AND_GET1(func, vec, arg1) {\
    u32 size = 0; \
    ##vec.clear(); \
    ##func(##arg1, &size, nullptr); \
    if(size > 0) { \
    ##vec.resize(size); \
    ##func(##arg1, &size, ##vec.data()); }\
}

#define COUNT_AND_GET2(func, vec, arg1, arg2) {\
    u32 size = 0; \
    ##vec.clear(); \
    ##func(##arg1, ##arg2, &size, nullptr); \
    if(size > 0) { \
    ##vec.resize(size); \
    ##func(##arg1, ##arg2, &size, ##vec.data()); }\
}

// Basic vec4 data
struct vec4
{
    u32 x; u32 y; u32 z; u32 w;
};

struct PhysicalDeviceProps
{
    VkPhysicalDeviceProperties              m_Properties;
    VkPhysicalDeviceFeatures                m_Features;
    VkPhysicalDeviceMemoryProperties        m_MemoryProperties;
    std::vector<VkQueueFamilyProperties>    m_QueueFamilyProperties;
    std::vector<VkLayerProperties>          m_LayerProperties;
    std::vector<VkExtensionProperties>      m_ExtensionProperties;
};

// Return device memory index that matches specified properties
u32 SelectMemoryHeapFrom(u32 memoryTypeBits, const VkPhysicalDeviceMemoryProperties& memoryProperties, VkMemoryPropertyFlags preferredProperties, VkMemoryPropertyFlags requiredProperties)
{
    assert((preferredProperties & requiredProperties) > 0);
    u32 selectedType = u32(-1);
    u32 memIndex = 0;
    while (memIndex < VK_MAX_MEMORY_TYPES && selectedType == u32(-1))
    {
        if (((memoryTypeBits & (1 << memIndex)) > 0)
            && ((memoryProperties.memoryTypes[memIndex].propertyFlags & preferredProperties) == preferredProperties))
        {
            // If it exactly matches my preferred properties, grab it.
            selectedType = memIndex;
        }
        ++memIndex;
    }

    if (selectedType == u32(-1))
    {
        memIndex = 0;
        while (memIndex < VK_MAX_MEMORY_TYPES && selectedType == u32(-1))
        {
            if (((memoryTypeBits & (1 << memIndex)) > 0)
                && ((memoryProperties.memoryTypes[memIndex].propertyFlags & requiredProperties) == requiredProperties))
            {
                // If it exactly matches my required properties, grab it.
                selectedType = memIndex;
            }
            ++memIndex;
        }
    }
    return selectedType;
}

// **** MAIN FUNCTION ****
void SampleCompute()
{
    // -------------------------------------
    // 1. Create Instance
    // -------------------------------------
    VkApplicationInfo appInfo = { VK_STRUCTURE_TYPE_APPLICATION_INFO, nullptr, "SampleCompute", 0, "MyEngine", 0, VK_API_VERSION_1_2 };
    VkInstanceCreateInfo instCreateInfo = { VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO, nullptr, 0, &appInfo, 0, nullptr, 0, nullptr };
    VkInstance instance = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateInstance(&instCreateInfo, nullptr, &instance))
        std::cout << "Instance creation failed!\n";


    // ---------------------------------------------------
    // 2. Enumerate physical devices and select 'best' one 
    // ---------------------------------------------------
    VkPhysicalDevice bestDevice = VK_NULL_HANDLE;
    PhysicalDeviceProps bestDeviceProps;
    {
        std::vector<VkPhysicalDevice> physicalDevices;
        COUNT_AND_GET1(vkEnumeratePhysicalDevices, physicalDevices, instance)
        assert(!physicalDevices.empty());

        std::vector< PhysicalDeviceProps> physicalDeviceProps(physicalDevices.size());
        for (u64 i = 0; i < physicalDevices.size(); ++i)
        {
            vkGetPhysicalDeviceProperties(physicalDevices[i], &physicalDeviceProps[i].m_Properties);
            vkGetPhysicalDeviceMemoryProperties(physicalDevices[i], &physicalDeviceProps[i].m_MemoryProperties);
            COUNT_AND_GET1(vkGetPhysicalDeviceQueueFamilyProperties, physicalDeviceProps[i].m_QueueFamilyProperties, physicalDevices[i])
            COUNT_AND_GET1(vkEnumerateDeviceLayerProperties, physicalDeviceProps[i].m_LayerProperties, physicalDevices[i])
            COUNT_AND_GET2(vkEnumerateDeviceExtensionProperties, physicalDeviceProps[i].m_ExtensionProperties, physicalDevices[i], nullptr)
        }

        u64 bestDeviceIndex = 0;
        for (u64 i = 1; i < physicalDevices.size(); ++i)
        {
            const bool isDiscrete = physicalDeviceProps[bestDeviceIndex].m_Properties.deviceType == VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU;
            const bool otherIsDiscrete = physicalDeviceProps[i].m_Properties.deviceType == VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU;
            if (isDiscrete && !otherIsDiscrete)
                continue;
            else if ((!isDiscrete && otherIsDiscrete)
                || (physicalDeviceProps[bestDeviceIndex].m_Properties.limits.maxFramebufferWidth < physicalDeviceProps[i].m_Properties.limits.maxFramebufferWidth))
                bestDeviceIndex = i;
        }

        bestDevice = physicalDevices[bestDeviceIndex];
        bestDeviceProps = physicalDeviceProps[bestDeviceIndex];
    }


    // ---------------------------------------------------
    // 3. Find queue family which support compute pipeline
    // ---------------------------------------------------
    u32 computeQueue = 0;
    while (computeQueue < bestDeviceProps.m_QueueFamilyProperties.size()
        && ((bestDeviceProps.m_QueueFamilyProperties[computeQueue].queueFlags & VK_QUEUE_COMPUTE_BIT) != VK_QUEUE_COMPUTE_BIT))
    {
        ++computeQueue;
    }
    assert(computeQueue < bestDeviceProps.m_QueueFamilyProperties.size());


    // -------------------------------
    // 4. Create logical device
    // -------------------------------
    VkDeviceQueueCreateInfo queueInfo = { VK_STRUCTURE_TYPE_DEVICE_QUEUE_CREATE_INFO, nullptr, 0, computeQueue, 1, nullptr };
    VkPhysicalDeviceFeatures features = {};
    VkDeviceCreateInfo createInfo = {
        VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO, nullptr, 0,
        1, &queueInfo,
        0, nullptr,
        0, nullptr,
        &features
    };
    VkDevice device = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateDevice(bestDevice, &createInfo, nullptr, &device))
        std::cout << "Logical Device creation failed\n";


    // -------------------------------
    // 5. Create data buffers
    // -------------------------------
    constexpr u64 elemCount = 25;
    constexpr u64 bufferSize = elemCount * sizeof(vec4);
    VkBufferCreateInfo bufferCreateInfo = {
            VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, nullptr, 0,
            bufferSize,
            VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT,
            VK_SHARING_MODE_EXCLUSIVE, 0, nullptr
    };

    VkBuffer inputBuffer = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateBuffer(device, &bufferCreateInfo, nullptr, &inputBuffer))
        std::cout << "Creating input buffer failed!\n";
    VkMemoryRequirements inputBufferMemory;
    vkGetBufferMemoryRequirements(device, inputBuffer, &inputBufferMemory);

    VkBuffer outputBuffer = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateBuffer(device, &bufferCreateInfo, nullptr, &outputBuffer))
        std::cout << "Creating output buffer failed!\n";
    VkMemoryRequirements outputBufferMemory;
    vkGetBufferMemoryRequirements(device, outputBuffer, &outputBufferMemory);


    // -------------------------------
    // 6. Allocate memory for buffers
    // -------------------------------
    u32 inputMemoryIndex = SelectMemoryHeapFrom(inputBufferMemory.memoryTypeBits, bestDeviceProps.m_MemoryProperties, 
        VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT, 
        VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT);
    VkMemoryAllocateInfo inputAllocationInfo = { VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO, nullptr, inputBufferMemory.size, inputMemoryIndex };
    VkDeviceMemory inputMemory = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkAllocateMemory(device, &inputAllocationInfo, nullptr, &inputMemory))
        std::cout << "Memory allocation of " << inputBufferMemory.size << " failed!\n";

    u32 outputMemoryIndex = SelectMemoryHeapFrom(outputBufferMemory.memoryTypeBits, bestDeviceProps.m_MemoryProperties, 
        VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT, 
        VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT);
    VkMemoryAllocateInfo outputAllocationInfo = { VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO, nullptr, outputBufferMemory.size, outputMemoryIndex };
    VkDeviceMemory outputMemory = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkAllocateMemory(device, &outputAllocationInfo, nullptr, &outputMemory))
        std::cout << "Memory allocation of " << outputBufferMemory.size << " failed!\n";


    // -------------------------------
    // 7. Bind buffers to memory
    // -------------------------------
    if (vkBindBufferMemory(device, inputBuffer, inputMemory, 0) != VK_SUCCESS)
        std::cout << "Input buffer binding failed!\n";

    if (vkBindBufferMemory(device, outputBuffer, outputMemory, 0) != VK_SUCCESS)
        std::cout << "Output buffer binding failed!\n";


    // ----------------------------------
    // 8. Map buffers and upload data
    // ----------------------------------
    vec4* inputData = nullptr;
    if (VK_SUCCESS != vkMapMemory(device, inputMemory, 0, VK_WHOLE_SIZE, 0, (void**)(&inputData)))
        std::cout << "Input memory mapping failed!\n";
    
    for (u32 i = 0; i < elemCount; ++i)
    {
        inputData[i].x = static_cast<u32>(rand() / (float)RAND_MAX * 100);
        inputData[i].y = static_cast<u32>(rand() / (float)RAND_MAX * 100);
        inputData[i].z = static_cast<u32>(rand() / (float)RAND_MAX * 100);
        inputData[i].w = static_cast<u32>(rand() / (float)RAND_MAX * 100);
        std::cout << inputData[i].x << ", " << inputData[i].y << ", " << inputData[i].z << ", " << inputData[i].w << ", ";
    }
    std::cout << "\n\n\n";
    vkUnmapMemory(device, inputMemory);

    vec4* initialOutputData = nullptr;
    if (VK_SUCCESS != vkMapMemory(device, outputMemory, 0, VK_WHOLE_SIZE, 0, (void**)(&initialOutputData)))
        std::cout << "Output memory mapping failed!\n";
    for (u32 i = 0; i < elemCount; ++i)
    {
        initialOutputData[i].x = 2; initialOutputData[i].z = 2; initialOutputData[i].y = 2; initialOutputData[i].w = 2;
    }
    vkUnmapMemory(device, outputMemory);


    // ----------------------------------
    // 9. Create shader/pipeline layout
    // ----------------------------------
    std::vector<VkDescriptorSetLayoutBinding> bindings = {
            { 0, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 1, VK_SHADER_STAGE_COMPUTE_BIT, nullptr },
            { 1, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 1, VK_SHADER_STAGE_COMPUTE_BIT, nullptr }
    };
    VkDescriptorSetLayoutCreateInfo layoutInfo = { VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, nullptr, 0, 2, bindings.data() };
    VkDescriptorSetLayout descriptorLayout = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateDescriptorSetLayout(device, &layoutInfo, nullptr, &descriptorLayout))
        std::cout << "Descriptor Layout creation failed!\n";

    // Create pipeline layout
    VkPipelineLayoutCreateInfo pipelineCreateInfo = { VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, nullptr, 0, 1, &descriptorLayout, 0, nullptr };
    VkPipelineLayout layout = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreatePipelineLayout(device, &pipelineCreateInfo, nullptr, &layout))
        std::cout << "Pipeline Layout creation failed\n";


    // --------------------------------------------------
    // 10. Load shader source and create shader module
    // --------------------------------------------------
    std::ifstream file("ComputeShader.spv", std::ifstream::binary);
    u64 size = 0;
    if (!file.is_open())
        std::cout << "Can't open shader!\n";
    
    file.seekg(0, file.end);
    size = file.tellg();
    file.seekg(0);
    char* shaderSrc = new char[size];
    file.read(shaderSrc, size);

    VkShaderModuleCreateInfo shaderCreateInfo = { VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, nullptr, 0, size, reinterpret_cast<u32*>(shaderSrc) };
    VkShaderModule shader = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateShaderModule(device, &shaderCreateInfo, nullptr, &shader))
        std::cout << "Shader Module creation failed\n";
    delete[] shaderSrc;
    

    // ----------------------------------
    // 10.5. Create descriptor sets
    // ----------------------------------
    VkDescriptorPoolSize descriptorPoolSize = { VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 2 };
    VkDescriptorPoolCreateInfo descriptorPoolCreateInfo = {
          VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO, nullptr, 0,
          1, 1, &descriptorPoolSize };
    VkDescriptorPool descriptorPool = VK_NULL_HANDLE;
    vkCreateDescriptorPool(device, &descriptorPoolCreateInfo, 0, &descriptorPool);

    VkDescriptorSetAllocateInfo descriptorSetAllocateInfo = {
          VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO, 0,
          descriptorPool, 1, &descriptorLayout
    };
    VkDescriptorSet descriptorSet;
    vkAllocateDescriptorSets(device, &descriptorSetAllocateInfo, &descriptorSet);

    VkDescriptorBufferInfo inputBufferDescriptorInfo = { inputBuffer, 0, VK_WHOLE_SIZE };
    VkDescriptorBufferInfo outputBufferDescriptorInfo = { outputBuffer, 0, VK_WHOLE_SIZE };
    VkWriteDescriptorSet writeDescriptorSet[2] = {
          {
            VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, 0, descriptorSet,
            0, 0, 1,
            VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
            0, &inputBufferDescriptorInfo, 0
          },
          {
            VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, 0, descriptorSet, 
            1, 0, 1,
            VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
            0, &outputBufferDescriptorInfo, 0
          }
    };

    vkUpdateDescriptorSets(device, 2, writeDescriptorSet, 0, nullptr);
    
    // -------------------------------
    // 11. Create compute pipeline
    // -------------------------------
    const char* entryPointName = "main";
    VkComputePipelineCreateInfo computeCreateInfo = {
            VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO, nullptr, 0,
            {
                VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO, nullptr, 0,
                VK_SHADER_STAGE_COMPUTE_BIT, shader,
                entryPointName, nullptr
            },
            layout, VK_NULL_HANDLE, 0
    };

    VkPipeline pipeline = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateComputePipelines(device, VK_NULL_HANDLE, 1, &computeCreateInfo, nullptr, &pipeline))
        std::cout << "Compute Pipeline creation failed!\n";


    // ------------------------------------------------
    // 12. Create Command Pool and Command Buffer
    // --------------------------------------------------
    VkCommandPoolCreateInfo poolInfo = { VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO, nullptr, 0, computeQueue };
    VkCommandPool cmdPool = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateCommandPool(device, &poolInfo, nullptr, &cmdPool))
        std::cout << "Command Pool creation failed!\n";

    VkCommandBufferAllocateInfo cmdBufferInfo = {
            VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO, nullptr,
            cmdPool, VK_COMMAND_BUFFER_LEVEL_PRIMARY, 1
    };
    VkCommandBuffer cmdBuffer = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkAllocateCommandBuffers(device, &cmdBufferInfo, &cmdBuffer))
        std::cout << "Command buffer allocation failed!\n";

    // ---------------------------
    // 13. Run compute shader
    // ---------------------------
    VkCommandBufferUsageFlags flags = VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT;
    VkCommandBufferBeginInfo beginInfo = { VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO, nullptr, VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT, nullptr };
    vkBeginCommandBuffer(cmdBuffer, &beginInfo);
    vkCmdBindDescriptorSets(cmdBuffer, VK_PIPELINE_BIND_POINT_COMPUTE, layout, 0, 1, &descriptorSet, 0, 0);
    vkCmdBindPipeline(cmdBuffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipeline);
    vkCmdDispatch(cmdBuffer, 8, 1, 1);
    
    vkEndCommandBuffer(cmdBuffer);

    // -----------------------------------------
    // 14. Submit command buffer (with fence)
    // -----------------------------------------
    VkFenceCreateInfo fenceCreateInfo = { VK_STRUCTURE_TYPE_FENCE_CREATE_INFO, nullptr, (VkFenceCreateFlags)0 };
    VkFence fence = VK_NULL_HANDLE;
    if (VK_SUCCESS != vkCreateFence(device, &fenceCreateInfo, nullptr, &fence))
        std::cout << "Fence creation failed!\n";

    VkQueue queue = VK_NULL_HANDLE;
    vkGetDeviceQueue(device, computeQueue, 0, &queue);

    VkSubmitInfo submitInfo = { 
        VK_STRUCTURE_TYPE_SUBMIT_INFO, nullptr, 0, nullptr, 0,
        1, &cmdBuffer, 0, nullptr
    };
    VkResult result = vkQueueSubmit(queue, 1, &submitInfo, fence);

    // Wait for everything finished
    if (result == VK_SUCCESS)
    {
        result = vkQueueWaitIdle(queue);
    }
    vkWaitForFences(device, 1, &fence, VK_TRUE, u64(-1));

    // ---------------------------------
    // 15. Grab and display results
    // ---------------------------------
    vec4* resultData = nullptr;
    if (VK_SUCCESS != vkMapMemory(device, outputMemory, 0, VK_WHOLE_SIZE, 0, (void**)(&resultData)))
        std::cout << "Output memory mapping failed!\n";
    for (u32 i = 0; i < elemCount; ++i)
    {
        std::cout << resultData[i].x << ", " << resultData[i].y << ", " << resultData[i].z << ", " << resultData[i].w << ", ";
    }
    std::cout << "\n\n\n";
    vkUnmapMemory(device, outputMemory);

    // ------------------------
    // 16. Resources Cleanup
    // ------------------------
    vkFreeCommandBuffers(device, cmdPool, 1, &cmdBuffer);
    vkDestroyCommandPool(device, cmdPool, nullptr);
    vkDestroyFence(device, fence, nullptr);
    vkDestroyPipeline(device, pipeline, nullptr);
    vkDestroyPipelineLayout(device, layout, nullptr);
    vkDestroyShaderModule(device, shader, nullptr);
    vkDestroyDescriptorSetLayout(device, descriptorLayout, nullptr);

    vkDestroyBuffer(device, inputBuffer, nullptr);
    vkDestroyBuffer(device, outputBuffer, nullptr);
    vkFreeMemory(device, inputMemory, nullptr);
    vkFreeMemory(device, outputMemory, nullptr);

    if (VK_SUCCESS != vkDeviceWaitIdle(device))
        std::cout << "Can't wait for device to idle\n";
    vkDestroyDevice(device, nullptr);
    vkDestroyInstance(instance, nullptr);
}

Answer 1

I think the problem might be mis-synchronization, specifically missing memory domain operation . Some platform might not like it...

At the end of your command buffer you need this special pipeline barrier that transitions the writes from device domain to the host domain:

VkBufferMemoryBarrier outbuffDependency = {};
outbuffDependency.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER;
outbuffDependency.srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT;
outbuffDependency.dstAccessMask = VK_ACCESS_HOST_READ_BIT;
outbuffDependency.buffer = outputBuffer;
outbuffDependency.size = VK_WHOLE_SIZE;

vkCmdPipelineBarrier(
    cmdBuffer,
    VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_HOST_BIT,
    (VkDependencyFlags)0,
    0, nullptr,
    1, &outbuffDependency,
    0, nullptr
);
    

---

Vulkan has a distinct concept of memory domains. There is a host domain, and there is a device domain. Same memory can have different state in each domain. Eg that memory write is visible in device domain does not mean it is also visible in the host domain.

A fence (or vk*WaitIdle ) does not include memory domain operation as warned in the specification:

Note

Signaling a fence and waiting on the host does not guarantee that the results of memory accesses will be visible to the host, as the access scope of a memory dependency defined by a fence only includes device access. A memory barrier or other memory dependency must be used to guarantee this. See the description of host access types for more information.

Only thing that does include domain operation is a memory dependency with VK_PIPELINE_STAGE_HOST_BIT , or vkQueueSubmit (which you did use with your inputBuffer to transfer it from host to device domain).

Validation layers cannot reasonably catch this error, because they have no way to know (without some intrusive OS debug features) if you actually read from the buffer via mapped pointer.

Answer 2

SO, it's finally working:)
I've made so many changes when trying soe stuff out that at some point my input buffer was bound as a uniform buffer...
Now that it is back as a storage buffer and descriptor sets are created and updated properly I get the expected output.
The memory barrier are not mandatory but I guess this is good practise when I'd have a more complicated example with multiple passes with different usage for buffers.
Thanks all for your help, it really helps me figuring out all those little details that can make a Vulkan implementation work or fail.