qf/quakeforge
0
0
Fork 0
mirror of https://git.code.sf.net/p/quake/quakeforge synced 2024-11-15 01:11:27 +00:00
Code Issues Projects Releases Packages Wiki Activity
quakeforge/libs/video/renderer/vulkan/deferred.plist

474 lines
11 KiB
Text
Raw Normal View History

[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
{
images = {
depth = {
imageType = VK_IMAGE_TYPE_2D; //FIXME short form is 2d...
format = x8_d24_unorm_pack32;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
usage = depth_stencil_attachment|input_attachment|transient_attachment;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
initialLayout = undefined;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
color = {
imageType = VK_IMAGE_TYPE_2D;
format = r8g8b8a8_unorm;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
usage = color_attachment|input_attachment|transient_attachment;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
initialLayout = undefined;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
emission = {
imageType = VK_IMAGE_TYPE_2D;
format = r16g16b16a16_sfloat;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
usage = color_attachment|input_attachment|transient_attachment;
initialLayout = undefined;
};
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
normal = {
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
imageType = VK_IMAGE_TYPE_2D;
format = r16g16b16a16_sfloat;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
usage = color_attachment|input_attachment|transient_attachment;
initialLayout = undefined;
};
position = {
imageType = VK_IMAGE_TYPE_2D;
format = r32g32b32a32_sfloat;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
usage = color_attachment|input_attachment|transient_attachment;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
initialLayout = undefined;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
opaque = {
imageType = VK_IMAGE_TYPE_2D;
format = r8g8b8a8_unorm;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
usage = color_attachment|input_attachment|transient_attachment;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
initialLayout = undefined;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
translucent = {
imageType = VK_IMAGE_TYPE_2D;
format = r8g8b8a8_unorm;
samples = 1;
extent = {
width = $swapchain.extent.width;
height = $swapchain.extent.height;
depth = 1;
};
mipLevels = 1;
arrayLayers = 1;
tiling = optimal;
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
usage = color_attachment|input_attachment|transient_attachment;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
initialLayout = undefined;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
};
imageViews = {
depth = {
image = depth;
viewType = VK_IMAGE_VIEW_TYPE_2D;
format = $properties.images.depth.format;
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = depth;
levelCount = 1;
layerCount = 1;
};
};
color = {
image = color;
viewType = VK_IMAGE_VIEW_TYPE_2D;
format = $properties.images.color.format;
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = color;
levelCount = 1;
layerCount = 1;
};
};
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
emission = {
image = emission;
viewType = VK_IMAGE_VIEW_TYPE_2D;
format = $properties.images.emission.format;
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = color;
levelCount = 1;
layerCount = 1;
};
};
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
normal = {
image = normal;
viewType = VK_IMAGE_VIEW_TYPE_2D;
format = $properties.images.normal.format;
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = color;
levelCount = 1;
layerCount = 1;
};
};
position = {
image = position;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
viewType = VK_IMAGE_VIEW_TYPE_2D;
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
format = $properties.images.position.format;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = color;
levelCount = 1;
layerCount = 1;
};
};
opaque = {
image = opaque;
viewType = VK_IMAGE_VIEW_TYPE_2D;
format = $properties.images.opaque.format;
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = color;
levelCount = 1;
layerCount = 1;
};
};
translucent = {
image = translucent;
viewType = VK_IMAGE_VIEW_TYPE_2D;
format = $properties.images.translucent.format;
components = {
r = identity;
g = identity;
b = identity;
a = identity;
};
subresourceRange = {
aspectMask = color;
levelCount = 1;
layerCount = 1;
};
};
};
framebuffer = {
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
renderPass = $properties.renderpass;
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
attachments = (depth, color, emission, normal, position, opaque,
translucent, "$swapchain.views[$swapImageIndex]");
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
width = $swapchain.extent.width;
height = $swapchain.extent.height;
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
layers = 1;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
clearValues = (
{ depthStencil = { depth = 1; stencil = 0; }; },
{ color = "[0, 0, 0, 1]"; }, // color
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ color = "[0, 0, 0, 1]"; }, // emission
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
{ color = "[0, 0, 0, 1]"; }, // normal
{ color = "[0, 0, 0, 1]"; }, // position
[vulkan] Parse clear values And get the render pass working in general. Note that this is only the render pass and frame buffers: actual commands still have problems.
2021-02-23 05:37:48 +00:00
{ color = "[0, 0, 0, 1]"; }, // opaque
{ color = "[0, 0, 0, 0]"; }, // translucent
{ color = "[0, 0, 0, 1]"; }, // swapchain
);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
renderpass = {
attachments = (
{
format = $properties.images.depth.format;
samples = 1;
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
loadOp = clear;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = depth_stencil_attachment_optimal;
},
{
format = $properties.images.color.format;
samples = 1;
loadOp = dont_care;
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = color_attachment_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{
format = $properties.images.emission.format;
samples = 1;
loadOp = dont_care;
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = color_attachment_optimal;
},
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
{
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
format = $properties.images.normal.format;
samples = 1;
loadOp = dont_care;
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = color_attachment_optimal;
},
{
format = $properties.images.position.format;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
samples = 1;
loadOp = dont_care;
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = color_attachment_optimal;
},
{
format = $properties.images.opaque.format;
samples = 1;
loadOp = dont_care;
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = color_attachment_optimal;
},
{
format = $properties.images.translucent.format;
samples = 1;
loadOp = dont_care;
storeOp = dont_care;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = color_attachment_optimal;
},
{
format = $swapchain.format;
samples = 1;
loadOp = clear;
storeOp = store;
stencilLoadOp = dont_care;
stencilStoreOp = dont_care;
initialLayout = undefined;
finalLayout = present_src_khr;
},
);
subpasses = (
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
{ // 0 depth
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
pipelineBindPoint = graphics;
depthStencilAttachment = {
attachment = 0;
layout = depth_stencil_attachment_optimal;
};
},
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
{ // 1 translucent
pipelineBindPoint = graphics;
colorAttachments = (
{ // translucent
attachment = 6;
layout = color_attachment_optimal;
},
);
depthStencilAttachment = {
attachment = 0;
layout = depth_stencil_read_only_optimal;
};
preserveAttachments = (1, 2, 3, 4, 5);
},
{ // 2 g-buffer generation
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
pipelineBindPoint = graphics;
colorAttachments = (
{ // color
attachment = 1;
layout = color_attachment_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ // emission
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
attachment = 2;
layout = color_attachment_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ // normal
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
attachment = 3;
layout = color_attachment_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ // position
attachment = 4;
layout = color_attachment_optimal;
},
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
);
depthStencilAttachment = {
attachment = 0;
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
layout = depth_stencil_read_only_optimal;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
};
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
preserveAttachments = (6);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
},
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
{ // 3 lighting
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
pipelineBindPoint = graphics;
inputAttachments = (
{ // depth
attachment = 0;
layout = shader_read_only_optimal;
},
{ // color
attachment = 1;
layout = shader_read_only_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ // emission
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
attachment = 2;
layout = shader_read_only_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ // normal
[vulkan] Add a position buffer to the g-buffer While I could reconstruct the position from the screen coords and depth, this is easier and good enough for now. Reconstruction is an optimization thing.
2021-02-25 04:46:33 +00:00
attachment = 3;
layout = shader_read_only_optimal;
},
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
{ // position
attachment = 4;
layout = shader_read_only_optimal;
},
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
);
colorAttachments = (
{ // opaque
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
attachment = 5;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
layout = color_attachment_optimal;
},
);
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
preserveAttachments = (6);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
},
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
{ // 4 compose
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
pipelineBindPoint = graphics;
inputAttachments = (
{ // opaque
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
attachment = 5;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
layout = shader_read_only_optimal;
},
{ // translucent
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
attachment = 6;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
layout = shader_read_only_optimal;
},
);
colorAttachments = (
{ // swapchain
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
attachment = 7;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
layout = color_attachment_optimal;
},
);
[vulkan] Implement deferred emission (fullbrights) That was... easier than expected. A little more tedious that I would have liked, but my scripting system isn't perfect (I suspect it's best suited as the output of a code generator), and the C side could do with a little more automation.
2021-03-22 10:08:16 +00:00
preserveAttachments = (0, 1, 2, 3, 4);
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
},
);
dependencies = (
{
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
srcSubpass = 0; // depth
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
dstSubpass = 1; // translucent
[vulkan] Use more correct dependency flags I was reading about multi-pass rendering on mobile devices (https://developer.oculus.com/blog/loads-stores-passes-and-advanced-gpu-pipelines/) and discovered that I had used the wrong flags (but then, I think Graham Sellers had, too, since used his Vulkan Programming Guide as a reference). Doesn't seem to make any difference on desktop, but as there's no loss there, but potential gains on mobile, I'd say it's a win.
2021-04-18 09:56:31 +00:00
srcStageMask = late_fragment_tests;
dstStageMask = fragment_shader|early_fragment_tests;
srcAccessMask = depth_stencil_attachment_write;
dstAccessMask = input_attachment_read|depth_stencil_attachment_read;
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
dependencyFlags = by_region;
},
{
srcSubpass = 0; // depth
dstSubpass = 2; // g-buffer
[vulkan] Use more correct dependency flags I was reading about multi-pass rendering on mobile devices (https://developer.oculus.com/blog/loads-stores-passes-and-advanced-gpu-pipelines/) and discovered that I had used the wrong flags (but then, I think Graham Sellers had, too, since used his Vulkan Programming Guide as a reference). Doesn't seem to make any difference on desktop, but as there's no loss there, but potential gains on mobile, I'd say it's a win.
2021-04-18 09:56:31 +00:00
srcStageMask = late_fragment_tests;
dstStageMask = early_fragment_tests;
srcAccessMask = depth_stencil_attachment_write;
dstAccessMask = depth_stencil_attachment_read;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dependencyFlags = by_region;
},
{
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
srcSubpass = 2; // g-buffer
dstSubpass = 3; // lighting
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
srcStageMask = color_attachment_output;
dstStageMask = fragment_shader;
srcAccessMask = color_attachment_write;
[vulkan] Use more correct dependency flags I was reading about multi-pass rendering on mobile devices (https://developer.oculus.com/blog/loads-stores-passes-and-advanced-gpu-pipelines/) and discovered that I had used the wrong flags (but then, I think Graham Sellers had, too, since used his Vulkan Programming Guide as a reference). Doesn't seem to make any difference on desktop, but as there's no loss there, but potential gains on mobile, I'd say it's a win.
2021-04-18 09:56:31 +00:00
dstAccessMask = input_attachment_read;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dependencyFlags = by_region;
},
{
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
srcSubpass = 3; // lighting
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
dstSubpass = 4; // compose
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
srcStageMask = color_attachment_output;
dstStageMask = fragment_shader;
srcAccessMask = color_attachment_write;
[vulkan] Use more correct dependency flags I was reading about multi-pass rendering on mobile devices (https://developer.oculus.com/blog/loads-stores-passes-and-advanced-gpu-pipelines/) and discovered that I had used the wrong flags (but then, I think Graham Sellers had, too, since used his Vulkan Programming Guide as a reference). Doesn't seem to make any difference on desktop, but as there's no loss there, but potential gains on mobile, I'd say it's a win.
2021-04-18 09:56:31 +00:00
dstAccessMask = input_attachment_read;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dependencyFlags = by_region;
},
{
[vulkan] Fix sky depth issues There were actually several problems: translucency wasn't using or depending on the depth buffer, and the depth buffer wasn't marked as read-only in the g-buffer pass. Getting that correct seems to have given bigass1 a 0.5% boost (hard to say, could be the usual noise).
2021-03-23 02:26:24 +00:00
srcSubpass = 1; // translucent
[vulkan] Implement lighting and compose passes Lighting doesn't actually do lights yet, but it's producing pixels. Translucent seems to be working (2d draw uses it), and compose seems to be working.
2021-02-24 10:58:31 +00:00
dstSubpass = 4; // compose
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
srcStageMask = color_attachment_output;
dstStageMask = fragment_shader;
srcAccessMask = color_attachment_write;
[vulkan] Use more correct dependency flags I was reading about multi-pass rendering on mobile devices (https://developer.oculus.com/blog/loads-stores-passes-and-advanced-gpu-pipelines/) and discovered that I had used the wrong flags (but then, I think Graham Sellers had, too, since used his Vulkan Programming Guide as a reference). Doesn't seem to make any difference on desktop, but as there's no loss there, but potential gains on mobile, I'd say it's a win.
2021-04-18 09:56:31 +00:00
dstAccessMask = input_attachment_read;
[vulkan] Start moving towards a deferred renderer After getting lights even vaguely working for alias models, I realized that it just wasn't going to be feasible to do nice lighting with forward rendering. This gets the bulk of the work done for deferred rendering, but still need to sort out the shaders before any real testing can be done.
2021-02-14 02:35:06 +00:00
dependencyFlags = by_region;
},
);
};
}
Reference in a new issue Copy permalink