Mesa (master): radv: advertise VK_EXT_shader_atomic_float

[GitLab Mirror]

Module: Mesa
Branch: master
Commit: 6c1108d25b46ddf7789b2c9f9f7b0a79b3e279a0
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=6c1108d25b46ddf7789b2c9f9f7b0a79b3e279a0

Author: Samuel Pitoiset <samuel.pitoiset at gmail.com>
Date:   Mon Jul 20 18:38:17 2020 +0200

radv: advertise VK_EXT_shader_atomic_float

No hw support for float atomic add for buffer and (sparse) images.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
Reviewed-by: Daniel Schürmann <daniel at schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6000>

---

 src/amd/vulkan/radv_device.c      | 18 ++++++++++++++++++
 src/amd/vulkan/radv_extensions.py |  1 +
 src/amd/vulkan/radv_formats.c     |  4 +++-
 src/amd/vulkan/radv_shader.c      |  1 +
 4 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index b9969e4e5eb..c474c78eb21 100644
--- a/src/amd/vulkan/radv_device.c
+++ b/src/amd/vulkan/radv_device.c
@@ -1340,6 +1340,24 @@ void radv_GetPhysicalDeviceFeatures2(
 			features->robustImageAccess = true;
 			break;
 		}
+		case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_ATOMIC_FLOAT_FEATURES_EXT: {
+			VkPhysicalDeviceShaderAtomicFloatFeaturesEXT *features =
+				(VkPhysicalDeviceShaderAtomicFloatFeaturesEXT *)ext;
+			features->shaderBufferFloat32Atomics = true;
+			features->shaderBufferFloat32AtomicAdd = false;
+			features->shaderBufferFloat64Atomics = true;
+			features->shaderBufferFloat64AtomicAdd = false;
+			features->shaderSharedFloat32Atomics = true;
+			features->shaderSharedFloat32AtomicAdd = pdevice->rad_info.chip_class >= GFX8 &&
+								 (!pdevice->use_llvm || LLVM_VERSION_MAJOR >= 10);
+			features->shaderSharedFloat64Atomics = true;
+			features->shaderSharedFloat64AtomicAdd = false;
+			features->shaderImageFloat32Atomics = true;
+			features->shaderImageFloat32AtomicAdd = false;
+			features->sparseImageFloat32Atomics = false;
+			features->sparseImageFloat32AtomicAdd = false;
+			break;
+		}
 		default:
 			break;
 		}
diff --git a/src/amd/vulkan/radv_extensions.py b/src/amd/vulkan/radv_extensions.py
index 90b5ef41e83..2a3e995e309 100644
--- a/src/amd/vulkan/radv_extensions.py
+++ b/src/amd/vulkan/radv_extensions.py
@@ -153,6 +153,7 @@ EXTENSIONS = [
     Extension('VK_EXT_sample_locations',                  1, 'device->rad_info.chip_class < GFX10'),
     Extension('VK_EXT_sampler_filter_minmax',             1, True),
     Extension('VK_EXT_scalar_block_layout',               1, 'device->rad_info.chip_class >= GFX7'),
+    Extension('VK_EXT_shader_atomic_float',               1, True),
     Extension('VK_EXT_shader_demote_to_helper_invocation',1, 'LLVM_VERSION_MAJOR >= 9 || !device->use_llvm'),
     Extension('VK_EXT_shader_viewport_index_layer',       1, True),
     Extension('VK_EXT_shader_stencil_export',             1, True),
diff --git a/src/amd/vulkan/radv_formats.c b/src/amd/vulkan/radv_formats.c
index 2e753f47be8..9360fb31bcc 100644
--- a/src/amd/vulkan/radv_formats.c
+++ b/src/amd/vulkan/radv_formats.c
@@ -749,7 +749,9 @@ radv_physical_device_get_format_properties(struct radv_physical_device *physical
 		          VK_FORMAT_FEATURE_TRANSFER_DST_BIT;
 	}
 
-	if (format == VK_FORMAT_R32_UINT || format == VK_FORMAT_R32_SINT) {
+	if (format == VK_FORMAT_R32_UINT ||
+	    format == VK_FORMAT_R32_SINT ||
+	    format == VK_FORMAT_R32_SFLOAT) {
 		buffer |= VK_FORMAT_FEATURE_STORAGE_TEXEL_BUFFER_ATOMIC_BIT;
 		linear |= VK_FORMAT_FEATURE_STORAGE_IMAGE_ATOMIC_BIT;
 		tiled |= VK_FORMAT_FEATURE_STORAGE_IMAGE_ATOMIC_BIT;
diff --git a/src/amd/vulkan/radv_shader.c b/src/amd/vulkan/radv_shader.c
index 072201fe66b..3af15d4e844 100644
--- a/src/amd/vulkan/radv_shader.c
+++ b/src/amd/vulkan/radv_shader.c
@@ -377,6 +377,7 @@ radv_shader_compile_to_nir(struct radv_device *device,
 				.draw_parameters = true,
 				.float_controls = true,
 				.float16 = device->physical_device->rad_info.has_packed_math_16bit,
+				.float32_atomic_add = true,
 				.float64 = true,
 				.geometry_streams = true,
 				.image_ms_array = true,