Implement MUBUF instructions for shorts/bytes (#2856)

* implement loads/store instructions for types smaller than dwords

* initialize s16/s8 types

* set profile for int8/16/64

* also need to zero extend u8/u16 to u32 result

* document unrelated bugs with atomic fmin/max

* remove profile checks and simple emit for added opcodes

---------

Co-authored-by: georgemoralis <giorgosmrls@gmail.com>

src/video_core/renderer_vulkan/vk_instance.h

@@ -94,6 +94,21 @@ public:
         return features.shaderFloat64;
     }
 
+    /// Returns true if 64-bit ints are supported in shaders
+    bool IsShaderInt64Supported() const {
+        return features.shaderInt64;
+    }
+
+    /// Returns true if 16-bit ints are supported in shaders
+    bool IsShaderInt16Supported() const {
+        return features.shaderInt16;
+    }
+
+    /// Returns true if 8-bit ints are supported in shaders
+    bool IsShaderInt8Supported() const {
+        return vk12_features.shaderInt8;
+    }
+
     /// Returns true when VK_EXT_custom_border_color is supported
     bool IsCustomBorderColorSupported() const {
         return custom_border_color;

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -203,6 +203,9 @@ PipelineCache::PipelineCache(const Instance& instance_, Scheduler& scheduler_,
     profile = Shader::Profile{
         .supported_spirv = SpirvVersion1_6,
         .subgroup_size = instance.SubgroupSize(),
+        .support_int8 = instance.IsShaderInt8Supported(),
+        .support_int16 = instance.IsShaderInt16Supported(),
+        .support_int64 = instance.IsShaderInt64Supported(),
         .support_float64 = instance.IsShaderFloat64Supported(),
         .support_fp32_denorm_preserve = bool(vk12_props.shaderDenormPreserveFloat32),
         .support_fp32_denorm_flush = bool(vk12_props.shaderDenormFlushToZeroFloat32),