anv: Implement acceleration structure API

Rework: (Kevin)
- Properly setup bvh_layout
   Our bvh resides in contiguous memory and can be divided into two sections:
      1. anv_accel_struct_header, tightly followed by
      2. actual bvh, which starts with root node, followed by interleaving
         leaves or internal nodes.
- Update comments for some fields for BVH and nodes.
- Properly populate the UUIDs in serialization header
- separate header func into completely two paths based on compaction bit
- Encode rt_uuid at second VK_UUID_SIZE.
- Write query result at correct slot
- add assertion for a 4B alignment
- move bvh_layout to anv_bvh
- Use meson option to decide which files to compile
- The alignment of serialization size is not needed
- Change static_assert to STATIC_ASSERT and move them inside functions

Rework (Sagar)
- Use anv_cmd_buffer_update_buffer instead of MI to copy data

Rework (Lionel)
- Remove flush after builds, and add flush in copy before dispatch
- Handle the flushes in CmdWriteAccelerationStructuresPropertiesKHR properly

Co-authored-by: Kevin Chuang <[email protected]>
Co-authored-by: Sagar Ghuge <[email protected]>
Reviewed-by: Lionel Landwerlin <[email protected]>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31588>

Author: 2 people

meson.build

@@ -316,6 +316,10 @@ with_intel_vk_rt = get_option('intel-rt') \
   .disable_if(host_machine.cpu_family() != 'x86_64', error_message : 'Intel Ray Tracing is only supported on x86_64') \
   .allowed()
 
+if with_intel_vk_rt
+  with_intel_bvh_grl = get_option('intel-bvh-grl')
+endif
+
 with_any_intel = [
   with_gallium_crocus,
   with_gallium_i915,

meson_options.txt

@@ -657,6 +657,13 @@ option(
   description : 'Build the intel-clc compiler or use a system version.'
 )
 
+option(
+  'intel-bvh-grl',
+  type : 'boolean',
+  value : true,
+  description : 'Build the BVH structure using GRL.'
+)
+
 option(
   'install-intel-clc',
   type : 'boolean',

src/intel/vulkan/anv_device.c

@@ -985,14 +985,22 @@ VkResult anv_CreateDevice(
 
    anv_device_utrace_init(device);
 
-   result = anv_genX(device->info, init_device_state)(device);
+   result = vk_meta_device_init(&device->vk, &device->meta_device);
    if (result != VK_SUCCESS)
       goto fail_utrace;
 
+   result = anv_genX(device->info, init_device_state)(device);
+   if (result != VK_SUCCESS)
+      goto fail_meta_device;
+
+   simple_mtx_init(&device->accel_struct_build.mutex, mtx_plain);
+
    *pDevice = anv_device_to_handle(device);
 
    return VK_SUCCESS;
 
+ fail_meta_device:
+   vk_meta_device_finish(&device->vk, &device->meta_device);
  fail_utrace:
    anv_device_utrace_finish(device);
  fail_queues:
@@ -1118,6 +1126,12 @@ void anv_DestroyDevice(
    /* Do TRTT batch garbage collection before destroying queues. */
    anv_device_finish_trtt(device);
 
+   if (device->accel_struct_build.radix_sort) {
+      radix_sort_vk_destroy(device->accel_struct_build.radix_sort,
+                            _device, &device->vk.alloc);
+   }
+   vk_meta_device_finish(&device->vk, &device->meta_device);
+
    anv_device_utrace_finish(device);
 
    for (uint32_t i = 0; i < device->queue_count; i++)
@@ -1218,6 +1232,8 @@ void anv_DestroyDevice(
    pthread_cond_destroy(&device->queue_submit);
    pthread_mutex_destroy(&device->mutex);
 
+   simple_mtx_destroy(&device->accel_struct_build.mutex);
+
    ralloc_free(device->fp64_nir);
 
    anv_device_destroy_context_or_vm(device);

src/intel/vulkan/anv_private.h

@@ -102,6 +102,7 @@
 #include "vk_log.h"
 #include "vk_ycbcr_conversion.h"
 #include "vk_video.h"
+#include "vk_meta.h"
 
 #ifdef __cplusplus
 extern "C" {
@@ -1828,6 +1829,11 @@ enum anv_rt_bvh_build_method {
    ANV_BVH_BUILD_METHOD_NEW_SAH,
 };
 
+/* If serialization-breaking or algorithm-breaking changes are made,
+ * increment the digits at the end
+ */
+#define ANV_RT_UUID_MACRO             "ANV_RT_BVH_0001"
+
 struct anv_device_astc_emu {
     struct vk_texcompress_astc_state           *texcompress;
 
@@ -2102,6 +2108,14 @@ struct anv_device {
         */
        struct util_dynarray                      prints;
     } printf;
+
+    struct {
+       simple_mtx_t  mutex;
+       struct radix_sort_vk *radix_sort;
+       struct vk_acceleration_structure_build_args build_args;
+   } accel_struct_build;
+
+   struct vk_meta_device meta_device;
 };
 
 static inline uint32_t

src/intel/vulkan/bvh/meson.build

@@ -0,0 +1,50 @@
+# Copyright © 2022 Konstantin Seurer
+# Copyright © 2024 Intel Corporation
+# SPDX-License-Identifier: MIT
+
+# source file, output name, defines
+bvh_shaders = [
+  [
+    'encode.comp',
+    'encode',
+    [],
+  ],
+  [
+    'header.comp',
+    'header',
+    [],
+  ],
+  [
+    'copy.comp',
+    'copy',
+    []
+  ],
+]
+
+anv_bvh_include_dir = dir_source_root + '/src/intel/vulkan/bvh'
+
+anv_bvh_includes = files(
+  'anv_build_helpers.h',
+  'anv_build_interface.h',
+  'anv_bvh.h',
+)
+
+bvh_spv = []
+foreach s : bvh_shaders
+  command = [
+    prog_glslang, '-V', '-I' + vk_bvh_include_dir, '-I' + anv_bvh_include_dir, '--target-env', 'spirv1.5', '-x', '-o', '@OUTPUT@', '@INPUT@'
+  ]
+  command += glslang_quiet
+
+  foreach define : s[2]
+    command += '-D' + define
+  endforeach
+
+  bvh_spv += custom_target(
+    s[1] + '.spv.h',
+    input : s[0],
+    output : s[1] + '.spv.h',
+    command : command,
+    depend_files: [vk_bvh_includes, anv_bvh_includes],
+  )
+endforeach