OpenSubDiv: add support for an OpenGL evaluator

This evaluator is used in order to evaluate subdivision at render time, allowing for
faster renders of meshes with a subdivision surface modifier placed at the last
position in the modifier list.

When evaluating the subsurf modifier, we detect whether we can delegate evaluation
to the draw code. If so, the subdivision is first evaluated on the GPU using our own
custom evaluator (only the coarse data needs to be initially sent to the GPU), then,
buffers for the final `MeshBufferCache` are filled on the GPU using a set of
compute shaders. However, some buffers are still filled on the CPU side, if doing so
on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose
logic is hardly GPU compatible).

This is done at the mesh buffer extraction level so that the result can be readily used
in the various OpenGL engines, without having to write custom geometry or tesselation
shaders.

We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in
order to control the data layout, and interpolation. For example, we store vertex colors
as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float
types.

In order to still access the modified geometry on the CPU side, for use in modifiers
or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`.
Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will
create such a wrapper if possible. If the final subdivision surface is not needed on
the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used.

Enabling or disabling GPU subdivision can be done through the user preferences (under
Viewport -> Subdivision).

See patch description for benchmarks.

Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport

Differential Revision: https://developer.blender.org/D12406

1 files changed, 83 insertions, 0 deletions

diff --git a/source/blender/draw/intern/mesh_extractors/extract_mesh_vbo_edge_fac.cc b/source/blender/draw/intern/mesh_extractors/extract_mesh_vbo_edge_fac.cc
index 2e2444a8e3d..5ee34d7fdb2 100644
--- a/source/blender/draw/intern/mesh_extractors/extract_mesh_vbo_edge_fac.cc
+++ b/source/blender/draw/intern/mesh_extractors/extract_mesh_vbo_edge_fac.cc
@@ -25,6 +25,7 @@
 
 #include "GPU_capabilities.h"
 
+#include "draw_subdivision.h"
 #include "extract_mesh.h"
 
 namespace blender::draw {
@@ -216,6 +217,86 @@ static void extract_edge_fac_finish(const MeshRenderData *mr,
   MEM_SAFE_FREE(data->edge_loop_count);
 }
 
+/* Different function than the one used for the non-subdivision case, as we directly take care of
+ * the buggy AMD driver case. */
+static GPUVertFormat *get_subdiv_edge_fac_format()
+{
+  static GPUVertFormat format = {0};
+  if (format.attr_len == 0) {
+    if (GPU_crappy_amd_driver()) {
+      GPU_vertformat_attr_add(&format, "wd", GPU_COMP_F32, 1, GPU_FETCH_FLOAT);
+    }
+    else {
+      GPU_vertformat_attr_add(&format, "wd", GPU_COMP_U8, 1, GPU_FETCH_INT_TO_FLOAT_UNIT);
+    }
+  }
+  return &format;
+}
+
+static void extract_edge_fac_init_subdiv(const DRWSubdivCache *subdiv_cache,
+                                         const MeshRenderData *mr,
+                                         struct MeshBatchCache *cache,
+                                         void *buffer,
+                                         void *UNUSED(data))
+{
+  GPUVertBuf *edge_idx = cache->final.buff.vbo.edge_idx;
+  GPUVertBuf *pos_nor = cache->final.buff.vbo.pos_nor;
+  GPUVertBuf *vbo = static_cast<GPUVertBuf *>(buffer);
+  GPU_vertbuf_init_build_on_device(
+      vbo, get_subdiv_edge_fac_format(), subdiv_cache->num_subdiv_loops + mr->loop_loose_len);
+
+  /* Create a temporary buffer for the edge original indices if it was not requested. */
+  const bool has_edge_idx = edge_idx != nullptr;
+  GPUVertBuf *loop_edge_idx = nullptr;
+  if (has_edge_idx) {
+    loop_edge_idx = edge_idx;
+  }
+  else {
+    loop_edge_idx = GPU_vertbuf_calloc();
+    draw_subdiv_init_origindex_buffer(
+        loop_edge_idx,
+        static_cast<int *>(GPU_vertbuf_get_data(subdiv_cache->edges_orig_index)),
+        subdiv_cache->num_subdiv_loops,
+        0);
+  }
+
+  draw_subdiv_build_edge_fac_buffer(subdiv_cache, pos_nor, loop_edge_idx, vbo);
+
+  if (!has_edge_idx) {
+    GPU_vertbuf_discard(loop_edge_idx);
+  }
+}
+
+static void extract_edge_fac_loose_geom_subdiv(const DRWSubdivCache *subdiv_cache,
+                                               const MeshRenderData *UNUSED(mr),
+                                               const MeshExtractLooseGeom *loose_geom,
+                                               void *buffer,
+                                               void *UNUSED(data))
+{
+  if (loose_geom->edge_len == 0) {
+    return;
+  }
+
+  GPUVertBuf *vbo = static_cast<GPUVertBuf *>(buffer);
+
+  /* Make sure buffer is active for sending loose data. */
+  GPU_vertbuf_use(vbo);
+
+  uint offset = subdiv_cache->num_subdiv_loops;
+  for (int i = 0; i < loose_geom->edge_len; i++) {
+    if (GPU_crappy_amd_driver()) {
+      float loose_edge_fac[2] = {1.0f, 1.0f};
+      GPU_vertbuf_update_sub(vbo, offset * sizeof(float), sizeof(loose_edge_fac), loose_edge_fac);
+    }
+    else {
+      char loose_edge_fac[2] = {255, 255};
+      GPU_vertbuf_update_sub(vbo, offset * sizeof(char), sizeof(loose_edge_fac), loose_edge_fac);
+    }
+
+    offset += 2;
+  }
+}
+
 constexpr MeshExtract create_extractor_edge_fac()
 {
   MeshExtract extractor = {nullptr};
@@ -224,6 +305,8 @@ constexpr MeshExtract create_extractor_edge_fac()
   extractor.iter_poly_mesh = extract_edge_fac_iter_poly_mesh;
   extractor.iter_ledge_bm = extract_edge_fac_iter_ledge_bm;
   extractor.iter_ledge_mesh = extract_edge_fac_iter_ledge_mesh;
+  extractor.init_subdiv = extract_edge_fac_init_subdiv;
+  extractor.iter_loose_geom_subdiv = extract_edge_fac_loose_geom_subdiv;
   extractor.finish = extract_edge_fac_finish;
   extractor.data_type = MR_DATA_POLY_NOR;
   extractor.data_size = sizeof(MeshExtract_EdgeFac_Data);