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

5 files changed, 50 insertions, 4 deletions

diff --git a/source/blender/gpu/intern/gpu_context.cc b/source/blender/gpu/intern/gpu_context.cc
index 5af15d1bc3d..98714269402 100644
--- a/source/blender/gpu/intern/gpu_context.cc
+++ b/source/blender/gpu/intern/gpu_context.cc
@@ -186,6 +186,15 @@ void GPU_backend_exit()
   g_backend = nullptr;
 }
 
+eGPUBackendType GPU_backend_get_type()
+{
+  if (g_backend && dynamic_cast<GLBackend *>(g_backend) != nullptr) {
+    return GPU_BACKEND_OPENGL;
+  }
+
+  return GPU_BACKEND_NONE;
+}
+
 GPUBackend *GPUBackend::get()
 {
   return g_backend;
diff --git a/source/blender/gpu/intern/gpu_index_buffer.cc b/source/blender/gpu/intern/gpu_index_buffer.cc
index 3472cc24a74..895b2a8461b 100644
--- a/source/blender/gpu/intern/gpu_index_buffer.cc
+++ b/source/blender/gpu/intern/gpu_index_buffer.cc
@@ -74,11 +74,16 @@ void GPU_indexbuf_init(GPUIndexBufBuilder *builder,
 GPUIndexBuf *GPU_indexbuf_build_on_device(uint index_len)
 {
   GPUIndexBuf *elem_ = GPU_indexbuf_calloc();
-  IndexBuf *elem = unwrap(elem_);
-  elem->init_build_on_device(index_len);
+  GPU_indexbuf_init_build_on_device(elem_, index_len);
   return elem_;
 }
 
+void GPU_indexbuf_init_build_on_device(GPUIndexBuf *elem, uint index_len)
+{
+  IndexBuf *elem_ = unwrap(elem);
+  elem_->init_build_on_device(index_len);
+}
+
 void GPU_indexbuf_join(GPUIndexBufBuilder *builder_to, const GPUIndexBufBuilder *builder_from)
 {
   BLI_assert(builder_to->data == builder_from->data);
@@ -410,9 +415,19 @@ int GPU_indexbuf_primitive_len(GPUPrimType prim_type)
   return indices_per_primitive(prim_type);
 }
 
+void GPU_indexbuf_use(GPUIndexBuf *elem)
+{
+  unwrap(elem)->upload_data();
+}
+
 void GPU_indexbuf_bind_as_ssbo(GPUIndexBuf *elem, int binding)
 {
   unwrap(elem)->bind_as_ssbo(binding);
 }
 
+void GPU_indexbuf_update_sub(GPUIndexBuf *elem, uint start, uint len, const void *data)
+{
+  unwrap(elem)->update_sub(start, len, data);
+}
+
 /** \} */
diff --git a/source/blender/gpu/intern/gpu_index_buffer_private.hh b/source/blender/gpu/intern/gpu_index_buffer_private.hh
index ed7dd830c8c..adc0145f867 100644
--- a/source/blender/gpu/intern/gpu_index_buffer_private.hh
+++ b/source/blender/gpu/intern/gpu_index_buffer_private.hh
@@ -92,11 +92,15 @@ class IndexBuf {
     return is_init_;
   };
 
+  virtual void upload_data(void) = 0;
+
   virtual void bind_as_ssbo(uint binding) = 0;
 
   virtual const uint32_t *read() const = 0;
   uint32_t *unmap(const uint32_t *mapped_memory) const;
 
+  virtual void update_sub(uint start, uint len, const void *data) = 0;
+
  private:
   inline void squeeze_indices_short(uint min_idx, uint max_idx);
   inline uint index_range(uint *r_min, uint *r_max);
diff --git a/source/blender/gpu/intern/gpu_vertex_buffer.cc b/source/blender/gpu/intern/gpu_vertex_buffer.cc
index 5ed9648387f..dba31f501f2 100644
--- a/source/blender/gpu/intern/gpu_vertex_buffer.cc
+++ b/source/blender/gpu/intern/gpu_vertex_buffer.cc
@@ -144,6 +144,12 @@ void GPU_vertbuf_init_with_format_ex(GPUVertBuf *verts_,
   unwrap(verts_)->init(format, usage);
 }
 
+void GPU_vertbuf_init_build_on_device(GPUVertBuf *verts, GPUVertFormat *format, uint v_len)
+{
+  GPU_vertbuf_init_with_format_ex(verts, format, GPU_USAGE_DEVICE_ONLY);
+  GPU_vertbuf_data_alloc(verts, v_len);
+}
+
 GPUVertBuf *GPU_vertbuf_duplicate(GPUVertBuf *verts_)
 {
   return wrap(unwrap(verts_)->duplicate());
@@ -313,6 +319,11 @@ GPUVertBufStatus GPU_vertbuf_get_status(const GPUVertBuf *verts)
   return unwrap(verts)->flag;
 }
 
+void GPU_vertbuf_tag_dirty(GPUVertBuf *verts)
+{
+  unwrap(verts)->flag |= GPU_VERTBUF_DATA_DIRTY;
+}
+
 uint GPU_vertbuf_get_memory_usage()
 {
   return VertBuf::memory_usage;
@@ -323,12 +334,17 @@ void GPU_vertbuf_use(GPUVertBuf *verts)
   unwrap(verts)->upload();
 }
 
+void GPU_vertbuf_wrap_handle(GPUVertBuf *verts, uint64_t handle)
+{
+  unwrap(verts)->wrap_handle(handle);
+}
+
 void GPU_vertbuf_bind_as_ssbo(struct GPUVertBuf *verts, int binding)
 {
   unwrap(verts)->bind_as_ssbo(binding);
 }
 
-void GPU_vertbuf_update_sub(GPUVertBuf *verts, uint start, uint len, void *data)
+void GPU_vertbuf_update_sub(GPUVertBuf *verts, uint start, uint len, const void *data)
 {
   unwrap(verts)->update_sub(start, len, data);
 }
diff --git a/source/blender/gpu/intern/gpu_vertex_buffer_private.hh b/source/blender/gpu/intern/gpu_vertex_buffer_private.hh
index 9531c2c1a5f..2f46295f45a 100644
--- a/source/blender/gpu/intern/gpu_vertex_buffer_private.hh
+++ b/source/blender/gpu/intern/gpu_vertex_buffer_private.hh
@@ -68,6 +68,8 @@ class VertBuf {
   void upload(void);
   virtual void bind_as_ssbo(uint binding) = 0;
 
+  virtual void wrap_handle(uint64_t handle) = 0;
+
   VertBuf *duplicate(void);
 
   /* Size of the data allocated. */
@@ -96,7 +98,7 @@ class VertBuf {
     }
   }
 
-  virtual void update_sub(uint start, uint len, void *data) = 0;
+  virtual void update_sub(uint start, uint len, const void *data) = 0;
   virtual const void *read() const = 0;
   virtual void *unmap(const void *mapped_data) const = 0;