just doing a naive layered compute for meow

src/vkprocessor.rs

@@ -184,11 +184,9 @@ impl<'a> VkProcessor<'a> {
         -> Box<dyn GpuFuture> {
 
         // take the canvas frame and create the vertex buffers
+        // TODO: This performs gpu buffer creation. Shouldn't be in hotpath
         self.canvas.draw(canvas_frame);
 
-        // ditto with the compuframe
-        self.compute_state.compute(compute_frame);
-
         let mut framebuffers =
             self.canvas.window_size_dependent_setup(&self.swapchain_images.clone().unwrap().clone());
 
@@ -229,6 +227,9 @@ impl<'a> VkProcessor<'a> {
                 .copy_buffer_to_image(self.compute_image.clone().unwrap().clone().rw_buffers.get(0).unwrap().clone(),
                                       self.compute_image.clone().unwrap().clone().get_swap_buffer().clone()).unwrap();
 
+        // ditto with the compuframe
+        let mut command_buffer = self.compute_state.compute_commands(compute_frame, command_buffer, &self.canvas);
+
         let mut command_buffer = self.canvas.draw_commands(command_buffer, framebuffers, image_num);
 
         let command_buffer = command_buffer.build().unwrap();