Rate-limit drags from the start of the previous render, not its end

During a drag the render loop skipped a render until 0.9 x the usual
iteration time had passed since last_render_t, and last_render_t was
stamped when a render finished. A new render therefore started every
(render time + 0.9 x iteration time), about half the image rate the FPS
readout shows: at Max W 960 on ZeroGPU one image per ~100 ms instead of
~52 ms. Stamp last_render_t when the render starts.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

viser_app/viewer.py

@@ -252,6 +252,11 @@ def render_loop(
             time.sleep(0.020)
             continue
         iter_t0 = time.monotonic()
+        # Rate-limit clock: stamped when this render STARTS. Stamped at the end,
+        # the check above made the next render wait another 0.9 x iteration
+        # time after this one finished, so frames came out at about half the
+        # rate the FPS readout shows.
+        last_render_t = iter_t0
 
         intr, c2w_r, c2w_t = viser_cam_to_dvsm(
             position, wxyz, fov_y, aspect, device="cuda", dtype=torch.float32
@@ -274,7 +279,6 @@ def render_loop(
             time.sleep(0.1)
             continue
         dt_ms = (time.perf_counter() - t0) * 1000.0
-        last_render_t = time.monotonic()
 
         rgb = (
             img[0, 0].clamp(0.0, 1.0).mul(255.0).to(torch.uint8)