SpotTiles_MatchWholeFrame_GPU: wait for the upload before the squares are cut

cudaMemcpy from pageable memory may return before the copy lands on the card, and
ImagePreprocessorBufferGPU gathers its squares on a non-blocking stream that does not wait for the
NULL stream, so on a loaded GPU the squares could be cut from the previous frame. Test-only race;
the pre-scan's engines synchronise their own streams before the buffer is read.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/SpotTilesTest.cpp

@@ -135,6 +135,9 @@ TEST_CASE("SpotTiles_MatchWholeFrame_GPU", "[SpotWidth][GPU]") {
     const auto upload = [&](Frame &f) -> ImagePreprocessorBuffer & {
         REQUIRE(cudaMemcpy(device.getGPUBuffer(), f.image.data(), f.image.size() * sizeof(int32_t),
                            cudaMemcpyHostToDevice) == cudaSuccess);
+        // From pageable memory cudaMemcpy may return before the copy has landed, and the buffer cuts
+        // its squares on a non-blocking stream that does not wait for the NULL stream.
+        REQUIRE(cudaDeviceSynchronize() == cudaSuccess);
         return device;
     };
     CheckFrames(frames, upload, SPOT_TILE_HALF, FOOTPRINT_TILE_HALF);