Add benchmark for CUDA blocks

[NicolasHug]

Slight overhead for now

~/dev/torchcodec-cuda (5bb18bfd*) » python benchmarks/bench_blocks.py --device=cuda                                                                                nicolashug@nicolashug-fedora-PW0H326Y

Video: /tmp/bench_blocks_testsrc2_1280x720_30fps_10s.mp4  (testsrc2 1280x720 30fps 10s)
Device: cuda

Method                             Time (ms)     CPU %   vs VideoDecoder
------------------------------------------------------------------------
VideoDecoder                   185.2 +/- 0.5        13             1.00x
sequential                     195.5 +/- 1.2        41             0.95x
demux+decode || cc             193.2 +/- 5.1        51             0.96x
demux || decode+cc             197.6 +/- 7.5        55             0.94x
demux || decode || cc          196.5 +/- 3.9        83             0.94x

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/meta-pytorch/torchcodec/1603

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

⏳ No Failures, 13 Pending

As of commit 1238887 with merge base 5bb18bf ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.