re-initialize inputs on each execution in executor_runner

[swolchok]

See NOTE for reasoning; request changes if wrong or unclear.

[swolchok]

Stack from ghstack (oldest at bottom):

• Use parallel_for in functional_util's apply_unary_map_fun #9348
• Use parallel_for in elementwise_util #9243
• Make BroadcastIndexesRange efficient if there is no broadcasting #9298
• executor_runner: print total execution time #9342
• -> re-initialize inputs on each execution in executor_runner #9340

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/9340

• 📄 Preview Python docs built from this PR

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

✅ No Failures

As of commit ac789ab with merge base f351ec2 ():
💚 Looks good so far! There are no failures yet. 💚

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

[shoumikhin]

@swolchok ptal if this trunk CI failure is related https://hud.pytorch.org/pytorch/executorch/commit/f80594037f1fcdc28d4f6bab826f7d3d5be594bd

[swolchok]

@swolchok ptal if this trunk CI failure is related https://hud.pytorch.org/pytorch/executorch/commit/f80594037f1fcdc28d4f6bab826f7d3d5be594bd

seems unlikely as it's a failure during building

[swolchok]

https://hud.pytorch.org/hud/pytorch/executorch/main/1?per_page=50&name_filter=arm-reference&mergeLF=true says it started the commit before this one