tt-smi reports higher than expected chips

[ttmchiou]

On T3K system

running

tt-smi-metal -s

tt-smi version 2.2.0

https://github.com/tenstorrent/tt-metal/actions/runs/10693593480/job/29690283051
under the "Set up runner" step under t3k ttnn tests

[warthog9]

Please confirm with latest version this may be resolved

[TTDRosen]

That behaviour is expected, new chips are detected before their uniqueness is determined. This is done this way to work around issues where specific ethernet cores have had garbage written into their lower addresses or there was some event which causes the chip id to be unreadable. Once a chip is determined to have already been detected it is removed from the detected count which is why the final number is correct, one way to improve this would be to split the candidate chips and unique chips into two separate counts.