Adriano Di Florio

please test (currently reviewing it, let's just refresh the test)

Hi @tomalin, in principle I believe you could force push your new squashed branch into this (with `-f` in `git push`). This would keep this PR open as it is...

Also (maybe you already know this but it's worth to mention) you could squash the commits you want directly on this branch without having to create a new one. You...

please test

Hi, I was doing some investigations on top of those done by @ferencek and I've found that for the detID that are failing something weird happens. By modifying `SiPixelRawToClusterGPUKernel.cu` with...

You are right. Sorry for the **huge** noise. For the warning I imagine that's `nvcc` not being that smart or me not noticing it.

Having tested the configuration posted by @missirol above with a dummy fix in which we sort the digis in `SiPixelDigisClustersFromSoA` by the raw id: ```diff --git a/RecoLocalTracker/SiPixelClusterizer/plugins/SiPixelDigisClustersFromSoA.cc b/RecoLocalTracker/SiPixelClusterizer/plugins/SiPixelDigisClustersFromSoA.cc index d36c345ecf0..d5d2ae8a0c6...

Also: is there a wf where we may test these updates?

hold - waiting for comments to be addressed - not needed for 12_5 production

+upgrade - nothing to comment from the upgrade point of view: backport of https://github.com/cms-sw/cmssw/pull/38967