Dan Riley

> In fact a lot of weird bit patterns are interpreted as nan. A NaN has exponent all 1s and the significand/mantissa non-zero (zero mantissa is infinity), so there are...

> [#48207](https://github.com/cms-sw/cmssw/pull/48207) converts this segmentation fault into an assertion failure Assertion failure [has been observed in IBs](https://cmssdt.cern.ch/SDT/cgi-bin/logreader/el8_amd64_gcc12/CMSSW_15_1_NONLTO_X_2025-06-04-1100/pyRelValMatrixLogs/run/34034.0_TTbar_14TeV+Run4D120/step2_TTbar_14TeV+Run4D120.log#/59-59): ``` cmsRun: src/RecoLocalCalo/HGCalRecProducers/plugins/HGCalRecHitWorkerSimple.cc:187: virtual void HGCalRecHitWorkerSimple::run(const edm::Event&, const HGCUncalibratedRecHitCollection&, HGCRecHitCollection&): Assertion `thickness -...

We don't have earlier ASAN builds, as far as I know. My recollection is that WF 138.3 has been failing in ASAN for at least a few months, but there's...

It's probably hitting the stack size limit, and then timing out trying to generate the backtrace

[UBSAN](https://cmssdt.cern.ch/SDT/cgi-bin/logreader/el8_amd64_gcc12/CMSSW_15_1_UBSAN_X_2025-06-18-2300/pyRelValMatrixLogs/run/24034.0_TTbar_14TeV+Run4D96/step2_TTbar_14TeV+Run4D96.log#/137-137) has a bunch more index out of bounds (since it doesn't immediately terminate): ``` src/L1Trigger/L1CaloTrigger/plugins/Phase2L1CaloPFClusterEmulator.cc:163:13: runtime error: load of address 0x152a4f1ea830 with insufficient space for an object of type...

UBSAN builds have more info, see for example [25234.0 step 2](https://cmssdt.cern.ch/SDT/cgi-bin/logreader/el8_amd64_gcc12/CMSSW_15_1_UBSAN_X_2025-07-11-2300/pyRelValMatrixLogs/run/25234.0_TTbar_14TeV+Run4D99/step2_TTbar_14TeV+Run4D99.log#/51-51). It looks like [the first PR](https://github.com/cms-sw/cmssw/pull/48478) fixed the eta direction, but phi still has issues. There are reports at...

Note that the file compiles fine by itself, the error is from LTO in the link phase. That suggests that, if real, the error somehow involves crossing file boundaries in...

I tried stubbing out the routines called by launchKernels(), and found that there are compilation errors in four routines in RecoPixelVertexing/PixelTriplets/plugins/CAHitNtupletGeneratorKernelsImpl.h, specifically kernel_connect(), kernel_countMultiplicity(), kernel_fillMultiplicity(), and kernel_fillHitDetIndices(). For (at least)...

If I comment out either of these two lines: https://github.com/cms-sw/cmssw/blob/df27e39abd260bb8bebfc1cf5fb66be5a88fbb0e/RecoPixelVertexing/PixelTriplets/plugins/CAHitNtupletGeneratorKernels.cc#L228-L229 then compilation succeeds through the LTO intermediate steps. The link ultimately fails due to missing symbols, but it gets past...

I'm running valgrind memcheck on 136.793 and 11834.21 (I was already doing 11834.21 for #42700, I've updated to CMSSW_14_0_X_2023-11-16-1100 to cover this too)