Add a new concurrent vertex algorithm based on DBSCAN (that reuses parts of the pixel cluster algorithm).
It still needs to be tuned to reach at least the performance of current DivisiveClustering algorithm used at HLT.

On K40c and P100 we get a segmentation violation in
PixelVertexHeterogeneousProducer::produceGPUCuda(...).

…kport cms-sw#25163) (#202)

Backport "Migrate tracker local reconstruction and pixel tracking to Tasks" (cms-sw#25163) to the Patatrack branch:
  - migrate RecoLocalTracker_cff to Tasks;
  - migrate RecoPixelVertexing_cff to Tasks;
  - keeping sequences to avoid massive migration (for now).

Port and optimise the full workflow from pixel raw data to pixel tracks and vertices to GPUs.
Clean the pixel n-tuplets with the "fishbone" algorithm (only on GPUs).

Other changes:
  - recover the Riemann fit updates lost during the merge with CMSSW 10.4.x;
  - speed up clustering and track fitting;
  - minor bug fix to avoid trivial regression with the optimized fit.

Make unit tests that require a CUDA device skip the test and exit
succesfully if the CUDA runtime is not available, or no CUDA devices
are available.

* Add DQM for pixel vertices

* Add pT>0.9GeV pixel track collections to MTV

* Add dzPV0p1, Pt0to1, Pt1 variants of pixel track DQM

Add two alternative (faster) track clusterizers: one based on DBSCAN,
and one "by density"; use the latter by default.

Allow all pixel triplets, but protect the vertex from triplets.

Use a larger-then-needed nearest neightbours array to allow for possible
duplicate pixels, as a pixel can appear more than once in the same event.

Use a separate workspace for temporary data.

  - port the whole pixel workflow to new heterogeneous framework
  - implement a legacy cluster to SoA converter for the pixel RecHits
  - update the vertex producer to run on CPU as well as GPU

…vice (#364)

To reduce dependencies on edm::Service, and to make CUDAService less
of a collection of everything, split off from it:
  - the CUDAEventCache
  - the CUDAStreamCache
  - the caching allocators

Other changes:
  - clean up unnecessary use of CUDAService
  - fix maxCachedFraction, add debug printouts
  - add make_*_unique_uninitialized that avoid the static_assert

Implement a wrapper to launch a CUDA kernel without using the
non-standard CUDA <<<...>>> syntax, based on the cudaLaunchKernel
library function.

Implement a similar wrapper for cudaLaunchCooperativeKernel.

Migrate code base from cuda::launch to cudautils::launch.

…389)

Replace cuda::stream_t<> with cudaStream_t in client code
Replace cuda::event_t with cudaEvent_t in the client code
Clean up BuildFiles

Otherwise there are possibilities for weird races, e.g. combination of
non-ExternalWork producers, consumed-but-not-read CUDAProducts, CUDA
streams executing work later than expected (= on the next event).

…_ptrs (#396)

Replace cuda::memory::device::make_unique() calls with cudautils::make_device_unique()
Replace cuda::memory::host::make_unique() with cudautils::make_host_unique()

Replaces the usage of cuda::device::count(), cuda::device::get(), cuda::device::set() and cuda::device::current::get() with native CUDA calls.

Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.

Major changes:
  - restructure the RecoPixelVertexing/PixelVertexFinding package;
  - update the interface of PixelCPEFast.

Clean up the Patatrack code base following the comments received during the integration into the upstream release.

Currently tracks the changes introduced due to
   - cms-sw#29109: Patatrack integration - trivial changes (1/N)
   - cms-sw#29110: Patatrack integration - common tools (2/N)

List of changes:
 * Remove unused files
 * Fix compilation warnings
 * Fix AtomicPairCounter unit test
 * Rename the cudaCompat namespace to cms::cudacompat
 * Remove extra semicolon
 * Move SimpleVector and VecArray to the cms::cuda namespace
 * Add missing dependency
 * Move HistoContainer, AtomicPairCounter, prefixScan and radixSort to the cms::cuda namespace
 * Remove rule exception for HeterogeneousCore
 * Fix code rule violations:
    - replace using namespace cms::cuda in test/OneToManyAssoc_t.h .
    - add an exception for cudaCompat.h:
      cudaCompat relies on defining equivalent symbols to the CUDA
      intrinsics in the cms::cudacompat namespace, and pulling them in the
      global namespace when compiling device code without CUDA.
* Protect the headers to compile only with a CUDA compiler

Clean up instances of using namespace ... from header files,
following the comments from the upstream integration.

Make the size consistent with the buffer in track SoA.

…sw#30350)

Backport cms-sw#30105: add ECAL-only workflows for data.
Backport cms-sw#30136: add HCAL-only workflows for MC and data.

Remove the possibility of changing the grid size used by the
cms::cudacompat layer, and make it a constant equal to {1, 1, 1}.

This avoids a thread-related problem caused by TBB using worker threads
where the grid size had not been initialised.

The kernel for pixel clustering need to be rewritten to support a
one-dimensional grid to run on the CPU.
Currently they are only used on the GPU in the Patatrack workflows, but
they are exercised on the CPU by the gpuClustering_t tests; those tests
have been commented out until the kernels can be updated.

Address the pixel local reconstruction review comments.

General clean up of the pixel local reconstruction code:
  - remove commented out and obsolete code and data members
  - use named constants more consistently
  - update variable names to follow the coding rules and for better consistency
  - use member initializer lists in the constructors
  - allow `if constexpr` in CUDA code
  - use `std::size` instead of hardcoding the array size
  - convert iterator-based loops to range-based loops
  - replace `cout` and `printf` with `LogDebug` or `LogWarning`
  - use put tokens
  - reorganise the auto-generated cfi files and use them more consistently
  - adjust code after rearranging an `#ifdef GPU_DEBUG` block
  - apply code formatting
  - other minor changes

Improve comments:
  - improve comments and remove obsolete ones
  - clarify comments and types regarding `HostProduct`
  - update comments about `GPU_SMALL_EVENTS` being kept for testing purposes
  - add notes about the original cpu code

Reuse some more common code:
  - move common pixel cluster code to `PixelClusterizerBase`
  - extend the `SiPixelCluster` constructor

Rename classes and modules for better consistency:
  - remove the `TrackingRecHit2DCUDA.h` and `gpuClusteringConstants.h` forwarding headers
  - rename `PixelRecHits` to `PixelRecHitGPUKernel`
  - rename SiPixelRecHitFromSOA to SiPixelRecHitFromCUDA
  - rename `siPixelClustersCUDAPreSplitting` to `siPixelClustersPreSplittingCUDA`
  - rename `siPixelRecHitsCUDAPreSplitting` to `siPixelRecHitsPreSplittingCUDA`
  - rename `siPixelRecHitsLegacyPreSplitting` to `siPixelRecHitsPreSplittingLegacy`
  - rename `siPixelRecHitHostSoA` to `siPixelRecHitSoAFromLegacy`

Re-apply changes from cms-sw#29805 that were lost in the Patatrack branch.

Updat EDM access:
  - switch to consumes() scheme for event setup;
  - simplify some event data access.

Style fixes:
  - make class member private & fixed problematic cast;
  - format of comments for clang-tidy;
  - chang to enum class to avoid creating a namespace (usage becomes: pixelTrack::Quality::loose);
  - add article reference in comment (it was already further down in the file);
  - fix member functions and classes capitalization;
  - fix one letter or upper case variable names in formulas (trying to keep the naming from the reference article).

Avoid some code repetitions.

Split PixelVertexProducerCUDA produce() method into two methods, for running on the CPU and on the GPU.

Update access to EDM handles and event collections.

General code clean up:
  - use named constants for kernel grid dimensions;
  - replace commented out code with #ifdef-based conditionals;
  - update data member names to follow the coding rules;
  - fix include guard names and replace relative with absolute includes.

Apply code formatting.

Use std::clamp(...) in device code now that CUDA supports c++17.
Name reused constants in the vertex fitting and splitting.

…ction code (#613)