![]() |
Project
|
Compile-time launch geometry of the ITS tracking kernels, per GPU family. Poor man's RTC to be removed/reworked entirely once we can use Gabriele's tuner. More...
Go to the source code of this file.
Classes | |
| struct | o2::its::gpu::KernelOccupancy |
| Minimum resident blocks per compute unit to request when no per-kernel measurement exists. More... | |
Namespaces | |
| namespace | o2 |
| a couple of static helper functions to create timestamp values for CCDB queries or override obsolete objects | |
| namespace | o2::its |
| namespace | o2::its::gpu |
Functions | |
| constexpr KernelOccupancy | o2::its::gpu::uniformOccupancy (int minBlocks) |
| Use the same occupancy floor for every kernel when no per-kernel measurements are available. | |
| constexpr int | o2::its::gpu::gridBlocks (int residentBlocksPerComputeUnit) |
| Number of blocks in a grid whose depth is residentBlocksPerComputeUnit blocks per CU. | |
| constexpr int | o2::its::gpu::gridThreads (int residentBlocksPerComputeUnit) |
| Number of threads covered by a grid whose depth is residentBlocksPerComputeUnit blocks per CU. | |
| constexpr bool | o2::its::gpu::residentCoversFloor () |
| The grid must provide at least as many blocks per CU as the corresponding occupancy floor. | |
Variables | |
| constexpr int | o2::its::gpu::ComputeUnits = 60 |
| constexpr int | o2::its::gpu::WarpSize = 64 |
| constexpr int | o2::its::gpu::GPUThreads = 256 |
| constexpr int | o2::its::gpu::DefaultBlocksPerComputeUnit = 4 |
| constexpr int | o2::its::gpu::MaxBlocksPerComputeUnit = 10 |
| constexpr KernelOccupancy | o2::its::gpu::MinBlocks = uniformOccupancy(1) |
| NVIDIA: unmeasured. | |
| constexpr KernelOccupancy | o2::its::gpu::ResidentBlocks = uniformOccupancy(DefaultBlocksPerComputeUnit) |
Compile-time launch geometry of the ITS tracking kernels, per GPU family. Poor man's RTC to be removed/reworked entirely once we can use Gabriele's tuner.
Definition in file LaunchGeometry.h.