LLVM: lib/Target/AMDGPU/GCNILPSched.cpp File Reference (original) (raw)
Go to the source code of this file.
| Namespaces | |
|---|---|
| namespace | llvm |
| This is an optimization pass for GlobalISel generic memory operations. |
| Macros | |
|---|---|
| #define | DEBUG_TYPE "machine-scheduler" |
| Functions | |
|---|---|
| static unsigned | CalcNodeSethiUllmanNumber (const SUnit *SU, std::vector< unsigned > &SUNumbers) |
| CalcNodeSethiUllmanNumber - Compute Sethi Ullman number. | |
| static unsigned | closestSucc (const SUnit *SU) |
| closestSucc - Returns the scheduled cycle of the successor which is closest to the current cycle. | |
| static unsigned | calcMaxScratches (const SUnit *SU) |
| calcMaxScratches - Returns an cost estimate of the worse case requirement for scratch registers, i.e. | |
| static int | BUCompareLatency (const SUnit *left, const SUnit *right) |
| std::vector< const SUnit * > | llvm::makeGCNILPScheduler (ArrayRef< const SUnit * > BotRoots, const ScheduleDAG &DAG) |
◆ DEBUG_TYPE
#define DEBUG_TYPE "machine-scheduler"
◆ BUCompareLatency()
◆ calcMaxScratches()
calcMaxScratches - Returns an cost estimate of the worse case requirement for scratch registers, i.e.
number of data dependencies.
Definition at line 121 of file GCNILPSched.cpp.
References llvm::SUnit::Preds.