I'm introducing a new machine model to simultaneously allow simple
subtarget CPU descriptions and support new features of
MachineScheduler.
MachineModel has three categories of data:
1) Basic properties for coarse grained instruction cost model.
2) Scheduler Read/Write resources for simple per-opcode and operand cost model (TBD).
3) Instruction itineraties for detailed per-cycle reservation tables.
These will all live side-by-side. Any subtarget can use any
combination of them. Instruction itineraries will not change in the
near term. In the long run, I expect them to only be relevant for
in-order VLIW machines that have complex contraints and require a
precise scheduling/bundling model. Once itineraries are only actively
used by VLIW-ish targets, they could be replaced by something more
appropriate for those targets.
This tablegen backend rewrite sets things up for introducing
MachineModel type #2: per opcode/operand cost model.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159891 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/ARM/ARM.td b/lib/Target/ARM/ARM.td
index d332d20..cd3c0e0 100644
--- a/lib/Target/ARM/ARM.td
+++ b/lib/Target/ARM/ARM.td
@@ -204,13 +204,13 @@
FeatureDSPThumb2]>;
// V7a Processors.
-def : Processor<"cortex-a8", CortexA8Itineraries,
+def : ProcessorModel<"cortex-a8", CortexA8Model,
[ProcA8, HasV7Ops, FeatureNEON, FeatureDB,
FeatureDSPThumb2, FeatureHasRAS]>;
-def : Processor<"cortex-a9", CortexA9Itineraries,
+def : ProcessorModel<"cortex-a9", CortexA9Model,
[ProcA9, HasV7Ops, FeatureNEON, FeatureDB,
FeatureDSPThumb2, FeatureHasRAS]>;
-def : Processor<"cortex-a9-mp", CortexA9Itineraries,
+def : ProcessorModel<"cortex-a9-mp", CortexA9Model,
[ProcA9, HasV7Ops, FeatureNEON, FeatureDB,
FeatureDSPThumb2, FeatureMP,
FeatureHasRAS]>;
diff --git a/lib/Target/ARM/ARMScheduleA8.td b/lib/Target/ARM/ARMScheduleA8.td
index 61de00a..56197d4 100644
--- a/lib/Target/ARM/ARMScheduleA8.td
+++ b/lib/Target/ARM/ARMScheduleA8.td
@@ -22,11 +22,7 @@
//
// Dual issue pipeline represented by A8_Pipe0 | A8_Pipe1
//
-def CortexA8Itineraries : MultiIssueItineraries<
- 2, // IssueWidth
- -1, // MinLatency - OperandCycles are interpreted as MinLatency.
- 2, // LoadLatency - overriden by OperandCycles.
- 10, // HighLatency - currently unused.
+def CortexA8Itineraries : ProcessorItineraries<
[A8_Pipe0, A8_Pipe1, A8_LSPipe, A8_NPipe, A8_NLSPipe],
[], [
// Two fully-pipelined integer ALU pipelines
@@ -1061,3 +1057,18 @@
InstrStage<1, [A8_NPipe], 0>,
InstrStage<2, [A8_NLSPipe]>], [4, 1, 2, 2, 3, 3, 1]>
]>;
+
+// ===---------------------------------------------------------------------===//
+// This following definitions describe the simple machine model which
+// will replace itineraries.
+
+// Cortex-A8 machine model for scheduling and other instruction cost heuristics.
+def CortexA8Model : SchedMachineModel {
+ let IssueWidth = 2; // 2 micro-ops are dispatched per cycle.
+ let MinLatency = -1; // OperandCycles are interpreted as MinLatency.
+ let LoadLatency = 2; // Optimistic load latency assuming bypass.
+ // This is overriden by OperandCycles if the
+ // Itineraries are queried instead.
+
+ let Itineraries = CortexA8Itineraries;
+}
diff --git a/lib/Target/ARM/ARMScheduleA9.td b/lib/Target/ARM/ARMScheduleA9.td
index 1677ba6..738974e 100644
--- a/lib/Target/ARM/ARMScheduleA9.td
+++ b/lib/Target/ARM/ARMScheduleA9.td
@@ -11,6 +11,10 @@
//
//===----------------------------------------------------------------------===//
+// ===---------------------------------------------------------------------===//
+// This section contains legacy support for itineraries. This is
+// required until SD and PostRA schedulers are replaced by MachineScheduler.
+
//
// Ad-hoc scheduling information derived from pretty vague "Cortex-A9 Technical
// Reference Manual".
@@ -31,11 +35,7 @@
// Bypasses
def A9_LdBypass : Bypass;
-def CortexA9Itineraries : MultiIssueItineraries<
- 2, // IssueWidth - FIXME: A9_Issue0, A9_Issue1 are now redundant.
- 0, // MinLatency - FIXME: for misched, remove InstrStage for OOO operations.
- 2, // LoadLatency - optimistic, assumes bypass, overriden by OperandCycles.
- 10, // HighLatency - currently unused.
+def CortexA9Itineraries : ProcessorItineraries<
[A9_Issue0, A9_Issue1, A9_Branch, A9_ALU0, A9_ALU1, A9_AGU, A9_NPipe, A9_MUX0,
A9_LSUnit, A9_DRegsVFP, A9_DRegsN],
[A9_LdBypass], [
@@ -1874,3 +1874,21 @@
InstrStage<2, [A9_NPipe]>],
[4, 1, 2, 2, 3, 3, 1]>
]>;
+
+// ===---------------------------------------------------------------------===//
+// This following definitions describe the simple machine model which
+// will replace itineraries.
+
+// Cortex-A9 machine model for scheduling and other instruction cost heuristics.
+def CortexA9Model : SchedMachineModel {
+ let IssueWidth = 2; // 2 micro-ops are dispatched per cycle.
+ let MinLatency = 0; // Data dependencies are allowed within dispatch groups.
+ let LoadLatency = 2; // Optimistic load latency assuming bypass.
+ // This is overriden by OperandCycles if the
+ // Itineraries are queried instead.
+
+ let Itineraries = CortexA9Itineraries;
+}
+
+// TODO: Add Cortex-A9 processor and scheduler resources.
+