gh-104584: Baby steps towards generating and executing traces (#105924)

Added a new, experimental, tracing optimizer and interpreter (a.k.a. "tier 2"). This currently pessimizes, so don't use yet -- this is infrastructure so we can experiment with optimizing passes. To enable it, pass ``-Xuops`` or set ``PYTHONUOPS=1``. To get debug output, set ``PYTHONUOPSDEBUG=N`` where ``N`` is a debug level (0-4, where 0 is no debug output and 4 is excessively verbose).

All of this code is likely to change dramatically before the 3.13 feature freeze. But this is a first step.

3 files changed, 35 insertions, 0 deletions

diff --git a/Include/cpython/optimizer.h b/Include/cpython/optimizer.h
index b2d173f..2664f5b 100644
--- a/Include/cpython/optimizer.h
+++ b/Include/cpython/optimizer.h
@@ -45,6 +45,7 @@ extern _PyOptimizerObject _PyOptimizer_Default;
 
 /* For testing */
 PyAPI_FUNC(PyObject *)PyUnstable_Optimizer_NewCounter(void);
+PyAPI_FUNC(PyObject *)PyUnstable_Optimizer_NewUOpOptimizer(void);
 
 #define OPTIMIZER_BITS_IN_COUNTER 4
 
diff --git a/Include/internal/pycore_uops.h b/Include/internal/pycore_uops.h
new file mode 100644
index 0000000..0e88d7e
--- /dev/null
+++ b/Include/internal/pycore_uops.h
@@ -0,0 +1,31 @@
+#ifndef Py_INTERNAL_UOPS_H
+#define Py_INTERNAL_UOPS_H
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#ifndef Py_BUILD_CORE
+#  error "this header requires Py_BUILD_CORE define"
+#endif
+
+#define _Py_UOP_MAX_TRACE_LENGTH 16
+
+typedef struct {
+    int opcode;
+    uint64_t operand;  // Sometimes oparg, sometimes a cache entry
+} _PyUOpInstruction;
+
+typedef struct {
+    _PyExecutorObject base;
+    _PyUOpInstruction trace[_Py_UOP_MAX_TRACE_LENGTH];  // TODO: variable length
+} _PyUOpExecutorObject;
+
+_PyInterpreterFrame *_PyUopExecute(
+    _PyExecutorObject *executor,
+    _PyInterpreterFrame *frame,
+    PyObject **stack_pointer);
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_INTERNAL_UOPS_H */
diff --git a/Include/pystats.h b/Include/pystats.h
index 034bf05..54c9b8d 100644
--- a/Include/pystats.h
+++ b/Include/pystats.h
@@ -71,6 +71,9 @@ typedef struct _object_stats {
     uint64_t type_cache_dunder_misses;
     uint64_t type_cache_collisions;
     uint64_t optimization_attempts;
+    uint64_t optimization_traces_created;
+    uint64_t optimization_traces_executed;
+    uint64_t optimization_uops_executed;
 } ObjectStats;
 
 typedef struct _stats {