i#6426: Add schedule_stats tool (#6442)

Adds a new analysis tool "schedule_stats" which computes characteristics
of a core-sharded schedule. This is meant to help in understanding and
studying the effects of varying the scheduler parameters.

This initial version counts instructions, blocking and all syscalls,
voluntary and direct switches, and direct requests, and computes derived
statistics CSPKI, I/CS, %voluntary, and %direct switches. Failed direct
requests can be inferred.

The tool also records a string of letters representing input threads to
help visualize the thread interleaving on each core. The number of
instructions per leter is controlled by a new parameter -schedule_stats_print_every.
This string visualization has its limits as more than 26 inputs will
wrap around, but this approach has been useful in our unit tests and
other small runs.

Adds a regression test which runs the checked-in threadsig trace.

Issue: #6426

clients/drcachesim/CMakeLists.txt

@@ -166,6 +166,7 @@ add_exported_library(drmemtrace_syscall_mix STATIC tools/syscall_mix.cpp)
 add_exported_library(drmemtrace_view STATIC tools/view.cpp)
 add_exported_library(drmemtrace_func_view STATIC tools/func_view.cpp)
 add_exported_library(drmemtrace_invariant_checker STATIC tools/invariant_checker.cpp)
+add_exported_library(drmemtrace_schedule_stats STATIC tools/schedule_stats.cpp)
 
 target_link_libraries(drmemtrace_invariant_checker drdecode)
 
@@ -273,7 +274,8 @@ configure_DynamoRIO_standalone(drcachesim)
 target_link_libraries(drcachesim drmemtrace_simulator drmemtrace_reuse_distance
   drmemtrace_histogram drmemtrace_reuse_time drmemtrace_basic_counts
   drmemtrace_opcode_mix drmemtrace_syscall_mix drmemtrace_view drmemtrace_func_view
-  drmemtrace_raw2trace directory_iterator drmemtrace_invariant_checker)
+  drmemtrace_raw2trace directory_iterator drmemtrace_invariant_checker
+  drmemtrace_schedule_stats)
 if (UNIX)
     target_link_libraries(drcachesim dl)
 endif ()
@@ -346,6 +348,7 @@ install_client_nonDR_header(drmemtrace tools/histogram_create.h)
 install_client_nonDR_header(drmemtrace tools/reuse_time_create.h)
 install_client_nonDR_header(drmemtrace tools/basic_counts_create.h)
 install_client_nonDR_header(drmemtrace tools/opcode_mix_create.h)
+install_client_nonDR_header(drmemtrace tools/schedule_stats_create.h)
 install_client_nonDR_header(drmemtrace tools/syscall_mix_create.h)
 install_client_nonDR_header(drmemtrace simulator/cache_simulator.h)
 install_client_nonDR_header(drmemtrace simulator/cache_simulator_create.h)
@@ -573,6 +576,7 @@ restore_nonclient_flags(drmemtrace_func_view)
 restore_nonclient_flags(drmemtrace_record_filter)
 restore_nonclient_flags(drmemtrace_analyzer)
 restore_nonclient_flags(drmemtrace_invariant_checker)
+restore_nonclient_flags(drmemtrace_schedule_stats)
 
 # We need to pass /EHsc and we pull in libcmtd into drcachesim from a dep lib.
 # Thus we need to override the /MT with /MTd.
@@ -638,6 +642,7 @@ add_win32_flags(drmemtrace_func_view)
 add_win32_flags(drmemtrace_record_filter)
 add_win32_flags(drmemtrace_analyzer)
 add_win32_flags(drmemtrace_invariant_checker)
+add_win32_flags(drmemtrace_schedule_stats)
 add_win32_flags(directory_iterator)
 add_win32_flags(test_helpers)
 if (WIN32 AND DEBUG)
@@ -809,7 +814,7 @@ if (BUILD_TESTS)
     drmemtrace_histogram drmemtrace_reuse_time drmemtrace_basic_counts
     drmemtrace_opcode_mix drmemtrace_syscall_mix drmemtrace_view drmemtrace_func_view
     drmemtrace_raw2trace directory_iterator drmemtrace_invariant_checker
-    drmemtrace_analyzer)
+    drmemtrace_schedule_stats drmemtrace_analyzer)
   if (UNIX)
     target_link_libraries(tool.drcachesim.core_sharded dl)
   endif ()

clients/drcachesim/analyzer_multi.cpp

@@ -58,6 +58,7 @@
 #include "tools/invariant_checker.h"
 #include "tools/invariant_checker_create.h"
 #include "tools/opcode_mix_create.h"
+#include "tools/schedule_stats_create.h"
 #include "tools/syscall_mix_create.h"
 #include "tools/reuse_distance_create.h"
 #include "tools/reuse_time_create.h"
@@ -427,6 +428,9 @@ analyzer_multi_t::create_analysis_tool_from_options(const std::string &simulator
                                      op_verbose.get_value());
     } else if (simulator_type == INVARIANT_CHECKER) {
         return create_invariant_checker();
+    } else if (simulator_type == SCHEDULE_STATS) {
+        return schedule_stats_tool_create(op_schedule_stats_print_every.get_value(),
+                                          op_verbose.get_value());
     } else {
         auto tool = create_external_tool(simulator_type);
         if (tool == nullptr) {

clients/drcachesim/common/options.cpp

@@ -34,6 +34,7 @@
 
 #include "options.h"
 
+#include <cstdint>
 #include <string>
 
 #include "dr_api.h" // For IF_X86_ELSE.
@@ -460,7 +461,7 @@ droption_t<std::string>
                       "Specifies the types of simulators, separated by a colon (\":\").",
                       "Predefined types: " CPU_CACHE ", " MISS_ANALYZER ", " TLB
                       ", " REUSE_DIST ", " REUSE_TIME ", " HISTOGRAM ", " BASIC_COUNTS
-                      ", or " INVARIANT_CHECKER
+                      ", " INVARIANT_CHECKER ", or " SCHEDULE_STATS
                       ". The external types: name of a tool identified by a "
                       "name.drcachesim config file in the DR tools directory.");
 
@@ -854,5 +855,11 @@ droption_t<std::string>
                          "Path with stored as-traced schedule for replay.");
 #endif
 
+// Schedule_stats options.
+droption_t<uint64_t>
+    op_schedule_stats_print_every(DROPTION_SCOPE_ALL, "schedule_stats_print_every", 5000,
+                                  "A letter is printed every N instrs",
+                                  "A letter is printed every N instrs or N waits");
+
 } // namespace drmemtrace
 } // namespace dynamorio

clients/drcachesim/common/options.h

@@ -35,13 +35,7 @@
 #ifndef _OPTIONS_H_
 #define _OPTIONS_H_ 1
 
-#define REPLACE_POLICY_NON_SPECIFIED ""
-#define REPLACE_POLICY_LRU "LRU"
-#define REPLACE_POLICY_LFU "LFU"
-#define REPLACE_POLICY_FIFO "FIFO"
-#define PREFETCH_POLICY_NEXTLINE "nextline"
-#define PREFETCH_POLICY_NONE "none"
-#define CPU_CACHE "cache"
+// Tool names (for -simulator_type option).
 #define MISS_ANALYZER "miss_analyzer"
 #define TLB "TLB"
 #define HISTOGRAM "histogram"
@@ -53,10 +47,21 @@
 #define VIEW "view"
 #define FUNC_VIEW "func_view"
 #define INVARIANT_CHECKER "invariant_checker"
+#define SCHEDULE_STATS "schedule_stats"
+
+// Constants used by specific tools.
+#define REPLACE_POLICY_NON_SPECIFIED ""
+#define REPLACE_POLICY_LRU "LRU"
+#define REPLACE_POLICY_LFU "LFU"
+#define REPLACE_POLICY_FIFO "FIFO"
+#define PREFETCH_POLICY_NEXTLINE "nextline"
+#define PREFETCH_POLICY_NONE "none"
+#define CPU_CACHE "cache"
 #define CACHE_TYPE_INSTRUCTION "instruction"
 #define CACHE_TYPE_DATA "data"
 #define CACHE_TYPE_UNIFIED "unified"
 #define CACHE_PARENT_MEMORY "memory"
+
 // The expected pattern for a single_op_value is:
 //     function_name|function_id|arguments_num
 // where function_name can contain spaces (for instance, C++ namespace prefix)
@@ -191,6 +196,7 @@ extern dynamorio::droption::droption_t<std::string> op_record_file;
 extern dynamorio::droption::droption_t<std::string> op_replay_file;
 extern dynamorio::droption::droption_t<std::string> op_cpu_schedule_file;
 #endif
+extern dynamorio::droption::droption_t<uint64_t> op_schedule_stats_print_every;
 
 } // namespace drmemtrace
 } // namespace dynamorio

clients/drcachesim/tests/schedule_stats_nopreempt.templatex

@@ -0,0 +1,76 @@
+Schedule stats tool results:
+Total counts:
+           4 cores
+           8 threads
+      638938 instructions
+           5 total context switches
+   0.0078255 CSPKI \(context switches per 1000 instructions\)
+      127788 instructions per context switch
+           5 voluntary context switches
+           0 direct context switches
+      100.00% voluntary switches
+        0.00% direct switches
+         161 system calls
+           2 maybe-blocking system calls
+           0 direct switch requests
+           0 waits
+Core #0 counts:
+           . threads
+      *[0-9]* instructions
+           . total context switches
+   0.0[0-9.]* CSPKI \(context switches per 1000 instructions\)
+      *[0-9]* instructions per context switch
+           . voluntary context switches
+           0 direct context switches
+      100.00% voluntary switches
+        0.00% direct switches
+         *[0-9]* system calls
+           . maybe-blocking system calls
+           0 direct switch requests
+           0 waits
+Core #1 counts:
+           2 threads
+      *[0-9]* instructions
+           . total context switches
+   0.0[0-9.]* CSPKI \(context switches per 1000 instructions\)
+      *[0-9]* instructions per context switch
+           . voluntary context switches
+           0 direct context switches
+      100.00% voluntary switches
+        0.00% direct switches
+          .. system calls
+           . maybe-blocking system calls
+           0 direct switch requests
+           0 waits
+Core #2 counts:
+           2 threads
+      *[0-9]* instructions
+           1 total context switches
+   0.0[0-9.]* CSPKI \(context switches per 1000 instructions\)
+      *[0-9]* instructions per context switch
+           1 voluntary context switches
+           0 direct context switches
+      100.00% voluntary switches
+        0.00% direct switches
+          .. system calls
+           . maybe-blocking system calls
+           0 direct switch requests
+           0 waits
+Core #3 counts:
+           2 threads
+      *[0-9]* instructions
+           1 total context switches
+   0.0[0-9.]* CSPKI \(context switches per 1000 instructions\)
+      *[0-9]* instructions per context switch
+           1 voluntary context switches
+           0 direct context switches
+      100.00% voluntary switches
+        0.00% direct switches
+          .. system calls
+           . maybe-blocking system calls
+           0 direct switch requests
+           0 waits
+Core #0 schedule: [A-H,]*
+Core #1 schedule: [A-H,]*
+Core #2 schedule: [A-H,]*
+Core #3 schedule: [A-H,]*