#include "os_binary.hpp"
#include "os_time.hpp"
+#include "os_thread.hpp"
#include "image.hpp"
#include "trace_callset.hpp"
#include "trace_dump.hpp"
#include "retrace.hpp"
-static bool wait = false;
+static bool waitOnFinish = false;
static const char *comparePrefix = NULL;
static const char *snapshotPrefix = NULL;
static unsigned dumpStateCallNo = ~0;
+retrace::Retracer retracer;
+
namespace retrace {
trace::Parser parser;
+trace::Profiler profiler;
int verbosity = 0;
bool debug = true;
-bool profiling = false;
-
+bool dumpingState = false;
bool doubleBuffer = true;
bool coreProfile = false;
+bool profiling = false;
+bool profilingGpuTimes = false;
+bool profilingCpuTimes = false;
+bool profilingPixelsDrawn = false;
-static unsigned frameNo = 0;
+unsigned frameNo = 0;
+unsigned callNo = 0;
void
}
+/**
+ * Take/compare snapshots.
+ */
static void
takeSnapshot(unsigned call_no) {
assert(snapshotPrefix || comparePrefix);
}
+/**
+ * Retrace one call.
+ *
+ * Take snapshots before/after retracing (as appropriate) and dispatch it to
+ * the respective handler.
+ */
static void
-mainLoop() {
- retrace::Retracer retracer;
+retraceCall(trace::Call *call) {
+ bool swapRenderTarget = call->flags &
+ trace::CALL_FLAG_SWAP_RENDERTARGET;
+ bool doSnapshot = snapshotFrequency.contains(*call) ||
+ compareFrequency.contains(*call);
+
+ // For calls which cause rendertargets to be swaped, we take the
+ // snapshot _before_ swapping the rendertargets.
+ if (doSnapshot && swapRenderTarget) {
+ if (call->flags & trace::CALL_FLAG_END_FRAME) {
+ // For swapbuffers/presents we still use this
+ // call number, spite not have been executed yet.
+ takeSnapshot(call->no);
+ } else {
+ // Whereas for ordinate fbo/rendertarget changes we
+ // use the previous call's number.
+ takeSnapshot(call->no - 1);
+ }
+ }
- addCallbacks(retracer);
+ callNo = call->no;
+ retracer.retrace(*call);
- long long startTime = 0;
+ if (doSnapshot && !swapRenderTarget)
+ takeSnapshot(call->no);
- startTime = os::getTime();
- trace::Call *call;
+ if (call->no >= dumpStateCallNo &&
+ dumpState(std::cout)) {
+ exit(0);
+ }
+}
- while ((call = retrace::parser.parse_call())) {
- bool swapRenderTarget = call->flags & trace::CALL_FLAG_SWAP_RENDERTARGET;
- bool doSnapshot =
- snapshotFrequency.contains(*call) ||
- compareFrequency.contains(*call)
- ;
-
- // For calls which cause rendertargets to be swaped, we take the
- // snapshot _before_ swapping the rendertargets.
- if (doSnapshot && swapRenderTarget) {
- if (call->flags & trace::CALL_FLAG_END_FRAME) {
- // For swapbuffers/presents we still use this call number,
- // spite not have been executed yet.
- takeSnapshot(call->no);
- } else {
- // Whereas for ordinate fbo/rendertarget changes we use the
- // previous call's number.
- takeSnapshot(call->no - 1);
+
+class RelayRunner;
+
+
+/**
+ * Implement multi-threading by mimicking a relay race.
+ */
+class RelayRace
+{
+private:
+ /**
+ * Runners indexed by the leg they run (i.e, the thread_ids from the
+ * trace).
+ */
+ std::vector<RelayRunner*> runners;
+
+public:
+ RelayRace();
+
+ ~RelayRace();
+
+ RelayRunner *
+ getRunner(unsigned leg);
+
+ inline RelayRunner *
+ getForeRunner() {
+ return getRunner(0);
+ }
+
+ void
+ run(void);
+
+ void
+ passBaton(trace::Call *call);
+
+ void
+ finishLine();
+
+ void
+ stopRunners();
+};
+
+
+/**
+ * Each runner is a thread.
+ *
+ * The fore runner doesn't have its own thread, but instead uses the thread
+ * where the race started.
+ */
+class RelayRunner
+{
+private:
+ friend class RelayRace;
+
+ RelayRace *race;
+
+ unsigned leg;
+
+ os::mutex mutex;
+ os::condition_variable wake_cond;
+
+ /**
+ * There are protected by the mutex.
+ */
+ bool finished;
+ trace::Call *baton;
+
+ os::thread thread;
+
+ static void *
+ runnerThread(RelayRunner *_this);
+
+public:
+ RelayRunner(RelayRace *race, unsigned _leg) :
+ race(race),
+ leg(_leg),
+ finished(false),
+ baton(0)
+ {
+ /* The fore runner does not need a new thread */
+ if (leg) {
+ thread = os::thread(runnerThread, this);
+ }
+ }
+
+ /**
+ * Thread main loop.
+ */
+ void
+ runRace(void) {
+ os::unique_lock<os::mutex> lock(mutex);
+
+ while (1) {
+ while (!finished && !baton) {
+ wake_cond.wait(lock);
}
+
+ if (finished) {
+ break;
+ }
+
+ assert(baton);
+ trace::Call *call = baton;
+ baton = 0;
+
+ runLeg(call);
}
- retracer.retrace(*call);
+ if (0) std::cerr << "leg " << leg << " actually finishing\n";
- if (doSnapshot && !swapRenderTarget) {
- takeSnapshot(call->no);
+ if (leg == 0) {
+ race->stopRunners();
}
+ }
- if (call->no >= dumpStateCallNo &&
- dumpState(std::cout)) {
- exit(0);
+ /**
+ * Interpret successive calls.
+ */
+ void
+ runLeg(trace::Call *call) {
+ /* Consume successive calls for this thread. */
+ do {
+ assert(call);
+ assert(call->thread_id == leg);
+ retraceCall(call);
+ delete call;
+ call = parser.parse_call();
+ } while (call && call->thread_id == leg);
+
+ if (call) {
+ /* Pass the baton */
+ assert(call->thread_id != leg);
+ flushRendering();
+ race->passBaton(call);
+ } else {
+ /* Reached the finish line */
+ if (0) std::cerr << "finished on leg " << leg << "\n";
+ if (leg) {
+ /* Notify the fore runner */
+ race->finishLine();
+ } else {
+ /* We are the fore runner */
+ finished = true;
+ }
}
+ }
+
+ /**
+ * Called by other threads when relinquishing the baton.
+ */
+ void
+ receiveBaton(trace::Call *call) {
+ assert (call->thread_id == leg);
- delete call;
+ mutex.lock();
+ baton = call;
+ mutex.unlock();
+
+ wake_cond.signal();
}
- // Reached the end of trace
- flushRendering();
+ /**
+ * Called by the fore runner when the race is over.
+ */
+ void
+ finishRace() {
+ if (0) std::cerr << "notify finish to leg " << leg << "\n";
+
+ mutex.lock();
+ finished = true;
+ mutex.unlock();
+
+ wake_cond.signal();
+ }
+};
+
+
+void *
+RelayRunner::runnerThread(RelayRunner *_this) {
+ _this->runRace();
+ return 0;
+}
+
+
+RelayRace::RelayRace() {
+ runners.push_back(new RelayRunner(this, 0));
+}
+
+
+RelayRace::~RelayRace() {
+ assert(runners.size() >= 1);
+ std::vector<RelayRunner*>::const_iterator it;
+ for (it = runners.begin(); it != runners.end(); ++it) {
+ RelayRunner* runner = *it;
+ if (runner) {
+ delete runner;
+ }
+ }
+}
+
+
+/**
+ * Get (or instantiate) a runner for the specified leg.
+ */
+RelayRunner *
+RelayRace::getRunner(unsigned leg) {
+ RelayRunner *runner;
+
+ if (leg >= runners.size()) {
+ runners.resize(leg + 1);
+ runner = 0;
+ } else {
+ runner = runners[leg];
+ }
+ if (!runner) {
+ runner = new RelayRunner(this, leg);
+ runners[leg] = runner;
+ }
+ return runner;
+}
+
+
+/**
+ * Start the race.
+ */
+void
+RelayRace::run(void) {
+ trace::Call *call;
+ call = parser.parse_call();
+ if (!call) {
+ /* Nothing to do */
+ return;
+ }
+
+ RelayRunner *foreRunner = getForeRunner();
+ if (call->thread_id == 0) {
+ /* We are the forerunner thread, so no need to pass baton */
+ foreRunner->baton = call;
+ } else {
+ passBaton(call);
+ }
+
+ /* Start the forerunner thread */
+ foreRunner->runRace();
+}
+
+
+/**
+ * Pass the baton (i.e., the call) to the appropriate thread.
+ */
+void
+RelayRace::passBaton(trace::Call *call) {
+ if (0) std::cerr << "switching to thread " << call->thread_id << "\n";
+ RelayRunner *runner = getRunner(call->thread_id);
+ runner->receiveBaton(call);
+}
+
+
+/**
+ * Called when a runner other than the forerunner reaches the finish line.
+ *
+ * Only the fore runner can finish the race, so inform him that the race is
+ * finished.
+ */
+void
+RelayRace::finishLine(void) {
+ RelayRunner *foreRunner = getForeRunner();
+ foreRunner->finishRace();
+}
+
+
+/**
+ * Called by the fore runner after finish line to stop all other runners.
+ */
+void
+RelayRace::stopRunners(void) {
+ std::vector<RelayRunner*>::const_iterator it;
+ for (it = runners.begin() + 1; it != runners.end(); ++it) {
+ RelayRunner* runner = *it;
+ if (runner) {
+ runner->finishRace();
+ }
+ }
+}
+
+
+static void
+mainLoop() {
+ addCallbacks(retracer);
+
+ long long startTime = 0;
+ frameNo = 0;
+
+ startTime = os::getTime();
+
+ RelayRace race;
+ race.run();
long long endTime = os::getTime();
float timeInterval = (endTime - startTime) * (1.0 / os::timeFrequency);
" average of " << (frameNo/timeInterval) << " fps\n";
}
- if (wait) {
+ if (waitOnFinish) {
waitForInput();
} else {
- exit(0);
+ return;
}
}
static void
usage(const char *argv0) {
std::cout <<
- "Usage: " << argv0 << " [OPTION] TRACE\n"
+ "Usage: " << argv0 << " [OPTION] TRACE [...]\n"
"Replay TRACE.\n"
"\n"
" -b benchmark mode (no error checking or warning messages)\n"
- " -p profiling mode (run whole trace, dump profiling info)\n"
+ " -pcpu cpu profiling (cpu times per call)\n"
+ " -pgpu gpu profiling (gpu times per draw call)\n"
+ " -ppd pixels drawn profiling (pixels drawn per draw call)\n"
" -c PREFIX compare against snapshots\n"
" -C CALLSET calls to compare (default is every frame)\n"
" -core use core profile\n"
" -S CALLSET calls to snapshot (default is every frame)\n"
" -v increase output verbosity\n"
" -D CALLNO dump state at specific call no\n"
- " -w wait on final frame\n";
+ " -w waitOnFinish on final frame\n";
}
extern "C"
int main(int argc, char **argv)
{
+ using namespace retrace;
+
assert(compareFrequency.empty());
assert(snapshotFrequency.empty());
} else if (!strcmp(arg, "-b")) {
retrace::debug = false;
retrace::verbosity = -1;
- } else if (!strcmp(arg, "-p")) {
- retrace::debug = false;
- retrace::profiling = true;
- retrace::verbosity = -1;
} else if (!strcmp(arg, "-c")) {
comparePrefix = argv[++i];
if (compareFrequency.empty()) {
}
} else if (!strcmp(arg, "-D")) {
dumpStateCallNo = atoi(argv[++i]);
+ dumpingState = true;
retrace::verbosity = -2;
} else if (!strcmp(arg, "-core")) {
retrace::coreProfile = true;
} else if (!strcmp(arg, "-v")) {
++retrace::verbosity;
} else if (!strcmp(arg, "-w")) {
- wait = true;
+ waitOnFinish = true;
+ } else if (arg[1] == 'p') {
+ retrace::debug = false;
+ retrace::profiling = true;
+ retrace::verbosity = -1;
+
+ if (!strcmp(arg, "-pcpu")) {
+ retrace::profilingCpuTimes = true;
+ } else if (!strcmp(arg, "-pgpu")) {
+ retrace::profilingGpuTimes = true;
+ } else if (!strcmp(arg, "-ppd")) {
+ retrace::profilingPixelsDrawn = true;
+ }
} else {
std::cerr << "error: unknown option " << arg << "\n";
usage(argv[0]);
}
retrace::setUp();
+ if (retrace::profiling) {
+ retrace::profiler.setup(retrace::profilingCpuTimes, retrace::profilingGpuTimes, retrace::profilingPixelsDrawn);
+ }
for ( ; i < argc; ++i) {
if (!retrace::parser.open(argv[i])) {
retrace::parser.close();
}
- retrace::cleanUp();
+ // XXX: X often hangs on XCloseDisplay
+ //retrace::cleanUp();
return 0;
}