boinc/client/sim.C

745 lines
22 KiB
C

// This file is part of BOINC.
// http://boinc.berkeley.edu
// Copyright (C) 2008 University of California
//
// BOINC is free software; you can redistribute it and/or modify it
// under the terms of the GNU Lesser General Public License
// as published by the Free Software Foundation,
// either version 3 of the License, or (at your option) any later version.
//
// BOINC is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
// See the GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with BOINC. If not, see <http://www.gnu.org/licenses/>.
// BOINC client simulator.
//
// usage:
// sim [--duration x] [--delta x] [--dirs dir ...]
// duration = simulation duration (default 86400)
// delta = simulation time step (default 10)
//
// If no dirs are specified:
// reads input files
// sim_projects.xml, sim_host.xml, sim_prefs.xml, cc_config.xml
// and does simulation, generating output files
// sim_log.txt, sim_out.html
//
// If dirs are specified, chdir into each directory in sequence,
// do the above for each one, and write summary info to stdout
#include "error_numbers.h"
#include "str_util.h"
#include "util.h"
#include "log_flags.h"
#include "filesys.h"
#include "network.h"
#include "client_msgs.h"
#include "../sched/edf_sim.h"
#include "sim.h"
#define SCHED_RETRY_DELAY_MIN 60 // 1 minute
#define SCHED_RETRY_DELAY_MAX (60*60*4) // 4 hours
#ifdef _WIN32
#define SIM_EXEC "..\\boincsim"
#else
#define SIM_EXEC "../sim"
#endif
CLIENT_STATE gstate;
NET_STATUS net_status;
bool user_active;
double duration = 86400, delta = 60;
FILE* logfile;
bool running;
double running_time = 0;
bool server_uses_workload = false;
bool dcf_dont_use;
bool dcf_stats;
bool dual_dcf;
bool cpu_sched_rr_only;
bool work_fetch_old;
int line_limit = 1000000;
SIM_RESULTS sim_results;
void SIM_PROJECT::update_dcf_stats(RESULT* rp) {
double raw_ratio = rp->final_cpu_time/rp->estimated_cpu_time_uncorrected();
// see http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Algorithm_III
++completed_task_count;
double delta = raw_ratio - completions_ratio_mean;
completions_ratio_mean += delta / completed_task_count;
completions_ratio_s += delta * ( raw_ratio - completions_ratio_mean);
if (completed_task_count > 1) {
completions_ratio_stdev = sqrt(completions_ratio_s / (completed_task_count - 1));
double required_stdev = (raw_ratio - completions_ratio_mean) / completions_ratio_stdev;
if (required_stdev > completions_required_stdevs) {
completions_required_stdevs = std::min(required_stdev, 7.0);
}
}
duration_correction_factor = completions_ratio_mean +
completions_required_stdevs * completions_ratio_stdev;
return;
}
// generate a job; pick a random app,
// and pick a FLOP count from its distribution
//
void CLIENT_STATE::make_job(SIM_PROJECT* p, WORKUNIT* wup, RESULT* rp) {
SIM_APP* ap1, *ap=0;
double net_fpops = host_info.p_fpops;
double x = drand();
for (unsigned int i=0; i<apps.size();i++) {
ap1 = (SIM_APP*)apps[i];
if (ap1->project != p) continue;
x -= ap1->weight;
if (x <= 0) {
ap = ap1;
break;
}
}
if (!ap) {
printf("ERROR-NO APP\n");
exit(1);
}
rp->clear();
rp->project = p;
rp->wup = wup;
sprintf(rp->name, "%s_%d", p->project_name, p->result_index++);
wup->project = p;
wup->rsc_fpops_est = ap->fpops_est;
double ops = ap->fpops.sample();
if (ops < 0) ops = 0;
rp->final_cpu_time = ops/net_fpops;
rp->report_deadline = now + ap->latency_bound;
}
// process ready-to-report results
//
void CLIENT_STATE::handle_completed_results() {
char buf[256];
vector<RESULT*>::iterator result_iter;
result_iter = results.begin();
while (result_iter != results.end()) {
RESULT* rp = *result_iter;
if (rp->ready_to_report) {
sprintf(buf, "result %s reported; %s<br>",
rp->name,
(gstate.now > rp->report_deadline)?
"<font color=#cc0000>MISSED DEADLINE</font>":
"<font color=#00cc00>MADE DEADLINE</font>"
);
SIM_PROJECT* spp = (SIM_PROJECT*)rp->project;
if (gstate.now > rp->report_deadline) {
sim_results.cpu_wasted += rp->final_cpu_time;
sim_results.nresults_missed_deadline++;
spp->project_results.nresults_missed_deadline++;
spp->project_results.cpu_wasted += rp->final_cpu_time;
} else {
sim_results.cpu_used += rp->final_cpu_time;
sim_results.nresults_met_deadline++;
spp->project_results.nresults_met_deadline++;
spp->project_results.cpu_used += rp->final_cpu_time;
}
gstate.html_msg += buf;
delete rp;
result_iter = results.erase(result_iter);
} else {
result_iter++;
}
}
}
// convert results in progress to IP_RESULTs,
// and get an initial schedule for them
//
void CLIENT_STATE::get_workload(vector<IP_RESULT>& ip_results) {
for (unsigned int i=0; i<results.size(); i++) {
RESULT* rp = results[i];
double x = rp->estimated_cpu_time_remaining(false);
if (x == 0) continue;
IP_RESULT ipr(rp->name, rp->report_deadline, x);
ip_results.push_back(ipr);
}
init_ip_results(work_buf_min(), ncpus, ip_results);
}
bool CLIENT_STATE::simulate_rpc(PROJECT* _p) {
char buf[256];
SIM_PROJECT* p = (SIM_PROJECT*) _p;
static double last_time=0;
vector<IP_RESULT> ip_results;
int infeasible_count = 0;
double diff = now - last_time;
if (diff && diff < host_info.connection_interval) {
msg_printf(NULL, MSG_INFO, "simulate_rpc: too soon %f < %f", diff, host_info.connection_interval);
return false;
}
last_time = now;
sprintf(buf, "RPC to %s; asking for %f<br>", p->project_name, p->work_request);
html_msg += buf;
handle_completed_results();
if (server_uses_workload) {
get_workload(ip_results);
}
bool sent_something = false;
double work_left = p->work_request;
while (work_left > 0) {
RESULT* rp = new RESULT;
WORKUNIT* wup = new WORKUNIT;
make_job(p, wup, rp);
if (server_uses_workload) {
IP_RESULT c(rp->name, rp->report_deadline, rp->final_cpu_time);
if (check_candidate(c, ncpus, ip_results)) {
ip_results.push_back(c);
} else {
delete rp;
delete wup;
if (++infeasible_count > p->max_infeasible_count) {
p->min_rpc_time = now + 1;
break;
}
}
}
sent_something = true;
rp->set_state(RESULT_FILES_DOWNLOADED, "simulate_rpc");
results.push_back(rp);
sprintf(buf, "got job %s: CPU time %.2f, deadline %s<br>",
rp->name, rp->final_cpu_time, time_to_string(rp->report_deadline)
);
html_msg += buf;
work_left -= p->duration_correction_factor*wup->rsc_fpops_est/host_info.p_fpops;
}
if (p->work_request > 0 && !sent_something) {
p->backoff();
} else {
p->nrpc_failures = 0;
}
p->work_request = 0;
request_schedule_cpus("simulate_rpc");
request_work_fetch("simulate_rpc");
return true;
}
void SIM_PROJECT::backoff() {
nrpc_failures++;
double backoff = calculate_exponential_backoff(
nrpc_failures, SCHED_RETRY_DELAY_MIN, SCHED_RETRY_DELAY_MAX
);
min_rpc_time = gstate.now + backoff;
}
bool CLIENT_STATE::scheduler_rpc_poll() {
PROJECT *p;
msg_printf(NULL, MSG_INFO, "RPC poll start");
p = next_project_sched_rpc_pending();
if (p) {
return simulate_rpc(p);
}
p = find_project_with_overdue_results();
if (p) {
return simulate_rpc(p);
}
p = next_project_need_work();
if (p) {
return simulate_rpc(p);
}
msg_printf(NULL, MSG_INFO, "RPC poll: nothing to do");
return false;
}
bool ACTIVE_TASK_SET::poll() {
unsigned int i;
char buf[256];
bool action = false;
static double last_time = 0;
double diff = gstate.now - last_time;
if (diff < 1.0) return false;
last_time = gstate.now;
SIM_PROJECT* p;
if (!running) return false;
for (i=0; i<gstate.projects.size(); i++) {
p = (SIM_PROJECT*) gstate.projects[i];
p->idle = true;
sprintf(buf, "%s STD: %f min RPC<br>",
p->project_name, p->short_term_debt,
time_to_string(p->min_rpc_time)
);
gstate.html_msg += buf;
}
int n=0;
for (i=0; i<active_tasks.size(); i++) {
ACTIVE_TASK* atp = active_tasks[i];
switch (atp->task_state()) {
case PROCESS_EXECUTING:
atp->cpu_time_left -= diff;
atp->current_cpu_time += diff;
RESULT* rp = atp->result;
double cpu_time_used = rp->final_cpu_time - atp->cpu_time_left;
atp->fraction_done = cpu_time_used/rp->final_cpu_time;
atp->checkpoint_wall_time = gstate.now;
if (atp->cpu_time_left <= 0) {
atp->set_task_state(PROCESS_EXITED, "poll");
rp->exit_status = 0;
rp->ready_to_report = true;
gstate.request_schedule_cpus("ATP poll");
gstate.request_work_fetch("ATP poll");
sprintf(buf, "result %s finished<br>", rp->name);
gstate.html_msg += buf;
action = true;
}
((SIM_PROJECT*)rp->project)->idle = false;
n++;
}
}
if (n < gstate.ncpus) {
sim_results.cpu_idle += diff*(gstate.ncpus-n);
}
if (n > gstate.ncpus) {
sprintf(buf, "TOO MANY JOBS RUNNING");
gstate.html_msg += buf;
}
for (i=0; i<gstate.projects.size(); i++) {
p = (SIM_PROJECT*) gstate.projects[i];
if (p->idle) {
p->idle_time += diff;
p->idle_time_sumsq += diff*(p->idle_time*p->idle_time);
} else {
p->idle_time = 0;
}
}
running_time += diff;
return action;
}
int SIM_APP::parse(XML_PARSER& xp) {
char tag[256];
bool is_tag;
int retval;
weight = 1;
while(!xp.get(tag, sizeof(tag), is_tag)) {
if (!is_tag) return ERR_XML_PARSE;
if (!strcmp(tag, "/app")) {
return 0;
}
else if (xp.parse_double(tag, "latency_bound", latency_bound)) continue;
else if (xp.parse_double(tag, "fpops_est", fpops_est)) continue;
else if (xp.parse_double(tag, "weight", weight)) continue;
else if (!strcmp(tag, "fpops")) {
retval = fpops.parse(xp, "/fpops");
if (retval) return retval;
} else if (!strcmp(tag, "checkpoint_period")) {
retval = checkpoint_period.parse(xp, "/checkpoint_period");
if (retval) return retval;
} else if (xp.parse_double(tag, "working_set", working_set)) continue;
else {
printf("unrecognized: %s\n", tag);
return ERR_XML_PARSE;
}
}
return ERR_XML_PARSE;
}
// return the fraction of CPU time that was spent in violation of shares
// i.e., if a project got X and it should have got Y,
// add up |X-Y| over all projects, and divide by total CPU
//
double CLIENT_STATE::share_violation() {
unsigned int i;
double tot = 0, trs=0;
for (i=0; i<projects.size(); i++) {
SIM_PROJECT* p = (SIM_PROJECT*) projects[i];
tot += p->project_results.cpu_used + p->project_results.cpu_wasted;
trs += p->resource_share;
}
double sum = 0;
for (i=0; i<projects.size(); i++) {
SIM_PROJECT* p = (SIM_PROJECT*) projects[i];
double t = p->project_results.cpu_used + p->project_results.cpu_wasted;
double rs = p->resource_share/trs;
double rt = tot*rs;
sum += fabs(t - rt);
}
return sum/tot;
}
// "monotony" is defined as follows:
// for each project P, maintain R(P), the time since P last ran,
// let S(P) be the RMS of R(P).
// Let X = mean(S(P))/(sched_interval*nprojects)
// (the *nprojects reflects the fact that in the limit of nprojects,
// each one waits for a time to run proportional to nprojects)
// X varies from zero (no monotony) to infinity.
// X is one in the case of round-robin on 1 CPU.
// Let monotony = 1-1/(x+1)
//
double CLIENT_STATE::monotony() {
double sum = 0;
double schedint = global_prefs.cpu_scheduling_period();
unsigned int i;
for (i=0; i<projects.size(); i++) {
SIM_PROJECT* p = (SIM_PROJECT*) projects[i];
double avg_ss = p->idle_time_sumsq/running_time;
double s = sqrt(avg_ss);
sum += s;
}
int n = (int)projects.size();
double x = sum/(n*schedint*n);
double m = 1-(1/(x+1));
//printf("sum: %f; x: %f m: %f\n", sum, x, m);
return m;
}
// the CPU totals are there; compute the other fields
//
void SIM_RESULTS::compute() {
double total = cpu_used + cpu_wasted + cpu_idle;
cpu_wasted_frac = cpu_wasted/total;
cpu_idle_frac = cpu_idle/total;
share_violation = gstate.share_violation();
monotony = gstate.monotony();
}
// top-level results (for aggregating multiple simulations)
//
void SIM_RESULTS::print(FILE* f, const char* title) {
if (title) {
fprintf(f, "%s: ", title);
}
fprintf(f, "wasted_frac %f idle_frac %f share_violation %f monotony %f\n",
cpu_wasted_frac, cpu_idle_frac, share_violation, monotony
);
}
void SIM_RESULTS::parse(FILE* f) {
fscanf(f, "wasted_frac %lf idle_frac %lf share_violation %lf monotony %lf",
&cpu_wasted_frac, &cpu_idle_frac, &share_violation, &monotony
);
}
void SIM_RESULTS::add(SIM_RESULTS& r) {
cpu_wasted_frac += r.cpu_wasted_frac;
cpu_idle_frac += r.cpu_idle_frac;
share_violation += r.share_violation;
monotony += r.monotony;
}
void SIM_RESULTS::divide(int n) {
cpu_wasted_frac /= n;
cpu_idle_frac /= n;
share_violation /= n;
monotony /= n;
}
void SIM_RESULTS::clear() {
memset(this, 0, sizeof(*this));
}
void SIM_PROJECT::print_results(FILE* f, SIM_RESULTS& sr) {
double t = project_results.cpu_used + project_results.cpu_wasted;
double gt = sr.cpu_used + sr.cpu_wasted;
fprintf(f, "%s: share %.2f total CPU %2f (%.2f%%)\n"
" used %.2f wasted %.2f\n"
" met %d missed %d\n",
project_name, resource_share,
t, (t/gt)*100,
project_results.cpu_used,
project_results.cpu_wasted,
project_results.nresults_met_deadline,
project_results.nresults_missed_deadline
);
}
char* colors[] = {
"#ffffdd",
"#ffddff",
"#ddffff",
"#ddffdd",
"#ddddff",
"#ffdddd",
};
static int outfile_num=0;
void CLIENT_STATE::html_start(bool show_prev) {
char buf[256];
sprintf(buf, "sim_out_%d.html", outfile_num++);
html_out = fopen(buf, "w");
setbuf(html_out, 0);
fprintf(html_out, "<h2>Simulator output</h2>\n");
if (show_prev) {
fprintf(html_out,
"<a href=sim_out_%d.html>Previous file</a><p>\n",
outfile_num-2
);
}
fprintf(html_out,
"<a href=sim_log.txt>message log</a><p>"
"<table border=1><tr><th>Time</th>\n"
);
for (int i=0; i<ncpus; i++) {
fprintf(html_out,
"<th>CPU %d<br><font size=-2>Job name and estimated time left<br>color denotes project<br>* means EDF mode</font></th>", i
);
}
fprintf(html_out, "<th>Notes</th></tr>\n");
}
void CLIENT_STATE::html_rec() {
static int line_num=0;
fprintf(html_out, "<tr><td>%s</td>", time_to_string(now));
if (!running) {
for (int j=0; j<ncpus; j++) {
fprintf(html_out, "<td bgcolor=#aaaaaa>OFF</td>");
}
} else {
int n=0;
for (unsigned int i=0; i<active_tasks.active_tasks.size(); i++) {
ACTIVE_TASK* atp = active_tasks.active_tasks[i];
if (atp->task_state() == PROCESS_EXECUTING) {
SIM_PROJECT* p = (SIM_PROJECT*)atp->result->project;
fprintf(html_out, "<td bgcolor=%s>%s%s: %.2f</td>",
colors[p->index],
atp->result->rr_sim_misses_deadline?"*":"",
atp->result->name, atp->cpu_time_left
);
n++;
}
}
while (n<ncpus) {
fprintf(html_out, "<td>IDLE</td>");
n++;
}
}
fprintf(html_out, "<td><font size=-2>%s</font></td></tr>\n", html_msg.c_str());
html_msg = "";
if (++line_num == line_limit) {
line_num = 0;
html_end(true);
html_start(true);
}
}
void CLIENT_STATE::html_end(bool show_next) {
fprintf(html_out, "</table>");
if (show_next) {
fprintf(html_out,
"<p><a href=sim_out_%d.html>Next file</a>\n",
outfile_num
);
} else {
fprintf(html_out, "<pre>\n");
sim_results.compute();
sim_results.print(html_out);
print_project_results(html_out);
fprintf(html_out, "</pre>\n");
}
if (show_next) {
fprintf(html_out, "<p><a href=sim_out_last.html>Last file</a>\n");
} else {
char buf[256];
sprintf(buf, "sim_out_%d.html", outfile_num-1);
#ifndef _WIN32
symlink(buf, "sim_out_last.html");
#endif
}
fclose(html_out);
}
void CLIENT_STATE::simulate() {
bool action;
now = 0;
html_start(false);
while (1) {
running = host_info.available.sample(now);
while (1) {
action = active_tasks.poll();
if (running) {
action |= handle_finished_apps();
action |= possibly_schedule_cpus();
action |= enforce_schedule();
action |= compute_work_requests();
action |= scheduler_rpc_poll();
}
if (!action) break;
}
now += delta;
html_rec();
if (now > duration) break;
}
html_end(false);
}
void parse_error(char* file, int retval) {
printf("can't parse %s: %d\n", file, retval);
exit(1);
}
void help(char* prog) {
fprintf(stderr, "usage: %s\n"
"[--duration X]\n"
"[--delta X]\n"
"[--server_uses_workload]\n"
"[--dcf_dont_user]\n"
"[--dcf_stats]\n"
"[--dual_dcf]\n"
"[--cpu_sched_rr_only]\n"
"[--work_fetch_old]\n"
"[--dirs ...]\n",
prog
);
exit(1);
}
char* next_arg(int argc, char** argv, int& i) {
if (i >= argc) {
fprintf(stderr, "Missing command-line argument\n");
help(argv[0]);
}
return argv[i++];
}
#define PROJECTS_FILE "sim_projects.xml"
#define HOST_FILE "sim_host.xml"
#define PREFS_FILE "sim_prefs.xml"
#define SUMMARY_FILE "sim_summary.txt"
#define LOG_FILE "sim_log.txt"
int main(int argc, char** argv) {
int i, retval;
vector<std::string> dirs;
logfile = fopen("sim_log.txt", "w");
sim_results.clear();
for (i=1; i<argc;) {
char* opt = argv[i++];
if (!strcmp(opt, "--duration")) {
duration = atof(next_arg(argc, argv, i));
} else if (!strcmp(opt, "--delta")) {
delta = atof(next_arg(argc, argv, i));
} else if (!strcmp(opt, "--dirs")) {
while (i<argc) {
dirs.push_back(argv[i++]);
}
} else if (!strcmp(opt, "--server_uses_workload")) {
server_uses_workload = true;
} else if (!strcmp(opt, "--dcf_dont_use")) {
dcf_dont_use = true;
} else if (!strcmp(opt, "--dcf_stats")) {
dcf_stats = true;
} else if (!strcmp(opt, "--dual_dcf")) {
dual_dcf = true;
dcf_stats = true;
} else if (!strcmp(opt, "--cpu_sched_rr_only")) {
cpu_sched_rr_only = true;
} else if (!strcmp(opt, "--work_fetch_old")) {
work_fetch_old = true;
} else if (!strcmp(opt, "--line_limit")) {
line_limit = atoi(next_arg(argc, argv, i));
} else {
help(argv[0]);
}
}
if (duration <= 0) {
printf("non-pos duration\n");
exit(1);
}
if (delta <= 0) {
printf("non-pos delta\n");
exit(1);
}
if (dirs.size()) {
// If we need to do several simulations,
// use system() to do each one in a separate process,
// because there are lots of static variables and we need to ensure
// that they start off with the right initial values
//
unsigned int i;
SIM_RESULTS total_results;
total_results.clear();
for (i=0; i<dirs.size(); i++) {
std::string dir = dirs[i];
retval = chdir(dir.c_str());
if (retval) {
fprintf(stderr, "can't chdir into %s: ", dir.c_str());
perror("chdir");
continue;
}
char buf[256];
sprintf(
buf, "%s --duration %f --delta %f > %s",
SIM_EXEC, duration, delta, SUMMARY_FILE
);
retval = system(buf);
if (retval) {
printf("simulation in %s failed\n", dir.c_str());
exit(1);
}
FILE* f = fopen(SUMMARY_FILE, "r");
sim_results.parse(f);
fclose(f);
sim_results.print(stdout, dir.c_str());
total_results.add(sim_results);
chdir("..");
}
total_results.divide((int)(dirs.size()));
total_results.print(stdout, "Total");
} else {
read_config_file();
int retval;
bool flag;
retval = gstate.parse_projects(PROJECTS_FILE);
if (retval) parse_error(PROJECTS_FILE, retval);
retval = gstate.parse_host(HOST_FILE);
if (retval) parse_error(HOST_FILE, retval);
retval = gstate.global_prefs.parse_file(PREFS_FILE, "", flag);
if (retval) parse_error(PREFS_FILE, retval);
gstate.set_ncpus();
gstate.request_work_fetch("init");
gstate.simulate();
sim_results.compute();
// print machine-readable first
sim_results.print(stdout);
// then other
gstate.print_project_results(stdout);
}
}