peer
Back to the main page.
- 42 - maxNumCompThreads not available on all matlab versions
- 109 - compilation problems because inet_ntop is missing on windows
- 110 - tcpserver thread cannot be stopped on windows
- 111 - discover thread does not want to start when compiled on matlab2009+msvc2008
- 135 - allowuser not updated
- 136 - the peer should announce that the tcpserver is NOT running
- 137 - implement udpserver
- 139 - smartcpu is triggered simultaneously on multiple slaves that share a linux node
- 140 - busy peerslave should announce the memreq of the job instead of the memavail
- 141 - there should be a mechanism for logging the output of the command-line peerslaves
- 142 - peer should announce timavail=0 for zombie and master
- 144 - peer slaves seem to switch to zombie while the master is sending a job
- 145 - documentation in peer.m should be more general introduction
- 146 - cmd line peerslave should use syslog() for output, instead of fprintf
- 148 - display error on headless slaves
- 149 - there is no way to cancel jobs that are submitted but not yet completed
- 150 - create a consistent implementation of the timers and timeout options
- 151 - cmd-line peerslave PANICs too easily
- 152 - pass global variables to peerslave
- 153 - implement refuseuser, refusehost, refusegroup
- 154 - add a description field in the host package
- 158 - improve scheduling in peerfeval
- 159 - wrong peerlist
- 162 - fallback to cellfun if there are no slaves
- 163 - using sleep in a thread causes the whole process to sleep
- 164 - peercellfun sends private directories in path, slave gives warning
- 165 - random number generator initializes to the same value on every matlab startup
- 166 - fractional time in seconds is rounded off incorrectly
- 167 - peerlist is empty on stand-alone OS X computer
- 169 - path info not handled properly
- 171 - unknown path during peerexec
- 173 - nice level
- 176 - peercellfun for sourceanalysis
- 183 - include job number in rethrown error inside peerget
- 219 - change syslog output to avoid triggering hobbit/bb
- 223 - peermaster and peerslave don't launch
- 265 - resubmit control for variable compute times
- 268 - remove peer from list in case open_tcp_connection error
- 269 - problem in printing job information
- 275 - timallow/timreq have to be made consistent on master and slave
- 277 - peercellfun resubmission should be improved
- 278 - implement MaxBusy option in peercellfun
- 290 - peercellfun error: unknown function keyvalcheck
- 294 - there are idle slaves with timavail=0
- 295 - resubmitted jobs should be postponed, i.e. first submit all jobs once
- 296 - enginetimeout and zombietimeout should be configurable on the command line
- 299 - add persistent input variable checking to peer/private/getglobal
- 303 - switching from zombie to ilde slave should be delayed, e.g. 5 minutes
- 304 - the user should be allowed to specify the variance of timreq
- 306 - too many resubmissions with small nr of collected jobs
- 309 - fail to submit a job
- 326 - jobs result in "resubmitting because engine takes too much time to start"
- 327 - engines are not efficiently reused
- 330 - mintimreq should be user specified
- 331 - smartcpu should switch off the "smallest" peerslave first
- 332 - the sluggishness in the peerslave smartxxx estimates has to be reimplemented
- 333 - smartcpu switching is inconsistent
- 335 - warning backtrace in peercellfun not functional
- 360 - fix the umask problem with the public peerslaves
- 362 - peer2peer: not starting the engine of one the peerslaves causes peercellfun to given an error when using stoponerror = false
- 378 - rename killswitch into watchdog
- 382 - peerslave syslog output is inconcistent in case of error (job number seems to increment)
- 384 - unpredictable crashese during peer computing
- 422 - the watchdog should also allow for monitoring the memory usage
- 533 - implement RetryOnError
- 546 - matlab startup error
- 548 - incorrect timreq update
- 571 - peer resulting in matlab crash
- 581 - peercellfun should only print "updated timreq/memreq" immediately before submitting a new job
- 588 - strfind error
- 601 - no peerslaves found
- 643 - maintaining the multicast list can be made conditional on the status
- 751 - command line options are not correctly parsed
- 752 - udsserver fails to start on max os x
- 756 - watchdog is not properly enabled in peerslave cmdline executable
- 768 - peermaster allowhost should not be case sensitive
- 781 - Incorrect memavail reading peerslave command line executable
- 789 - peerslave crashes on startup
- 919 - compile getpid.c for different platforms and add to repository in appropriate way
- 920 - ensure walltime is in a format that qsub can understand
- 935 - implement qsublisten function
- 936 - make sure qsubexec's output becomes available in an atomic operation
- 956 - waittime error
- 1319 - implement an engine-based distributed computing toolbox
- 1537 - memtic/memtoc do not always provide accurate results, leading to jobs getting killed
- 2295 - peerlist receives no job-information, causing all jobs to be resubmitted after the set 60 seconds
- 2598 - Peermaster Mex issue
- 3243 - cannot compile "peer" code using Makefile on OSX