BOINC 4.09 is offiziel zum download freigegeben
- Ersteller Sir Ulli
- Erstellt am
Sir Ulli
Grand Admiral Special
zu finden wie immer
http://setiweb.ssl.berkeley.edu/sah/download.php
änderungen soweit bekannt
BOINC 4.09 release
I'm hoping this will be good enough for public deployment...
----- Rom
David 6 Sept 2004
- The Windows GUI and CLI versions were using different mechanisms
to ensure that only one instance of BOINC is running.
Factored this into a function check_unique_instance(),
which is used by all versions.
client/
file_names.C,h
main.C
win/
wingui_mainwindow.cpp
lib/
error_numbers.h
David 7 Sept 2004
- (from Ben Herndon) Some compilers optimize away sections of the
Whetstone
benchmark function, since their results aren't used outside
the function.
To deal with this, we added an global-scope array
(extern_array[])
and store the result of each section into it.
Also changed section 3 by initializint j from the
extern_array,
so the compiler can't optimize out all the conditionals.
client/
whetstone.C
David 7 Sept 2004
- change "upgrade" to query user before overwriting files
tools/
upgrade
David 7 Sept 2004
- recognize CPU type on HP (DEC) alpha (from Duane Krahn)
client/
hostinfo_unix.C
Rom 8 Sep 2004
- If we can lock the lockfile, also aquire the global mutex. If we
don't the screensaver
cannot properly detect that BOINC is executing.
client/
file_names.C
Jeff 09 Sep 2004
- Change the validator to handle the case where one or more
of the result files for a WU's result set can't be read.
Currently, this case results in this WU and its results
never getting looked at again - no science, no credit.
We distinguish 2 cases:
1) transient failure, e.g. an NFS mount is not working.
(can't read directory)
In this case check_set() returns a special code to the validator,
which arranges for the WU to get transitioned again in a few hours.
2) permanent failure (file doesn't exist)
The result's validate_state is set to ERROR (new state),
which is generally equivalent to INVALID
- add new WU arg to check_set()
(so it can access min_quorum)
This is coordinated seti_boinc / boinc change.
This note is repeated in the seti_boinc checkin_notes.
seti_boinc changes:
- get_result_file(), if it cannot read the result file,
tries to read the directory where the result file should be.
If it can read the directory, it returns ERR_FOPEN, as before.
If it cannot read the directory, it resturns ERR_OPENDIR.
ERR_OPENDIR in this context signals a possibly transient problem.
Upon a successful result read, it sets sah_result.have_result
(a new data member as of this change) to true.
- check_set(), upon return from get_result_file() :
- sets result.validate_state to VALIDATE_STATE_ERROR (a new
state)
and retval to zero if there is a nonzero retval from
check_set()
and it is not ERR_OPENDIR. It then continues.
- otherwise just continues.
check_set() then determines if any IO errors brought the result
count
below wu.min_quorum. If not, it continues. Otherwise it returns.
In all subsequent logic any results for which
sah_result.have_result == false are skipped.
Note that the result and the sah_result vectors are associated.
validate/
sah_result.h
sah_validate.cpp
sah_boinc_db.cpp
boinc changes:
- new validate state VALIDATE_STATE_ERROR.
- in the enumeration for check_set(),
we now include a clause to check for VALIDATE_STATE_INIT
in order to ignore results set to VALIDATE_STATE_ERROR.
This enumeration query now matches that for check_pair().
Note that the only possible retvals from check_set()
are now zero or ERR_OPENDIR.
Note that logic to retry possibly transient errors still needs
to be done.
Also left to be done is a transitioner change to subtract
the number of results in state VALIDATE_STATE_ERROR
from the count of active results.
The latter change is needed to stimulate the production
of additional results for the affected WU.
sched/
validator.C
db/
boinc_db.h
Rom 9 Sep 2004 ( for Sebastian Masch )
- Don't cause the core client to rewrite the state file
for every block transfered during a network operation.
client/
pers_file_xfer.C
David 9 Sept 2004
- Further refinements of validation to deal with
recoverable/nonrecoverable file read errors.
Revised the signatures of check_set() and check_pair(),
and revised the validator accordingly.
db/
boinc.db.h
sched/
sample_bitwise_validator.C
sample_trivial_validator.C
validate_util.C,h
validator.C
David 9 Sept 2004
- added try_fopen() (from Jeff Cobb):
like fopen(), but on tells you whether it's recoverable or not
lib/
util.C,h
David 9 Sept 2004
- Added support for stats access via WAP (for cell phones)
(from Carl Christensen)
David 9 Sept 2004
- In transitioner, count RESULT_OUTCOME_VALIDATE_ERROR as an error
sched/
transitioner.C
David 9 Sept 2004
- preliminary support for "locality scheduling" (for Einstein@home).
David 10 Sept 2004
- compile fix for Win
lib/
util.C
David 10 Sept 2004
- validator: slight change to check_set() and check_pair():
they no longer should update result records in case of
nonrecoverable error.
Just set the outcome to RESULT_OUTCOME_VALIDATE_ERROR,
and the framework program will do the updates.
sched/
validator.C
David 10 Sept 2004
- more changes for locality scheduling
Rom 10 Sept 2004
- NEw GUI: Put the glue back in so that we can make the various tabs
work.
clientgui/
MainDocument.cpp, .h
stdwx.h
Lana 26 August 2004
- Installation script provides better default daemons by starting
transitioner, feeder and file deleter deamons automatically
tools/
make_project
Lana 11 Sept 2004
- Created DB Purger that archives "DONE" workunits and their
results into XML archive files. Deletes those records from the
database.
Added delete_from_db function to DB_BASE
sched/
db_purge.C
db/
db_base.h
db_base.C
David 12 Sept 2004
- Added a queueing mechanism for shared-memory message passing,
class MSG_QUEUE.
This provides reliable, ordered message delivery.
I think process-control messages were being lost in some
cases.
- Replaced ad-hoc 1-level queueing mechanism for graphics requests
- Changed the logic for selecting app for screensaver a bit;
have new field bool ACTIVE_TASK::is_ss_app that keeps track
of which app is acting as the screensaver.
This fixes bugs in a few particular cases.
David 13 Sept 2004
- Preliminary checkin for "GUI URLs" feature
(allows projects to send URLs to be use as links in GUIs)
This is implemented in a new class GUI_URLS.
Its init() member reads template file.
Its get_gui_urls() member macro-substitutes user/team/host IDs
into the URLs, and produces XML to be included in scheduler replies.
- renamed replace_element() to replace_element_contents()
- moved try_fopen() from lib/util.C to sched/sched_util.C
- added old and new docs to CVS
David 13 Sept 2004
- When email account ID, show (on the next web page) the "from"
address of the email,
since some users may have email filters based on sender.
The "from" address is the constant EMAIL_FROM,
which should be defined in project.inc.
If this is missing, it uses URL_BASE
(which is not an email address, so projects should define
EMAIL_FROM)
David 13 Sept 2004
- small changes to screensaver logic, try to fix Rom's 2-CPU bug
client/
app_graphics.C
ss_logic.C
David 13 Sept 2004
- Change CPU scheduling logic so that the min debt is subtracted
from the debt of all active projects.
i.e. debt is normalized to be zero-based.
This fixes a bug where the debt of projects with work
keeps getting bigger and bigger,
so that when a new project is attached or gets work,
it takes a long time for it to run.
client/
cs_apps.C
Rom 14 Sept 2004
- GUIRPC: Remove memory leaks from the client-side GUI RPC's.
- GUIRPC: Make the API contracts consistant.
clientgui/
MainFrame.h
lib/
gui_rpc_client.C, .h
--- cut due to lenth-----
mfg
Sir Ulli
http://setiweb.ssl.berkeley.edu/sah/download.php
änderungen soweit bekannt
BOINC 4.09 release
I'm hoping this will be good enough for public deployment...
----- Rom
David 6 Sept 2004
- The Windows GUI and CLI versions were using different mechanisms
to ensure that only one instance of BOINC is running.
Factored this into a function check_unique_instance(),
which is used by all versions.
client/
file_names.C,h
main.C
win/
wingui_mainwindow.cpp
lib/
error_numbers.h
David 7 Sept 2004
- (from Ben Herndon) Some compilers optimize away sections of the
Whetstone
benchmark function, since their results aren't used outside
the function.
To deal with this, we added an global-scope array
(extern_array[])
and store the result of each section into it.
Also changed section 3 by initializint j from the
extern_array,
so the compiler can't optimize out all the conditionals.
client/
whetstone.C
David 7 Sept 2004
- change "upgrade" to query user before overwriting files
tools/
upgrade
David 7 Sept 2004
- recognize CPU type on HP (DEC) alpha (from Duane Krahn)
client/
hostinfo_unix.C
Rom 8 Sep 2004
- If we can lock the lockfile, also aquire the global mutex. If we
don't the screensaver
cannot properly detect that BOINC is executing.
client/
file_names.C
Jeff 09 Sep 2004
- Change the validator to handle the case where one or more
of the result files for a WU's result set can't be read.
Currently, this case results in this WU and its results
never getting looked at again - no science, no credit.
We distinguish 2 cases:
1) transient failure, e.g. an NFS mount is not working.
(can't read directory)
In this case check_set() returns a special code to the validator,
which arranges for the WU to get transitioned again in a few hours.
2) permanent failure (file doesn't exist)
The result's validate_state is set to ERROR (new state),
which is generally equivalent to INVALID
- add new WU arg to check_set()
(so it can access min_quorum)
This is coordinated seti_boinc / boinc change.
This note is repeated in the seti_boinc checkin_notes.
seti_boinc changes:
- get_result_file(), if it cannot read the result file,
tries to read the directory where the result file should be.
If it can read the directory, it returns ERR_FOPEN, as before.
If it cannot read the directory, it resturns ERR_OPENDIR.
ERR_OPENDIR in this context signals a possibly transient problem.
Upon a successful result read, it sets sah_result.have_result
(a new data member as of this change) to true.
- check_set(), upon return from get_result_file() :
- sets result.validate_state to VALIDATE_STATE_ERROR (a new
state)
and retval to zero if there is a nonzero retval from
check_set()
and it is not ERR_OPENDIR. It then continues.
- otherwise just continues.
check_set() then determines if any IO errors brought the result
count
below wu.min_quorum. If not, it continues. Otherwise it returns.
In all subsequent logic any results for which
sah_result.have_result == false are skipped.
Note that the result and the sah_result vectors are associated.
validate/
sah_result.h
sah_validate.cpp
sah_boinc_db.cpp
boinc changes:
- new validate state VALIDATE_STATE_ERROR.
- in the enumeration for check_set(),
we now include a clause to check for VALIDATE_STATE_INIT
in order to ignore results set to VALIDATE_STATE_ERROR.
This enumeration query now matches that for check_pair().
Note that the only possible retvals from check_set()
are now zero or ERR_OPENDIR.
Note that logic to retry possibly transient errors still needs
to be done.
Also left to be done is a transitioner change to subtract
the number of results in state VALIDATE_STATE_ERROR
from the count of active results.
The latter change is needed to stimulate the production
of additional results for the affected WU.
sched/
validator.C
db/
boinc_db.h
Rom 9 Sep 2004 ( for Sebastian Masch )
- Don't cause the core client to rewrite the state file
for every block transfered during a network operation.
client/
pers_file_xfer.C
David 9 Sept 2004
- Further refinements of validation to deal with
recoverable/nonrecoverable file read errors.
Revised the signatures of check_set() and check_pair(),
and revised the validator accordingly.
db/
boinc.db.h
sched/
sample_bitwise_validator.C
sample_trivial_validator.C
validate_util.C,h
validator.C
David 9 Sept 2004
- added try_fopen() (from Jeff Cobb):
like fopen(), but on tells you whether it's recoverable or not
lib/
util.C,h
David 9 Sept 2004
- Added support for stats access via WAP (for cell phones)
(from Carl Christensen)
David 9 Sept 2004
- In transitioner, count RESULT_OUTCOME_VALIDATE_ERROR as an error
sched/
transitioner.C
David 9 Sept 2004
- preliminary support for "locality scheduling" (for Einstein@home).
David 10 Sept 2004
- compile fix for Win
lib/
util.C
David 10 Sept 2004
- validator: slight change to check_set() and check_pair():
they no longer should update result records in case of
nonrecoverable error.
Just set the outcome to RESULT_OUTCOME_VALIDATE_ERROR,
and the framework program will do the updates.
sched/
validator.C
David 10 Sept 2004
- more changes for locality scheduling
Rom 10 Sept 2004
- NEw GUI: Put the glue back in so that we can make the various tabs
work.
clientgui/
MainDocument.cpp, .h
stdwx.h
Lana 26 August 2004
- Installation script provides better default daemons by starting
transitioner, feeder and file deleter deamons automatically
tools/
make_project
Lana 11 Sept 2004
- Created DB Purger that archives "DONE" workunits and their
results into XML archive files. Deletes those records from the
database.
Added delete_from_db function to DB_BASE
sched/
db_purge.C
db/
db_base.h
db_base.C
David 12 Sept 2004
- Added a queueing mechanism for shared-memory message passing,
class MSG_QUEUE.
This provides reliable, ordered message delivery.
I think process-control messages were being lost in some
cases.
- Replaced ad-hoc 1-level queueing mechanism for graphics requests
- Changed the logic for selecting app for screensaver a bit;
have new field bool ACTIVE_TASK::is_ss_app that keeps track
of which app is acting as the screensaver.
This fixes bugs in a few particular cases.
David 13 Sept 2004
- Preliminary checkin for "GUI URLs" feature
(allows projects to send URLs to be use as links in GUIs)
This is implemented in a new class GUI_URLS.
Its init() member reads template file.
Its get_gui_urls() member macro-substitutes user/team/host IDs
into the URLs, and produces XML to be included in scheduler replies.
- renamed replace_element() to replace_element_contents()
- moved try_fopen() from lib/util.C to sched/sched_util.C
- added old and new docs to CVS
David 13 Sept 2004
- When email account ID, show (on the next web page) the "from"
address of the email,
since some users may have email filters based on sender.
The "from" address is the constant EMAIL_FROM,
which should be defined in project.inc.
If this is missing, it uses URL_BASE
(which is not an email address, so projects should define
EMAIL_FROM)
David 13 Sept 2004
- small changes to screensaver logic, try to fix Rom's 2-CPU bug
client/
app_graphics.C
ss_logic.C
David 13 Sept 2004
- Change CPU scheduling logic so that the min debt is subtracted
from the debt of all active projects.
i.e. debt is normalized to be zero-based.
This fixes a bug where the debt of projects with work
keeps getting bigger and bigger,
so that when a new project is attached or gets work,
it takes a long time for it to run.
client/
cs_apps.C
Rom 14 Sept 2004
- GUIRPC: Remove memory leaks from the client-side GUI RPC's.
- GUIRPC: Make the API contracts consistant.
clientgui/
MainFrame.h
lib/
gui_rpc_client.C, .h
--- cut due to lenth-----
mfg
Sir Ulli
halbschnabel111
Lieutnant
Hast du schon die 4.06 probiert? Die funzt bei mir auf 3 Win98 SE Rechnern, die 4.09 hab ich noch nicht mit Win98 SE probiert, kommt aber demnächst.
hi!!
geht keine version auf meinem win98se. der hat eine wu gerechnet und seit dem gehts nich mehr. und nochwas ich schicke seit 4 tagen von anderen rechnern wus ab aber es werden keine creditz gutgeschrieben. deswegen habe ich erst mal aufgehört und mich anderen projekten gewidmet.
wader
geht keine version auf meinem win98se. der hat eine wu gerechnet und seit dem gehts nich mehr. und nochwas ich schicke seit 4 tagen von anderen rechnern wus ab aber es werden keine creditz gutgeschrieben. deswegen habe ich erst mal aufgehört und mich anderen projekten gewidmet.
wader
Terminatrix
Grand Admiral Special
Bei mir hat es auch ohne Probleme funktioniert. Einfach in das vorhandene Installationsverzeichniss installieren.
Habe von 4.05 auf 4.09 upgedatet.
Habe von 4.05 auf 4.09 upgedatet.
thelittlegreen
Lieutnant
Habe diese Woche 8 Kisten mit CPDN und 2 mit Seti auf 4.09 aufgerüstet.
Keinerlei Probleme. Einfach darüber installiert.
Ciao
thelittlegreen
Keinerlei Probleme. Einfach darüber installiert.
Ciao
thelittlegreen
- Mitglied seit
- 01.07.2000
- Beiträge
- 24.066
- Renomée
- 10.446
Kann ich nicht bestätigenOriginal geschrieben von AndyRTR
Ich habe seit der neuen Version das Problem, dass ich nicht mehr durch die Meldungen scrollen kann. Die Scrollleiste ist einfach nicht da
Bei Euch auch
Ich kann scrollen wie eh und je 
ComputerTho
Commander
gabs bei Messages in der letzten Version eine Scrolleiste? Ich seh in der neuen keine 
Sir Ulli
Grand Admiral Special
Original geschrieben von AndyRTR
Ich habe seit der neuen Version das Problem, dass ich nicht mehr durch die Meldungen scrollen kann. Die Scrollleiste ist einfach nicht da
Bei Euch auch
hatte ich mit einem Rechner auch, Problem ist aber zur zeit weg.
mfg
Sir Ulli
Sir Ulli
Grand Admiral Special
Terminatrix
Grand Admiral Special
Was heißt jetzt " next time you contact our servers"?
Ist damit gemeint beim nächsten mal hochladen einer WU oder muss man dem Project neu attachen?
Ist damit gemeint beim nächsten mal hochladen einer WU oder muss man dem Project neu attachen?
Sir Ulli
Grand Admiral Special
für die Linux Freunde unter uns
Ned Slider von AthlonMB.com, hat ne Beschreibung wie man den BOINC client unter Linux etwas optimiert
optimise Client for Linux on AMD-Systems
ausserdem erste bilder von der neuen LINUX GUI
http://uml.no-ip-here.net/ss/
mfg
Sir Ulli
Ned Slider von AthlonMB.com, hat ne Beschreibung wie man den BOINC client unter Linux etwas optimiert
optimise Client for Linux on AMD-Systems
ausserdem erste bilder von der neuen LINUX GUI
http://uml.no-ip-here.net/ss/
mfg
Sir Ulli
F
FalconFly
Guest
Selbst compilieren ist ja schoen, aber warum in aller Welt bietet nicht einer einen fertig compiled Client (mit den genannten Optimierungen) zum Download an
Daran hapert es (vor allem fuer nicht ausreichend Linux-bewanderte) immer wieder, Sources sind fuer normale User vollkommen uninteressant, es werden fertige Binaries gebraucht.
Daran hapert es (vor allem fuer nicht ausreichend Linux-bewanderte) immer wieder, Sources sind fuer normale User vollkommen uninteressant, es werden fertige Binaries gebraucht.
Sir Ulli
Grand Admiral Special
Original geschrieben von FalconFly
Selbst compilieren ist ja schoen, aber warum in aller Welt bietet nicht einer einen fertig compiled Client (mit den genannten Optimierungen) zum Download an
Daran hapert es (vor allem fuer nicht ausreichend Linux-bewanderte) immer wieder, Sources sind fuer normale User vollkommen uninteressant, es werden fertige Binaries gebraucht.
und user die in der Lage sind das auch zu machen, bin leider kein Programmierer,
darum call at all Users die das Know How haben und das auch für uns andere machen würden.
mfg
Sir Ulli
F
FalconFly
Guest
Hmm...
Ich muss mal sehen, ob ich den kompletten gcc auf einer meiner Linux-Kisten installiert habe.
(der ist normalerweise nie installiert, weil halt nicht notwending)
Habe im Moment allerdings leider nicht viel Zeit, kann also nichts versprechen.
Ich muss mal sehen, ob ich den kompletten gcc auf einer meiner Linux-Kisten installiert habe.
(der ist normalerweise nie installiert, weil halt nicht notwending)
Habe im Moment allerdings leider nicht viel Zeit, kann also nichts versprechen.
Die Linux Gui sieht nett aus von den Funktionen. Allerdings mal wieder im typischen gtk-look 
Wie weit ist denn der funktionale Stand? Hab die Project URL noch net gefunden. Und gibt es dieses BioncGui noch bzw. mal wieder aktualisiert? Das war während des Betatests schon mal ganz gut. Der Hit wäre natürlich: optimierte RPMs vom Client für verschiede Architekturen und ein KBoinc(war ja mal angekündigt).
Wie weit ist denn der funktionale Stand? Hab die Project URL noch net gefunden. Und gibt es dieses BioncGui noch bzw. mal wieder aktualisiert? Das war während des Betatests schon mal ganz gut. Der Hit wäre natürlich: optimierte RPMs vom Client für verschiede Architekturen und ein KBoinc(war ja mal angekündigt).
was ist denn los Leute.
Genauso machen wie es da steht und es funktioniert einwandfrei!
Die Benchmarks sind jetzt bei integer Speed besser als bei Windoof und bei float kommen sie in die Nähe das sah vorher viel viel schlimmer aus. Mal sehen wie die Zeiten sind.
In diesem Sinne.
Traut euch man kann nichts kaputt machen.
Ist der identische Rechner siehe Sig.
Suse Linux 9.1
Measured floating point speed 2242.14 million ops/sec
Measured integer speed 5786.78 million ops/sec
WinXP SP1
Measured floating point speed 2355.83 million ops/sec
Measured integer speed 5408.24 million ops/sec
Hierbei ist zu überlegen ob es nochwas bringt einen AMD-optimierten Kernel zu basteln da ja leider unter Suse 9.1 nur noch sogenannte Default-Kernel gibt. Bei 8.0, 8.1 8.2 und 9.0 gibt es die ja noch. Ich weiss nicht ob das nochmals einen Schub bringt.
Genauso machen wie es da steht und es funktioniert einwandfrei!
Die Benchmarks sind jetzt bei integer Speed besser als bei Windoof und bei float kommen sie in die Nähe das sah vorher viel viel schlimmer aus. Mal sehen wie die Zeiten sind.
In diesem Sinne.
Traut euch man kann nichts kaputt machen.
Ist der identische Rechner siehe Sig.
Suse Linux 9.1
Measured floating point speed 2242.14 million ops/sec
Measured integer speed 5786.78 million ops/sec
WinXP SP1
Measured floating point speed 2355.83 million ops/sec
Measured integer speed 5408.24 million ops/sec
Hierbei ist zu überlegen ob es nochwas bringt einen AMD-optimierten Kernel zu basteln da ja leider unter Suse 9.1 nur noch sogenannte Default-Kernel gibt. Bei 8.0, 8.1 8.2 und 9.0 gibt es die ja noch. Ich weiss nicht ob das nochmals einen Schub bringt.
Zuletzt bearbeitet:
F
FalconFly
Guest
Nun, das wissen wir bereits.
Ein Download link auf eine entsprechende Binary ist aber das, worauf all die warten, die es entweder nicht koennen oder einfach keine Zeit haben...
Ähnliche Themen
