- whether host is "reliable" for an app version
- whether host is eligible for single replication for an app version
- whether to use host scaling
In each case, the answer is yes if the number of
consecutive valid results is above a threshold.
This replaces existing "error rate" and "scale probation" mechanisms.
TODO: the # of consecutive valid results should also determine
a limit on jobs in progress for an app version.
Namely, if N is the threshold for host scaling, the limit should be
ndevices*(max(1, consecutive_valid - N))
The client currently doesn't supply enough
app version info to do this.
It could be approximated; that would give some protection
against cherry-picking.
- credit: more conservative formulas for combining claimed credit
among replicas.
If there are normal replicas, we use a "low average"
that weights each sample by the sum of the other samples.
Otherwise we use the min (not the average) of the approximate samples.
NOTE: a DB update is required
svn path=/trunk/boinc/; revision=21230
- daily quota mechanism
- reliable mechanism (accelerated retries)
- "trusted" mechanism (adaptive replication)
- scheduler: enforce host scale probation only for apps with
host_scale_check set.
- validator: do scale probation on invalid results
(need this in addition to error and timeout cases)
- feeder: update app version scales every 10 min, not 10 sec
- back-end apps: support --foo as well as -foo for options
Notes:
- If you have, say, cuda, cuda23 and cuda_fermi plan classes,
a host will have separate quotas for each one.
That means it could error out on 100 jobs for cuda_fermi,
and when its quota goes to zero,
error out on 100 jobs for cuda23, etc.
This is intentional; there may be cases where one version
works but not the others.
- host.error_rate and host.max_results_day are deprecated
TODO:
- the values in the app table for limits on jobs in progress etc.
should override rather than config.xml.
Implementation notes:
scheduler:
process_request():
read all host_app_versions for host at start;
Compute "reliable" and "trusted" for each one.
write modified records at end
get_app_version():
add "reliable_only" arg; if set, use only reliable versions
skip over-quota versions
Multi-pass scheduling: if have at least one reliable version,
do a pass for jobs that need reliable,
and use only reliable versions.
Then clear best_app_versions cache.
Score-based scheduling: for need-reliable jobs,
it will pick the fastest version,
then give a score bonus if that version happens to be reliable.
When get back a successful result from client:
increase daily quota
When get back an error result from client:
impose scale probation
decrease daily quota if not aborted
Validator:
when handling a WU, create a vector of HOST_APP_VERSION
parallel to vector of RESULT.
Pass it to assign_credit_set().
Make copies of originals so we can update only modified ones
update HOST_APP_VERSION error rates
Transitioner:
decrease quota on timeout
svn path=/trunk/boinc/; revision=21181
1) peak FLOPS (based on benchmarks or GPU attributes).
This does not change over time.
It's not adjusted on the basis of statistics.
It's not affected by wu.rsc_fpops_est.
It can be compared across projects.
versus
2) projected FLOPS: the scheduler's best guess as to what will satisfy
X * elapsed_time = wu.rsc_fpops_est;
this is used to make server-side runtime estimates,
and it's sent to the client and used for its runtime estimates.
It may be based on the (host, app version) elapsed time average.
My checkin [21153] mistakently confounded these two.
Notes:
1) app_plan() now must return both peak and projected FLOPS.
2) result.flops_estimate stores peak FLOPS
3) the <flops> field in app_info.xml files should be
projected FLOPS. But its accuracy is not important;
it's not used once the server has statistics
for the (host, app version)
svn path=/trunk/boinc/; revision=21164
TODO: remove related code
- validator: update wu.canonical_credit correctly.
However, this field should be deprecated.
- validator: check for error return from assign_credit_set().
svn path=/trunk/boinc/; revision=21096
(todo: do this for all daemons)
- validator: change cmdline args from -foo to --foo
(todo: do this for all daemons)
- validator: pass max_granted_credit to assign_credit_set()
svn path=/trunk/boinc/; revision=21093
are written to the DB.
The incremental approach was bogus.
New approach:
host_app_version: write directly; R/W interval is tiny
app_version: maintain an explicit list of update samples
for both PFC and credit.
When the validator flushes its app_version cache,
do careful updates.
Note: when using double fields in careful updates,
you can't test for equality. Use abs(new-old) < 1e-N
svn path=/trunk/boinc/; revision=21057
see http://boinc.berkeley.edu/trac/wiki/CreditNew
Projects will need to update DB and recompile all back-end programs.
Summary:
- new way of computing credit
- "reliable host" mechanism is per app version
- "host punishment" mechanism is per app version
- adjustment of wu.rsc_fpops_est provides the
equivalent of per app version DCF
- max jobs in progress is now per app
- max jobs per RPC is now per app
TODO:
- reliable mechanism:
- populate and use host_app_version.error_rate
- populate host_app_version.turnaround
- host punishment:
- populate host_app_version.max_jobs_per_day
- populate host_app_version.n_jobs_today
- use app.max_jobs_per_day_init
- job limits:
- use app.max_jobs_in_progress, max_gpu_jobs_in_progress
- use app.max_jobs_per_rpc
- adjust wu.rsc_fpops_est
- remove old credit stuff
fpops_cumulative, credit_multiplier
credit computation in scheduler
- AVERAGE class: use the Knuth algorithm (Wikipedia)
svn path=/trunk/boinc/; revision=21021
David Anderson