treat it as a "backup project":
fetch work from it only if there is an idle instance
and no other projects have work.
svn path=/trunk/boinc/; revision=20286
if job A is unstarted and EDF,
and there's a job B that is later in the list,
is started, has the same app version,
and has the same arrival time,
move A after B.
- client: remove the "temp_dcf" mechanism,
which had the same goal but didn't work.
- client: in computing overall debt for a project,
subtract a term that reflects pending work.
This should reduce repeated fetches from the same project.
- client simulator: tweaks
svn path=/trunk/boinc/; revision=20223
- a project overestimates job FLOP counts
- the client starts jobs in EDF mode
- as job progresses and fraction done increases,
its completion time estimate decreases until
it's no longer a deadline miss.
- job gets preempted by other job from that project;
you end up with lots of partly completed jobs.
Solution (I hope): if an app version has running jobs,
compute a "temp DCF" for the app version,
which is the min of dynamic/static estimates for its jobs.
Apply this scaling factor to completion time estimates
for unstarted jobs in RR simulation
- client: the estimation of remaining time of running jobs was wrong
(how did this bug survive so long?)
svn path=/trunk/boinc/; revision=20077
RAM to run job, but when we actually run the job
not enough GPU RAM is free, so the application fails.
This can cause a large number of jobs to fail.
Solution:
- app_plan() can specify the GPU RAM requirements of an app version.
This is passed to the client in a new field
<gpu_ram> of the <app_version> element.
- prior to starting or restarting a GPU app, the client
checks the amount of free RAM on the particular GPU.
If it's not enough for the app version,
the client doesn't start it,
and arranges for the scheduler to ignore it for 5 minutes
(by which point there might be more free GPU RAM)
Notes:
1) this change will have effect only when
both client and scheduler are updated.
2) the check is done in enforce_schedule(),
rather than schedule_cpus(),
because only at that point
have we assigned a specific GPU to the job.
3) there's another case to deal with:
a GPU app's malloc of GPU RAM fails in the middle of the job.
Currently the job fails.
I plan to add an API call boinc_temporary_exit(x) so
that the job can exit and potentially restart in x seconds.
(In principle this mechanism is sufficient for all cases,
but it could lead to a lot of starting/exiting,
so the current change is worthwhile).
svn path=/trunk/boinc/; revision=19864
It computed an "overall STD" as the sum of CPU and coprocs,
weighted by the coproc's speed, as we do for LTD.
This was the wrong idea; in the presence of GPUs,
STDs quickly get pushed to +- 1 day and are truncated there.
New scheme: STD is maintained per (resource type, project).
This fixes the above problem,
and it opens to door to round-robin scheduling of GPUs.
- client: the calculation of "anticipated debt" was scaling
by relative resource share.
This wasn't correct, seems to me.
- client: rename "debt" to "long_term_debt" in a few places
(but not in the client state file, for compatibility)
svn path=/trunk/boinc/; revision=19777
Old: it's based entirely on CPU time.
So a GPU project, whose app uses only a fraction
of a CPU, accrues positive debt.
This is OK if the project has only GPU apps,
since STD is not (currently) used for GPU scheduling.
But some projects have both CPU and GPU apps.
New: STD is based on total processing.
It has terms for each resource type.
The notion of "runnable resource share" is specific to a type.
Note: the notion of "resource share fraction" appears in
a couple of other places:
- it's passed to apps in app_init_data.xml
- it's passed in scheduler requests.
It should be broken down by resource type in these cases too.
Note to self: do this later.
svn path=/trunk/boinc/; revision=19762
set a "coproc_missing" flag rather than aborting the job.
If use removes a GPU board while there's a large queue of GPU jobs,
they'll stay queued (until their deadline passes).
Note: this doesn't fix the situation where user connects via
Remote Desktop while GPU jobs are running or queued.
We should check for Remote Desktop every minute or so, and stop GPU jobs.
svn path=/trunk/boinc/; revision=19287
to accept CPU, NVIDIA and ATI jobs.
These prefs are shown only where relevant:
e.g., only for processor types for which the project has app versions,
and if it has versions for only one type, no pref is shown.
These prefs affect both client and scheduler.
The client won't ask for work for a device blocked by prefs,
and the scheduler won't send it.
This replaces earlier optional project-specific prefs for
"no CPU jobs" and "no GPU jobs".
(However, these prefs continue to be honored on the server side).
- client: if NVIDIA driver is unknown, say that rather than 0
svn path=/trunk/boinc/; revision=19194
ones already running.
The problem: we considered a job as started if it has an ACTIVE_TASK.
However, we were creating ACTIVE_TASKS for jobs before deciding
to run them, because we needed a place to store the coproc reservations.
This caused the above bug, and also had the undesirable effect
of creating slot directories before they're needed.
Solution: store coprocessor reservations in RESULT
rather than ACTIVE_TASK.
svn path=/trunk/boinc/; revision=19129
e.g. the Milkyway@home ATI app, of which we can typically run
2 or 3 instances at once on a GPU.
Changes include:
- In APP_VERSION, don't use a COPROCS to represent the GPU
requirements; just use doubles ncudas and natis.
- sufficient_coprocs() etc. are no longer members of COPROCS
- in HOST_USAGE, ncudas and natis are doubles
- in scheduler request, req_instances is now a double
This checkin doesn't include the job scheduling logic,
i.e. assigning jobs to GPUs. That will follow.
svn path=/trunk/boinc/; revision=18868
2 * max(ncpus, ngpus);
show this in the state displayed by <work_fetch_debug>
- manager: show project-wide backoff in transfers tab
svn path=/trunk/boinc/; revision=18662
uploads and downloads.
I originally added this on 30 Sept 2005
and disabled it 2 weeks later because there were reports of problems.
However, we need this functionality
(e.g. on GPU hosts with hundreds of files to upload,
we need to back off after a few failures, not try all of them).
I added messages (<file_xfer_debug>) so you can see what's going on.
Fixes#932.
svn path=/trunk/boinc/; revision=18593
New approach: take the "ordered_schedule_results" list,
add running jobs that haven't finished their time slice,
and order the result appropriately.
Then run jobs in order until CPUs are filled.
Simpler and clearer than the old way.
svn path=/trunk/boinc/; revision=17992
- first schedule jobs projected to miss deadline in EDF order
- then schedule remaining jobs in FIFO order
This is intended to reduce the number of preemptions of coproc jobs,
and hence (since they are always preempted by quit)
to reduce the wasted time due to checkpoint gaps.
- client: the CPU scheduling policy made use of the number
of deadline misses in various places.
This should include only the deadline misses of CPU jobs.
So move "deadlines_missed" from RR_SIM_STATUS and PROJECT
to RSC_PROJECT_WORK_FETCH so that we have separate counts
for CPU and coproc jobs, and use the count for CPU jobs.
- GUI RPC: removed the rr_sim_deadlines_missed field
from project descriptor.
This is no longer meaningful, and it didn't seem to be used anywhere.
svn path=/trunk/boinc/; revision=17785
keep track of the largest WSS of tasks using it.
In checking whether tasks fit in RAM,
use this as an estimate for tasks that haven't started yet.
This avoids a situation where the client starts a lot of
tasks in sequence, only to find that each one doesn't fit in RAM.
svn path=/trunk/boinc/; revision=17765
CPU time is visible in task Properties.
- Manager: in task Properties, show final CPU and elapsed times
if job is finished
- client: honor backoff for account-manager-requested scheduler RPCs
- client: keep track final elapsed time for results
- GUI RPC: report final elapsed time
svn path=/trunk/boinc/; revision=17588
and passing them the corresponding --device N cmdline args.
This fixes a bug introduced in 17402 (Feb 26)
that broke the --device feature,
presumably causing problems on systems with multiple GPUs.
svn path=/trunk/boinc/; revision=17549
app versions in scheduler reply
- client: when reporting anonymous platform apps in sched request,
don't include <file_info>s (not relevant to server)
svn path=/trunk/boinc/; revision=17507
which of those files to include
- Modified MAC address check to work on some non-Linux unixes.
(mac_address.cpp)
- Added suggested change to "already attached to project" checking.
(ProjectInfoPage.cpp)
- changed includes of standard c header files to their c++ equivalents
(i.e. replaced <stdio.h> with <cstdio>) for namespace protection.
- replaced "using namespace std;" with more explicit "using std::function" in
several files.
- Fixed bug in checking whether the os is OS/2 and added conditional OS_OS2
to the build environment. (boinc_platform.m4,configure.ac)
- Changed build environment to not use -nostandardlibs unless we are using
G++ and static linkage is specified. (configure.ac)
- Added makefiles and package building files for solaris CSW package manager.
- Fixed bug with attempting to find login name using logname. (configure.ac)
- Added ifdef HAVE_* protection around some include files commonly found in
sys.
- Added support for unified binary for x86_64/i686-pc-solaris.
(cs_platforms.cpp)
- generate_host_cpid() now uses MAC address on non-linux unix.
(hostinfo_network.cpp)
- Macro BOINC_SET_COMPILE_FLAGS now doesn't check gcc only flags on non-gcc
compilers. (boinc_set_compile_flags.m4)
- Library compiles no longer depend upon the library extension or require
the library to be prefixed with lib.
- More fixes for fcgi builds.
- Added declaration of "struct ether_addr" and ether_ntoa(). Have not yet
implemented ether_ntoa() for machines that don't have it, or where it is
buggy. (unix_util.h)
- Added FCGI::perror() which calls FCGI_perror(). (boinc_fcgi.{h,cpp})
- Fixed library Makefiles so that all required headers get installed.
svn path=/trunk/boinc/; revision=17388
using a coprocessor we don't know about, ignore it
(and all results using that app_version will be flushed).
This deals with the situation where we have some GPU jobs,
but the GPU card is removed (previously this resulted in a crash).
This requires some code shuffling so that we check for coprocessors
before reading state file.
svn path=/trunk/boinc/; revision=17161
worked in the presence of coprocessors.
The simulator maintained per-project queues of pending jobs.
When a job finished (in the simulation) it would get
one or more jobs from that project's pending queue.
The problem: this could cause "holes" in the scheduling of GPUs,
and produce an erroneous nonzero shortfall for GPUs,
leading to infinite work fetch.
The solution: maintain a separate (per-resource, not per--project)
queue of pending coprocessor jobs.
When a coprocessor job finishes,
start pending jobs from the queue for that resource.
Another change: the simulator did strict reservation of coprocessors.
If there are 2 instances of CUDA,
and a 1-instance job is running in the simulation,
it wouldn't start an additional 2-instance job.
This also can cause erroneous nonzero shortfalls.
So instead, schedule coprocessors like CPUs, i.e. saturate them.
This can cause distorted completion time estimates,
but it's better than infinite work fetch.
svn path=/trunk/boinc/; revision=17093
- client: respect work-fetch backoff for non-CPU-intensive projects
- client: for non-CPU-intensive project, fetch new job
if no currently running jobs
- client: skip non-CPU-intensive projects in debt calculations
- manager: show resource backoff times correctly
svn path=/trunk/boinc/; revision=16998
(otherwise it doesn't work for coproc or multi-proc apps)
- client: in estimate of job completion time,
weight the estimate based on fraction done more heavily
(quadratic rather than linear)
svn path=/trunk/boinc/; revision=16603
as the basis for estimating job completion times.
This should improve estimates for GPU apps,
and prevent the DCF from getting messed up.
svn path=/trunk/boinc/; revision=16598
- scheduler: fix bug in adaptive replication:
if send an unreplicated job to untrusted host,
set both wu.target_nresults and wu.min_quorum to app.target_nresults.
svn path=/trunk/boinc/; revision=15762
This supports apps that can do variable amounts of computing;
they can boinc_finish() if their deadline is near.
Rom: please back-port.
svn path=/trunk/boinc/; revision=15395
and clear all timeout variables.
This should fix the situation where, say:
1) the user sets the system clock forward by a year;
2) all projects get their min_rpc_time set;
3) the user sets the system clock back to the correct time.
Previously, BOINC would not do anything for a year.
Note: a restart of BOINC is required to fix things.
It would be harder to do this on the fly.
svn path=/trunk/boinc/; revision=15314
David Anderson