BOINC:FAST 2017 conference
The report about this project was made at the BOINC:FAST 2017 conference. Here is the list of talks. You can find the amazing report about LHC@Home there as well as the reports about some new projects: Acoustics@Home (report), AndersonAttack@home (report), RakeSearch (report) and USPEX@Home (report). 12 Sep 2017, 16:18:10 UTC · Comment
We are trying to configure SSL encryption. This can cause a short-term server malfunction. We apologize for any inconvenience.
Nikolay has successfully configured SSL encryption for this project. Now, the most valuable communications are secured.
The emails are also SSL-encrypted. Those who previously had problems receiving emails from the project (password recovery, etc.) will now be able to receive them. Just don't forget to check the spam folder. 15 Aug 2017, 14:31:13 UTC · Comment
I’ve added badges for total credit (for users and teams) and total number of different COD entries processed (only for users).
The steps for total credit are:
The x-ray powder diffraction patterns of respective materials are shown on the badges (since the diffraction pattern of ordinary gold and silver look identical, I used the diffraction pattern of hexagonal gold nanostructured films for the gold).
The steps for total number of different COD entries processed are:
1,000; 2,000; …; 9,000 (bronze)
4 Aug 2017, 15:02:40 UTC
10,000; 20,000; …; 90,000 (silver)
100,000; 200,000; 300,000 (gold)
Active phase of the project is completed by 20% in the first 24 days
Only about 0.2% of the obtained results should be recalculated due to bugs in the early versions of the apps (the results obtained with the version 1.04 mostly). All other results are included in the final database. 12 Jul 2017, 20:28:27 UTC · Comment
Crystallography Open Database server is temporarily down
We get the input data for the WUs from the http://crystallography.net. So, while this server is down we can't generate new WUs.
Added a mirror.
It seems that http://crystallography.net/ is up again. But if the problem recurs in the future, the requests will be redirected to the mirror server, http://qiserver.ugr.es/cod/. 12 Jul 2017, 16:20:55 UTC · Comment
Source code of the client app is released
You can download it here. 28 Jun 2017, 21:18:48 UTC · Comment
CPU usage in GPU apps
The new CUDA app 1.03 does not use the CPU core anymore except for the computation of the atomic ensemble in the beginning.
The new OpenCL app 1.03 has a command line option --nowait which when specified reduces the CPU load to zero but at the same time makes the system almost unusable due to lagging. If you do not use the system while crunching, you may add this to your app_config.xml:
Replace 'opencl_ati_102_windows' with the appropriate plan class: opencl_ati_102_mac, opencl_ati_102_linux, opencl_intel_gpu_102_windows, opencl_intel_gpu_102_linux, opencl_intel_gpu_102_mac or opencl_nvidia_102_linux.
All OpenCL 1.03 versions are marked as beta due to potential memory leak problem.
The 1.03 app uses exactly the same amount of memory as 1.02 does, so there is no memory leak. The memory consumption depends on the WU. Those tasks which have 'solid_material' in their names may consume up to 1.5 GB of memory depending on the initial data. Please take this into consideration when launching multiple WUs in parallel.
I'll try to reduce memory consumption in the next version. 25 Jun 2017, 23:13:15 UTC · Comment
Maintenance is over!
The project is restarted with new native BOINC apps and additional computational scenario.
Now the large WUs can be assigned only to the GPU and the small ones only to the CPU.
GPU apps do not require dedicated CPU core by default anymore. If you notice any performance issues caused by this, consider using custom app configuration. Place this file in the project's folder.
The apps of this project do not make checkpoints. If stopped, the task will be restarted from the beginning. The task execution time should not exceed 10 minutes for moderate PCs. However, some tasks may take longer to complete on low-end PCs. This is particularly true for low-end GPUs. If the execution time of tasks is unacceptably high for your GPU, consider disabling it in the "Preferences for this project" in your profile page.
The project is in beta-testing stage. Bugs are more than likely. However, all the results obtained during this stage will be included in the final database If no critical errors are found.
Thank you for participating! 18 Jun 2017, 18:09:05 UTC · Comment
Project restart in June
The maintenance is almost over. The new native BOINC apps are being tested in the local network now. We expect the project to be restarted at the end of next week. This will be preceded by server and storage upgrade. The work on the database is still in progress and will continue in parallel with the computations.
Also, the paper “XANSONS for COD: a new small BOINC project in crystallography” was accepted for presentation at the BOINC:FAST’2017 conference which is to be held in Petrozavodsk, Russia in August. 5 Jun 2017, 21:46:33 UTC · Comment
XaNSoNS is published in SoftwareX
The article about XaNSoNS, the software this project is based on, has been published in the peer-reviewed scientific journal SoftwareX. The article is open-access and can be found here.
The paper was submitted to the journal about a month before this project was launched, so I was able to add only a brief mention of the BOINC project during the revision process. 7 Feb 2017, 19:27:29 UTC · Comment
The number of project team members has doubled!
Nikolay Khrapov agreed to join the project in 2017 as a developer of the diffraction pattern database. He will also help with the system administration. Nikolay is known to the BOINC community primarily as the system administrator for the SAT@Home project. Now things will get easier!
Happy New Year! 31 Dec 2016, 14:49:01 UTC · Comment
Alpha-testing is over! The project is stopped for long-term maintenance. I will post the progress of work from time to time. However, please do not expect much until the spring of 2017.
The work generator is stopped. BOINC server will wait for all WUs that are in-progress. The site and the forum will continue to work during the maintenance. However, the registration of new users will be suspended. The TODO list can be found here.
The alpha-testing was successful! Here are some results and conclusions.
CPU, single-threaded: < 10 nm
CPU, multi-threaded: 10 – 20 nm
GPU: > 20 nm
I am very grateful to all volunteers participating in the alpha-testing! I apologize for poor server performance, awful job scheduling, app crashes and bugs leading to a waste of computational resources. Thank you for your patience and understanding!
I will post the progress of work from time to time. However, please do not expect much until the spring of 2017.
Problems with the storage IOPS
The Amazon EBS storage expended all its IOPS credits and the performance dropped down from ~300 IOPS to 90 IOPS. The number of WUs is limited again. Sorry for that.
The 1TB EBS volume with constant performance of 3000 IOPS will be attached to the server starting from the beta stage. However there is no point to attach it now because alpha stage will be over in 2 days. 27 Nov 2016, 14:51:39 UTC · Comment
Server upgrade and the last stage of alpha-testing
24 Nov 2016, 20:38:45 UTC
Server will be upgraded to Amazon c4.large instance this Saturday. The upgrade should be easy but who knows what can go wrong.
When the server will be upgraded, the number of new WUs generated per hour will be increased 5-10 times. The final stage of alpha-testing will last for 2-3 days. Then the project will be suspended for long-term maintenance which will take a few months.
During the maintenance:
Update: Server upgrade was successful. The number of new WUs generated per hour is increased tenfold. However, the parameters of size_regulator are to be optimized...
24 Nov 2016, 20:38:45 UTC
Issues with integrated AMD GPUs
For the unknown reasons the OpenCL app does not work on integrated AMD GPUs.
I created a thread on this issue. The owners of the integrated AMD GPUs are invited to take part in a small test.
Update: The new OpenCL version 1.02 is provided for AMD (windows only). Please, test whether the issue is solved.
Update 2: This problem is solved in OpenCL version 1.02 23 Nov 2016, 20:14:53 UTC · Comment
Master URL has been changed from IP to domain name
This should solve this issue. I hope this will not break anything.
I apology for any inconvenience caused. 22 Nov 2016, 8:43:03 UTC · Comment
Multi-size application feature is finally implemented
Server will try to send large jobs to fast devices and small jobs to slow devices. However, this feature requires one day or so to start working properly. 21 Nov 2016, 20:33:08 UTC · Comment
Troubles with Nvidia GTX 570 and 580
I've noticed that CUDA app generates wrong results on Nvidia GTX 570 and 580 (multiple hosts). The reason is unknown to me. Maybe, there are some issues with the SASS code generated for CUDA devices with compute capability (CC) 2.0. I've provided the new executable (windows only), which does not include SASS code for CC 2.0, so all devices with CC 2.0 and 2.1 are forced to build the SASS code from the PTX now.
Update: Another GPU with CC 2.0 is affected - GTX 480M. The patch did not help.
Update 2: It seems, that some Nvidia GPUs with CC 2.1 are also affected but not all of them.
Update 3: I've created a separate thread on this issue.
Update 4: These problems seems to be solved in the version 1.06. 20 Nov 2016, 12:54:29 UTC · Comment
Bug fixing and changes in GPU plan classes Also, to address this issue, all GPU versions now require a dedicated CPU core. Please, let me know, if this will cause problems.
Two bugs were fixed thanks to Steve Hawker* report:
a bug preventing the mt_mac app version from running on Mac hosts with no GCC runtime,
a bug preventing all OpenCL versions from querying some old AMD GPUs if they can’t report their wavefront width value (CL_DEVICE_WAVEFRONT_WIDTH_AMD).
Also, to address this issue, all GPU versions now require a dedicated CPU core. Please, let me know, if this will cause problems.19 Nov 2016, 18:54:11 UTC · Comment
Statistics per app version for last 24 hours
Here are Pass/Fail rates per app version - link.
All fails for mt_mac (OpenMP-multithreaded for Mac) are for the same host. All fails for opencl_ati_102_mac are also for the same host. Maybe, there are problems with these app versions but better statistics is needed.
The most problematic app version so far is opencl_intel_gpu_102_windows. The detailed investigation is needed.
Other app versions perform good enough for an alpha stage.
About 1% of COD entries are already processed. The project will be suspended for a long-term (few months) maintenance after 5-7% of COD entries will be processed. The amount of collected data should be more than enough for further project development. 16 Nov 2016, 17:07:39 UTC · Comment
Generation of new WUs is restarted
The number of new WUs per hour may vary until the optimal value is found.
The value of target_nresults is set to the value of min_quorum again to reduce the number of results aborted by project.
Some users may get EXIT_TIME_LIMIT_EXCEEDED errors for the results obtained on slow machines. This happens because the job parameters are not optimized yet. I am sorry for that. 16 Nov 2016, 11:38:43 UTC · Comment
Generation of new WUs is temporary suspended due to server overload
It will be restored tomorrow with a limit of about 30,000 new WUs per day. 15 Nov 2016, 22:57:32 UTC · Comment
Restart after maintenance
15 Nov 2016, 15:26:10 UTC
The project is restarted after the maintenance.
Attention! Do not run this project on integrated AMD GPUs. This may cause a display driver crash!
15 Nov 2016, 15:26:10 UTC
The project is stopped for maintenance for a couple of days.
Thank you very much for participating and testing! The project is stopped for maintenance until the proper archiving of the calculated diffraction patterns is implemented. This may take a couple of days. 11 Nov 2016, 20:15:55 UTC · Comment
"XANSONS for COD" BOINC project is launched in test mode. The development of this project will continue until the end of 2017. Then the project will be re-launched. 26 Oct 2016, 17:38:13 UTC · Comment