Message boards :
Number crunching :
SETI & AP apps/tasks checkpoints & HDD
Message board moderation
Author | Message |
---|---|
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
If a PC let run 48 (24 CPU & 24 GPU) SETI/AP project tasks simultaneously ... the HDD could be a bottleneck? Is there a difference between stock and opti apps, when done (AFAIK, (opti) SETI CPU/GPU & AP CPU set prefs & AP GPU +0.9%) and the size of the checkpoints? If I increase the checkpoint value from default 60 to 240 seconds, what does this mean for the PC? The usage of the system-RAM will increase? In this +180 seconds, where is the calculation progress hold temporarily? Or increase this the calculation times of the tasks? And so on ... How big is the data size (each SETI or AP task checkpoint) which need to transfer (over chip, cable ...) and save to HDD? A SAS 3.5" HDD (15,000RPM) would be the fastest/best solution (AFAIK, double costs than a 'normal' HDD)? A SSD wouldn't go because of the read/write cycles (lifetime), or? What's about a RAM-Drive, the checkpoints could be temporarily saved there, and maybe every 30 mins or so copied to HDD? How big should be the system-RAM for this at least? (Currently I calculate with 4x4GB for each CPU, so 32GB in whole. Enough?) OS Win8.1 Pro x64 (BTW, at least Pro Edition support multi-CPU, right?) Thanks. |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
24 GPU tasks? I'm thinking even the largest GPU's should only be running 4 tasks at a time, even at that you should probably free up 1 core per GPU, or at least 2 total. There was a lot of discussion about SSD here. http://setiathome.berkeley.edu/forum_thread.php?id=76413 Consensus was basically, you couldn't burn out a SSD on SETI for about 4 times (or even much more) the life of your computer. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
24 GPU tasks? I'm thinking even the largest GPU's should only be running 4 tasks at a time, even at that you should probably free up 1 core per GPU, or at least 2 total. As I mentioned in that thread. I am running about 1% wear a year on my SSD, but it isn't a machine that only runs BOINC. Both my Gaming PC & HTPC use the SSD for OS, SWAP, & BOINC. At work I have a 24 core server I run BOINC on a 2GB RAM Disk. It writes the RAM Disk to its image file once an hour. That machine has a great deal of disk activity so I don't want to add BOINC to its already heavy load. Even if it is only once a min. However, I have not checked the delta value when it writes the RAM Disk to its image file. EDIT: For windows there is a command line to check disk usage. I think it is total since PC was started/rebooted C:\>fsutil fsinfo statistics c: SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Drive acted only as BOINCdata storage, PC active few days: C:\_bench_AP.3>fsutil fsinfo statistics e: File System Type : NTFS UserFileReads : 1146 UserFileReadBytes : 52487680 UserDiskReads : 1183 UserFileWrites : 12919 UserFileWriteBytes : 225779712 UserDiskWrites : 14349 MetaDataReads : 1210 MetaDataReadBytes : 4956160 MetaDataDiskReads : 1303 MetaDataWrites : 17293 MetaDataWriteBytes : 80232448 MetaDataDiskWrites : 22871 MftReads : 174 MftReadBytes : 712704 MftWrites : 12035 MftWriteBytes : 55005184 Mft2Writes : 0 Mft2WriteBytes : 0 RootIndexReads : 0 RootIndexReadBytes : 0 RootIndexWrites : 0 RootIndexWriteBytes : 0 BitmapReads : 992 BitmapReadBytes : 4063232 BitmapWrites : 4711 BitmapWriteBytes : 22986752 MftBitmapReads : 2 MftBitmapReadBytes : 8192 MftBitmapWrites : 547 MftBitmapWriteBytes : 2240512 UserIndexReads : 93 UserIndexReadBytes : 380928 UserIndexWrites : 5535 UserIndexWriteBytes : 23007232 LogFileReads : 7 LogFileReadBytes : 28672 LogFileWrites : 40743 LogFileWriteBytes : 192643072 LogFileFull : 0 Summarizing: BOINC is not reader, BOINC is a writer :) |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
I general, I expect much bigger HDD activity from BOINC per se than from checkpointing apps. Maybe instead of building 24-GPU tasks in fly host one could consider to build 24-tasks in hour host. That is, use less number of parallel tasks but compute single task faster. HDD usage-wise it would mean one can effectively disable checkpointing for GPU tasks at all, if they run fast enough. For example, if one uses RAM-drive with it's "checkpointing" interval of 1 hour (as in post above) there is zero-sense to set app's checkpoint in less than 1 hour (provided BOINC runs 24/7 of course, and no suspend while gaming rule). If GPU app runs less than hour in such case it will never checkpoint, only write final results. And no, bigger checkpoint interval will not increase amount of used memory (for SETI apps). All data kept in memory anyway. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Summarizing: BOINC is not reader, BOINC is a writer :) Because BOINC is mainly writing recovery snapshot files - both application checkpoints, and client_state.xml - which are only read after a computer, BOINC, or application restart. About once a month, in my case. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Sure. We just have some proverb here that makes joke from cited phrase :) (In general, about person who not reads/listen but only speaks/writes). |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.