| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Uncovering Genome Mysteries
Thread: Too much disk writing |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 45
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
When running UGM, my computers write to disk every 6-8 seconds. Does this happen to other people?
|
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges: 
|
I don't notice it and my HD is pretty slow.
----------------------------------------Edit: Nevermind that :(  - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz [Edit 1 times, last edit by Falconet at Oct 25, 2014 5:33:55 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes. Others have reported this too.
I have speculated, based on observations, that the results file is continuously written to. Presumably this occurs once per sequence comparison, of which there are very many. The BOINC "Write to disk at most every" setting is used only for the checkpoint file. It seems that whoever programmed this chose to ignore that setting when it came to the results file. A mistake, IMHO. Maybe if enough people complain we might get an update for this when the techs fix the checkpoint bug. (But I wouldn't hold your breath.) |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I used iotop and can confirm that UGM is the project that is writing to disk.
|
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
Is it a problem? Each UGM slot appears to be writing on average about 2kB/sec on one of my Haswell machines, and about 1kB/sec on another. That is only half of the NTFS log file on Win7, not to mention all the other programs, and should be no problem.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
What reduces disk life is not data transfer rate but number of writes. If not using UGM, my machines don't write to disk every 6 seconds but every 20 minutes.
----------------------------------------From Wikipedia: Most manufacturers design the sliders to survive 50,000 contact cycles before the chance of damage on startup rises above 50%. However, the decay rate is not linear: when a disk is younger and has had fewer start-stop cycles, it has a better chance of surviving the next startup than an older, higher-mileage disk (as the head literally drags along the disk's surface until the air bearing is established). For example, the Seagate Barracuda 7200.10 series of desktop hard disks are rated to 50,000 start-stop cycles, in other words no failures attributed to the head-platter interface were seen before at least 50,000 start-stop cycles during testing. http://en.wikipedia.org/wiki/Hard_disk_drive_failure#Landing_zones I will stop crunching for UGM until this is fixed. [Edit 2 times, last edit by Former Member at Oct 26, 2014 9:18:31 AM] |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
Number of writes is not the same as start/stop cycles. The latter refers to when the disk goes into power down mode (stops spinning), and then starts up again. But the disk is constantly being written to when the machine is active (not in sleep mode), and hence is always spinning. Just look at Task Manager/Performance/Resource Monitor/Disk Activity to see all the writes.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Number of writes is not the same as start/stop cycles. Yes, but an isolated write on a stopped disk will make it start, and then stop again until the next isolated write. But the disk is constantly being written to when the machine is active (not in sleep mode) Maybe your machine does that. Mine doesn't. Just look at Task Manager/Performance/Resource Monitor/Disk Activity to see all the writes. I don't have that. As I wrote above, I used iotop and checked that UGM is the only program that writes to disk in a long time. I set my BOINC preferences to write to disk every 20 minutes. I don't use these computers for nothing more than BOINC crunching. Also, my disks are 5-7 years old. I don't want to force them needlessly. [Edit 1 times, last edit by Former Member at Oct 26, 2014 3:17:34 PM] |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
I don't have that. As I wrote above, I used iotop and checked that UGM is the only program that writes to disk in a long time. I set my BOINC preferences to write to disk every 20 minutes. I don't use these computers for nothing more than BOINC crunching. Also, my disks are 5-7 years old. I don't want to force them needlessly. You can set up a ramdisk and put the BOINC data folder on that; all the disk accesses (both reads and writes) will then be from main memory. I think in Linux it is a virtual disk; you will need enough RAM of course. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Mmm... yes, I can try that. Thanks. One of the computers has only 500 MB of RAM, but probably it will suffice.
----------------------------------------I'll investigate wich BOINC directories I need to chage. [Edit 1 times, last edit by Former Member at Oct 26, 2014 3:56:32 PM] |
||
|
|
|