Personal computing discussed
Moderators: renee, farmpuma, just brew it!
thegleek wrote:I haven't folded in so long... 1/2 decade perhaps? I may be interested in starting some boxes up again soon...
Haitch wrote:The barrier fix will also work for XFS.
The problem with the filesystems is specific to those that have barriers enabled, and is a problem of Stanfords making. After completing the WU the client 0's out two ( or more, don't recall all the murky details) of the temp files, one byte at a time, and some of the files are over 100MB. The filesystems with barriers enabled take a long time to do this. EXT3 with no barrier support does it a lot faster.
H.
just brew it! wrote:The way to get the really big points these days is with the GPU client. I tend not to buy high-end GPUs (I prefer ones with passive cooling)
just brew it! wrote:so I just run the Linux SMP client. Linux SMP tends to be "fire and forget" once you get it set up properly (though initial setup can be somewhat fiddly depending on distro); as long as you've got a reasonably recent multi-core CPU it posts some respectable points.
thegleek wrote:just brew it! wrote:The way to get the really big points these days is with the GPU client. I tend not to buy high-end GPUs (I prefer ones with passive cooling)
Any issues ever arise from running this on your video card 24/7? I have a NVIDIA GeForce GTS 450, is that adequate enough to fold?
Ragnar Dan wrote:That zeroing out thing is an odd way of working, indeed.
Ragnar Dan wrote:It seems like they need a rewrite, though, if they're making 100MB temp files these days.
Ragnar Dan wrote:I run the same video card (but with 930 MHz shader), and get about 9000 PPD on it with no trouble. I've been running it since sometime last November and it doesn't seem to have any problems.
just brew it! wrote:Ragnar Dan wrote:That zeroing out thing is an odd way of working, indeed.
Sounds like they want to remove all traces of the just-completed WU from the system for some reason. If that's truly their intent, what they're doing probably doesn't even accomplish that on a modern journaling file system.
just brew it! wrote:Ragnar Dan wrote:It seems like they need a rewrite, though, if they're making 100MB temp files these days.
100MB is nothing; done in a sane manner, it should take only a couple of seconds to write that much data on a modern system. 100MB just *sounds* like a lot because you're old and have first-hand memories of the days when systems had 640KB (or less) of RAM! (And yes, I'm getting frikkin' old too; don't remember if I'm older than you or not though... )
thegleek wrote:Ragnar Dan wrote:I run the same video card (but with 930 MHz shader), and get about 9000 PPD on it with no trouble. I've been running it since sometime last November and it doesn't seem to have any problems.
Oh god... SO many choices! WTH!
http://folding.stanford.edu/English/DownloadWinOther
WHICH one do I download?!?!
Ragnar Dan wrote:My thinking was more of the "this 100 MB meant something back a decade ago, but now it's a pittance and shouldn't be considered worth worrying about" sort, not the other way around.
Ragnar Dan wrote:And, the first machine I used had 4k of RAM, bucko. I'm not old. At least I didn't learn about the S-100 stuff until it was mostly over.
just brew it! wrote:Ragnar Dan wrote:My thinking was more of the "this 100 MB meant something back a decade ago, but now it's a pittance and shouldn't be considered worth worrying about" sort, not the other way around.
Ahh, OK. But they still need to be able to maintain state that can persist across client restarts and reboots (i.e. checkpoints), yes?
just brew it! wrote:Ragnar Dan wrote:And, the first machine I used had 4k of RAM, bucko. I'm not old. At least I didn't learn about the S-100 stuff until it was mostly over.
Touche!
Ragnar Dan wrote:But the Z-80 was where I first learned assembLY language, and used an assmblER to assemble my programs (yes that's a pet peeve of mine. Ahem). Fun stuff.
Ragnar Dan wrote:Columbia portable (suitcase machine).
thegleek wrote:Ragnar Dan wrote:Columbia portable (suitcase machine).
Back in '83, tt was the [url=http://en.wikipedia.org/wiki/Compaq_Portable]Compaq Portable[/rul] you are referring to!
I still have mine in 100% working condition!
thegleek wrote:But alas, my FIRST computer was a year prior to that computer, the Timex Sinclair 1000 ($99.95 deal!). The CPU was called a Zilog Z80A (@ 3.25Mhz), whatever that is!
Haitch wrote:This is a known issue if you install Lunux with an Ext4 file system, the default with Ubuntu and others.
Quickest fix is to reinstall with ext3. It may be also possible to fix the issues by remounting the filesystem with "nobarriers", but I don't have any instructions for that.
H.
just brew it! wrote:You can try adding the "nobarrier" option to your /etc/fstab file. If there's a line with a mount point of "/home", add it to the <options> column of that line, otherwise add it to the line with a mount point of "/". If the existing <options> entry says "defaults" then you should replace "defaults" with "nobarrier"; otherwise, append "nobarrier" to the existing option(s), separated from them by a comma. Reboot the system to have the new options take effect.
Haitch wrote:The barrier fix will also work for XFS.
The problem with the filesystems is specific to those that have barriers enabled, and is a problem of Stanfords making. After completing the WU the client 0's out two ( or more, don't recall all the murky details) of the temp files, one byte at a time, and some of the files are over 100MB. The filesystems with barriers enabled take a long time to do this. EXT3 with no barrier support does it a lot faster.
H.
[09:29:39] Completed 250000 out of 250000 steps (100%)
[09:29:51] DynamicWrapper: Finished Work Unit: sleep=10000
[09:30:01]
[09:30:01] Finished Work Unit:
[09:30:01] - Reading up to 52629024 from "work/wudata_01.trr": Read 52629024
[09:30:01] trr file hash check passed.
[09:30:01] - Reading up to 42749468 from "work/wudata_01.xtc": Read 42749468
[09:30:01] xtc file hash check passed.
[09:30:01] edr file hash check passed.
[09:30:01] logfile size: 284743
[09:30:01] Leaving Run
[09:30:02] - Writing 95833587 bytes of core data to disk...
[09:30:02] ... Done.
[09:52:35] - Shutting down core
[09:52:35]
[09:52:35] Folding@home Core Shutdown: FINISHED_UNIT
[09:54:38] CoreStatus = 64 (100)
[09:54:38] Unit 1 finished with 48 percent of time to deadline remaining.
Folding Status wrote:Project: 6098 (Run 5, Clone 72, Gen 5)
...
1593.00 point GRO-A3 (100 frames), 2364.87 PPD, 00:09:42/step, 02:35:12 remaining, 12027.54 total points w/bonus (17855.32 PPD)
Ragnar Dan wrote:That's quite impressive. I forget what you may have written about that proc's speed/OC if any,
just brew it! wrote:This pile o' parts is currently responsible for nearly half of my ~50K PPD F@h production.