Posts by wolfman1360

1) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 100701)
Posted 7 Mar 2021 by wolfman1360
Post:
I'd suggest to not suspend it at all and see if some issues occur....if a non-BOINC application needs 100% of the CPU, it will get it.


I might try the exclusion thing that Grant suggested. The reason that I turned off BOINC was that on this laptop, which is optimized for weight and not performance, the latency of the handoff from BOINC to Zoom was a problem, not that it made the handoff at all.

This is what I use, especially for meetings involving video and / or screen share. Just make sure you have enough memory to keep tasks suspended indefinitely without thrashing the swap. Not that Zoom uses a lot.
2) Message boards : Number crunching : Another instance of BOINC is running (Message 99943)
Posted 6 Dec 2020 by wolfman1360
Post:
So a problem I've had lately on Just a few of my Linux machines. Boinc doesn't actually start on system startup. I have to do systemctl restart boinc-client.
This then resulted in MLC at home giving me tons of mount errors telling me it was read only, yet Roseta works just fine. I followed the same procedure - sudo apt-get install boinc virtualbox. This is after adding the unofficial repo - ppa:costamagnagianfranco/boinc.
I have no idea how this has happened or what I did wrong here but attempting to reinstall did no good. With each restart Boinc must be manually restarted. I'm still very much new to Linux and don't really want to go through an entire OS install. I've tried chmod to the boinc directory, that did no good.
Anyone have any ideas?
3) Message boards : Number crunching : 3832 new hosts per day? (Message 99942)
Posted 6 Dec 2020 by wolfman1360
Post:
I might replace the i7-920 with one. Currently the dual Xeon 5520 seems to be holding its own even at a lower clock speed.
Pah! I've got four X5650 CPUs in two motherboards. Not the fastest things ever made, but they do 48 tasks at once between them. They're not even in cases, they're loose on a bookshelf, which alarmed someone once who told me I was being very unsociable by letting all the EM interference out. Maybe that's why my neighbour has installed a weird device on his shed to boost a mobile or Wifi signal :-)

RAC seems to be quicker about increasing since I performed benchmarks on (I think)? everything. Maybe that's just me.
It should do it itself when you install Boinc, change hardware, and also periodically, but I've heard that sometimes it forgets. Boinc is so buggy. Why do programmers expect the users to find their mistakes nowadays? Don't they do testing anymore? And why does this forum keep removing my double spacing between sentences?!

One day I'll have tons of money to burn and can grab me up 2 threadripper 5950x.
Don't, the 5950X costs 2.5 times more than the 3900XT, and only goes 1.38 times faster. Always best to buy about the 5th fastest CPU. Let the millionaires throw money at the new ones to pay the development cost.

The basement of my house is currently a nice warm area, warmer than upstairs and it's winter.
Mine are in the garage (5 soon to be 6). Unfortunately the garage is not (yet) connected to the house, so I'm heating the house with electricity and throwing Boinc heat out of the window of the garage. I must get the tarrif "changed" on that meter....

I have a 3700x and might get a 5600 when they're actually available here. But we'll ultimately see.
I haven't seen more than 360 credits per task at Roseta so I wonder if there is a bottleneck in how much a CPU actually gets done due to no AVX/SSE? i5-3570 and Ryzen 3700 get about the same credits per task.
Not that credits are the entire story. I'm just glad the science is getting done.
4) Message boards : Number crunching : 3832 new hosts per day? (Message 99931)
Posted 6 Dec 2020 by wolfman1360
Post:
Hi
My benchmarks are all odd with Linux - Ubuntu 1804.
I'm far too lazy to actually try and upgrade everything to 20.04 with such a small issue unless there is some super amazing things that it will do for me.
I need to retire a lot of these systems - my poor i7-920 and Xeon w3520. Trying to run them into the ground.


Actually you need to get another Xeon w3520 cpu and put them both in a dual cpu board and bang out twice as many tasks as you are right now.

I might replace the i7-920 with one. Currently the dual Xeon 5520 seems to be holding its own even at a lower clock speed.
RAC seems to be quicker about increasing since I performed benchmarks on (I think)? everything. Maybe that's just me.
One day I'll have tons of money to burn and can grab me up 2 threadripper 5950x.
The basement of my house is currently a nice warm area, warmer than upstairs and it's winter.
5) Message boards : Number crunching : 3832 new hosts per day? (Message 99904)
Posted 4 Dec 2020 by wolfman1360
Post:
Hi
My benchmarks are all odd with Linux - Ubuntu 1804.
I'm far too lazy to actually try and upgrade everything to 20.04 with such a small issue unless there is some super amazing things that it will do for me.
I need to retire a lot of these systems - my poor i7-920 and Xeon w3520. Trying to run them into the ground.
6) Message boards : Number crunching : 3832 new hosts per day? (Message 99877)
Posted 3 Dec 2020 by wolfman1360
Post:
How is it my i3-2130 is making more credits on work than an i7-4790 which is averaging about 90 or 100 credits per WU.
That’s at least in part because the i7 hasn’t sent Rosetta its performance measurements yet.

Advanced view » Tools menu » Run CPU benchmarks

Ugh. I thought Boinc did this automatically. One more bug.
Thanks for letting me know I'll go ahead and do this now.
7) Message boards : Number crunching : 23.3 GB RAM per Rosetta WU (Message 99867)
Posted 2 Dec 2020 by wolfman1360
Post:
Thanks for the heads up.
Going through my tasks to search for these.
Hopefully these get yanked from the server sooner than later. A lot of my machines can handle around 2 gb per thread. The Ryzen 3700x can handle about 4, as well as a few others.
8) Message boards : Number crunching : 3832 new hosts per day? (Message 99838)
Posted 30 Nov 2020 by wolfman1360
Post:
Left it at the default 8 hours. I'd guess they know best as far as what they want.
From a purely selfish point of view, I prefer ones with a variable time (fixed amount of processing), then I can see them flying through on a faster machine!

I can see that for sure.
It's hard to figure out amount of work done proportionate to each machine on this project since the credits vary so drastically from one type of workunit to the next. How is it my i3-2130 is making more credits on work than an i7-4790 which is averaging about 90 or 100 credits per WU. I guess as long as science is being done that's what matters.
9) Message boards : Number crunching : 3832 new hosts per day? (Message 99818)
Posted 30 Nov 2020 by wolfman1360
Post:

Right now I'm trying to figure out if running workunits for the default 8 hours or 24 hours is more worthwhile.


There's not much difference in the ratio of credits you get between 8 hour tasks or the 24 hour tasks, you just get 3 times as many credits for the 24 hours tasks.
I think you get credits for the number of results you produce within the task. This has a random element (some tasks are more difficult), and is related to how fast your processor is and how long you run the task for. I think they've set it to 8 hours so they get about the right number of results for each task - I remember one of the scientists saying something about all the task results produce a big picture so if you run yours for longer you're in effect making one of the pixels sharper. If a lot of people started running them longer, they'd get too much information, so they'd turn down the setting to 7 hours standard.

This makes sense. I have a few tasks that, regardless of the machine, get 66 credits or there about. I'm assuming this is the base credit a task can produce. I guess it depends on the specific task type to go by decoy count - some get 20, some get over 100.
Left it at the default 8 hours. I'd guess they know best as far as what they want.
10) Message boards : Number crunching : What are your tips for new Rosetta@home users? (Message 99809)
Posted 28 Nov 2020 by wolfman1360
Post:
Should I set both values to 0 e.g. keep at least and keep up to an additional?
Both to 0 or close to it, eg 0.01 & 0.01). If one project is out of work for a while, other projects will take it's place. When it gets work again, BOINC will process more of it for a while to balance out things according to your Resource share settings. So no need to cache work.

Thanks! Set that now and will see how it works out.
11) Message boards : Number crunching : What are your tips for new Rosetta@home users? (Message 99801)
Posted 28 Nov 2020 by wolfman1360
Post:
I'm trying to remember if Credit per machine is relative to its score on this project, or if it's luck of the draw. I think it's the latter and I should just ignore it.
My 4790 is doing much worse than either of my 4770s credit wise. Even my two little 3570s's are doing better, actually they're probably getting the most credit out of all my CPU's apart from the 1 Ryzen 3700x.
Running just a single project 24/7 with no cache it takes around 6-8 weeks for RAC to level off. The more projects you run, and the larger your cache, the longer it will take for your RAC to stabilise around it's nominal value. Set you cache to zero, and let things be for 2 months and your RAC for each system should get to it's nominal value by then end of that 2 months.

Just how much RAC you will get for any given system will depend heavily on clock speed, processor model & version, and how much time a system has to actually process work. 2 systems of the same specs will eventually end up with RACs within a few percent of each other, if they also have the same up time & processing time.
Thank you. Should I set both values to 0 e.g. keep at least and keep up to an additional?



Is it worth upping the runtime and running for 24 hours, or is 8 hours doing just fine for most folks?
There is no point in changing the runtime. The Credit per hour is (pretty much) the same regardless of how long the Task runs for- running 1 Task for 24 hours will result in the same Credit as running 3 Tasks of the same type for 8 hours each. However processing 3 different Tasks will provide more different useful results than running just the 1 Task over the same time frame as running the 3 different ones.
The project chose 8 hours because it gives them a suitable result in the time frame they want, so it's best to stick with that. If a Task needs more time, it will run for longer (up to an additional 10 hour limit).[/quote]
Exactly the info I was looking for. Right now I'm crunching this and WCG ARP, which take significantly longer than Roseta, as well as 1 or 2 CPDN (given l3 cache limitations with their N216 especially, and a lot of my processors aren't the most modern).
12) Message boards : Number crunching : 3832 new hosts per day? (Message 99799)
Posted 28 Nov 2020 by wolfman1360
Post:
The Tasks ready to send I assume are a buffer that the server maintains from which it can immediately service client requests for new work, with files staged ready for download. Total queued jobs on the front page is over 15 000 000 – so at a completion rate of around 25 000 per hour, we’re not at risk of running out just yet…

Signal 11 could be just about anything – faulty hardware, overheating, instability from overclocking, …
There do seem to be a lot of reports of trouble from hosts running Linux on AMD; I wonder whether there’s some obscure bug lurking somewhere.


The Ryzens appear to be fine and I had no problems with my Bulldozer but I couldn’t speak for the Opteron.

I have one bulldozer and 2 Ryzen's with 1 Ryzen running Windows and the other 2 processors running Ubuntu 1804.
Right now I'm trying to figure out if running workunits for the default 8 hours or 24 hours is more worthwhile.
The Opteron 6128 and 4122 are getting signal 11 errors on this project but are working just fine on TN Grid.
13) Message boards : Number crunching : What are your tips for new Rosetta@home users? (Message 99798)
Posted 28 Nov 2020 by wolfman1360
Post:
I'm trying to remember if Credit per machine is relative to its score on this project, or if it's luck of the draw. I think it's the latter and I should just ignore it.
My 4790 is doing much worse than either of my 4770s credit wise. Even my two little 3570s's are doing better, actually they're probably getting the most credit out of all my CPU's apart from the 1 Ryzen 3700x.
Is it worth upping the runtime and running for 24 hours, or is 8 hours doing just fine for most folks?
14) Message boards : Number crunching : 3832 new hosts per day? (Message 99790)
Posted 27 Nov 2020 by wolfman1360
Post:
Always right around 26000 tasks to send and 500000 in progress.
Either the task generator is barely keeping up or it just always keeps 26000 tasks pending.
It would be really nice if I could figure out why AMd Opteron processors (6128 in particular) keep getting signal 11 errors on this project. Probably time to retire the old fella anyway but still.
15) Message boards : Number crunching : horns named project files causing pgtables out of memory errors in boinc all others run fine (Message 99776)
Posted 26 Nov 2020 by wolfman1360
Post:
The real question is why do they allow a Pi at all? It is bound to fail at some point.

Not only Raspberry Pi, but also smarphone and pc with less than 4bg per core...

Because most WUs take 1 gb per core if not a bit less.
I wish there was a way to separate out these larger requirements for WUs on this project. I have quite a variety of machines and it's frustrating to have super huge WUs pop up unexpectedly. RPI or not, smartphone or not, everything should be able to contribute. Especially now.
16) Message boards : Number crunching : Underclocking (Message 98635)
Posted 23 Aug 2020 by wolfman1360
Post:
Hmm, this is very interesting! I might have to do this with my Ryzens.
Keep in mind, though, depending on the workunits, credit and decoy count may be far different. It's hard to gauge performance on here unless workunits are the same across all machines. Which is unfortunate.
17) Message boards : Number crunching : No Alt Platform (Message 98586)
Posted 19 Aug 2020 by wolfman1360
Post:
Does Boinc periodically run benchmarks on its own or is this something I should do myself every so often?
Thanks.
18) Message boards : Number crunching : Running multiple covid projects on multiple machines, how do you personally go about it? (Message 98521)
Posted 15 Aug 2020 by wolfman1360
Post:
Thank you for such valuable input, everyone! It is very much appreciated.
I've had to switch off some of my older machines, too - in particular a dual Opteron 6128.
I will more than likely attempt the nanoHUB project as well as a few others Jim1348 mentioned. I was giving all my resources to MLC previously, since it was a new project and I really like the- owner gives a lot of communication in regards to the status and tries to help wherever he can.

LHC - I should really give that a try again. Seems I had some issues the last time I attempted it.
Thanks again, everyone. Right now I just have everything connected to this project but I'll slowly start branching out in the future, especially when WCG comes out with GPU work.
19) Message boards : Number crunching : Running multiple covid projects on multiple machines, how do you personally go about it? (Message 98464)
Posted 10 Aug 2020 by wolfman1360
Post:
I spent all winter upgrading machines even before COVID, and then added a couple more after it hit (11 total). They are all on Ubuntu 18.04, except my main machine on Win7 64-bit.

My somewhat random thoughts:

    Normally I just devote all the cores of a machine to a given project after finding out which architecture it runs best on. Sometimes you have to mix projects for best efficiency, but that is not necessary for the ones I am running now except for WCG/ARP, which is why I am mixing it with Rosetta.
    I like short caches, and use the default of 0.1+0.5 days, except on some projects with long run times where I go to 0.25+0.5 or a little longer to ensure that they download properly.
    I am mainly on Rosetta (two Ryzen 3900X and two Ryzen 3950X) on almost all cores, but I am allowing WCG/ARP to run on 12 of them, so that leaves 100 for Rosetta.
    I do a lot of Folding too on GPUs, but am down to only one GTX 1070 at the moment until cooler weather arrives. By the way, the Ryzen 3900X and 3950X do great on the CPU version of Folding, but they are more valuable on Rosetta since I can use the GPUs on Folding.
    I think Folding and Rosetta do the most innovative and advanced science.

    However, even though they are working very hard, you won't seem them in use for the first round of anti-virals. The first round is actually in testing now, and there should be several available by the end of the year (look for the mono-clonal antibodies; they should do well).
    I do TN-Grid also on an i7-8700.



While I like WCG/OPN, I am waiting for the GPU version. I don't want to tie up my CPU cores on something that can run much faster on GPUs.
Insofar as I can determine, Ibercivis, is just autodock on CPUs. I think it is fairly duplicative of other efforts, and not nearly as fast as a GPU version would be.

But I still run some non-COVID projects. The COVID ones are really for a next-generation anti-viral, which is important but won't make it to market soon enough for this round.


Thank you for that - this is exactly what I'm looking for.
Is the reason you run TN grid on the 8700 due to it having better avx support?
Those are some nice Ryzen processors, there I wish I could afford that, though I'm hoping to get another 3600 or 3800x soon. Paired with the two 3rd gen ryzens are now 64 gb of it, which long term should be okay, but the ram was almost more expensive than the 3800x over here, let alone the 3600.
How do you run ARP on 12 threads consistently? Do you use an app_config or do you just have wcg set to 12 at once on the machines you run it on, with the resource shares for Rosetta and WCG at their default (50% each)?
I don't have much for GPU's - just an rx570 for right now.
What other projects are you running, CPU wise, out of curiosity? Have you tried the QuChem project, which I believe requires virtual box and is also working on Covid related science?
20) Message boards : Number crunching : Running multiple covid projects on multiple machines, how do you personally go about it? (Message 98457)
Posted 10 Aug 2020 by wolfman1360
Post:
Hi Wolfman,
As different Projects are approaching the problem from different directions, to develop new drugs or test the effectiveness of existing ones, I think it's good to back all the dogs in the fight. Because of those different strategies, I don't think it's possible to determine which hardware is "better" at which Project, other than Boinc credits awarded per day or per hour, but the virus doesn't care about credits!
Even with encouraging trials ongoing, I think there's still a long way to go and hopefully something even better than those candidates waiting to be discovered.
With lots of hosts, the simplest way might be to have each host doing different Projects rather than setting resource shares in Boinc and letting it switch between.
I have only 2 hosts active just now, both having been switched from LHC to doing purely Covid work (for an old uncle that I lost to it).
This one, running 3of4 threads, is doing Rosetta, WCG and Ibercivis (which is out of work again) but rather than letting Boinc switch mid-task, I have "Switch between ...." set to 1500 mins so that any individual task will run to completion.
The other one is running exclusively Folding@Home on 3of4 threads. That is not Boinc so Boinc doesn't see the threads as being used so tries to use them itself (unless restricted).
I don't have anything with a useable Graphics card so haven't tried TNG so others would have to input on balancing resources re.CPU/GPU.

Stay safe

Thank you both for the recommendations.
How much memory does Ibercivis take up as far as ram usage?
The other project which I am referencing can be found at TNGrid

I'll definitely take a look at all of these - I'll probably add the machines with more ram to Rosetta as it does like to use at least 1 GB per thread. I don't have many GPUs, however Folding will likely be added to the machines that can use it.

thanks for all the suggestions again.


Next 20



©2021 University of Washington
https://www.bakerlab.org