Outage notice

Message boards : News : Outage notice

To post messages, you must log in.

Previous · 1 . . . 7 · 8 · 9 · 10 · 11 · 12 · 13 · Next

AuthorMessage
mikey
Avatar

Send message
Joined: 5 Jan 06
Posts: 1894
Credit: 8,536,986
RAC: 9,927
Message 99572 - Posted: 5 Nov 2020, 0:53:54 UTC - in response to Message 99567.  

I didn't realise we were talking about a setting on the server. Whenever I join something like a project or play a new computer game, etc, etc, the first thing I do is adjust all the settings, so I'll have done that a long time ago when I used WCG in the past. Do you mean in the device profile, under project limits? All mine say unlimited, I probably did that a while back.


Yes, under the default device profile, Project Limit for African Rainfall Project showed "1" - now changed to "Unlimited"
I'd never been into the default profile option before for any reason, let alone set a custom default profile.
I've never really understood WCG so I've previously left well alone.[/quote]

Just be careful as they take alot of memory and running ie 24 of them on a Threadripper can quickly run out of memory.
ID: 99572 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1965
Credit: 38,162,917
RAC: 9,188
Message 99573 - Posted: 5 Nov 2020, 4:44:49 UTC - in response to Message 99572.  

Just be careful as they take alot of memory and running ie 24 of them on a Threadripper can quickly run out of memory.

I wish!
I don't think they're any bigger than the Rosetta tasks we've had here in the past, are they?
ID: 99573 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile robertmiles

Send message
Joined: 16 Jun 08
Posts: 1223
Credit: 13,824,497
RAC: 2,340
Message 99585 - Posted: 5 Nov 2020, 23:55:39 UTC - in response to Message 99573.  

Just be careful as they take alot of memory and running ie 24 of them on a Threadripper can quickly run out of memory.

I wish!
I don't think they're any bigger than the Rosetta tasks we've had here in the past, are they?

Didn't you notice that the message you replied to wasn't about Rosetta tasks?
ID: 99585 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
sombraguerrero

Send message
Joined: 5 Mar 18
Posts: 8
Credit: 617,193
RAC: 0
Message 99601 - Posted: 7 Nov 2020, 7:18:39 UTC - in response to Message 99573.  
Last modified: 7 Nov 2020, 7:20:09 UTC

WCG ARP units can get upwards of about 3/4 of a Gigabyte when they're fresh, possibly closer to a Gig at points. So it's not too much for most modern systems to handle, but it's enough that most people probably wouldn't want more than one. Rosetta units do tend to stay pretty steadily in the lower hundreds of Megabytes. It's just a question of knowing how much RAM you have and whether or not you tend to need RAM for a lot of other things.
ID: 99601 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Grant (SSSF)

Send message
Joined: 28 Mar 20
Posts: 1467
Credit: 14,316,187
RAC: 16,257
Message 99602 - Posted: 7 Nov 2020, 7:48:54 UTC - in response to Message 99601.  

Rosetta units do tend to stay pretty steadily in the lower hundreds of Megabytes.
At the moment. There have been times when most of them were around 1GB, with some around 3GB.
Allowing 1.3GB of RAM per Task means you won't run in to RAM issues even when there are lots of high RAM usage Tasks being run. If you've got less than 1GB of RAM per Task then you will run in to a lack of RAM issue at some point.
Grant
Darwin NT
ID: 99602 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1965
Credit: 38,162,917
RAC: 9,188
Message 99621 - Posted: 9 Nov 2020, 7:53:59 UTC - in response to Message 99585.  

Just be careful as they take alot of memory and running ie 24 of them on a Threadripper can quickly run out of memory.

I wish!
I don't think they're any bigger than the Rosetta tasks we've had here in the past, are they?

Didn't you notice that the message you replied to wasn't about Rosetta tasks?

Er… yes. That's why I said "they" (the African Rainfall tasks) are no bigger than the large Rosetta tasks we had here.
My point being, if I handled the 1.4Gb & 2.3Gb Rosetta tasks, I'd be fine with <1Gb African Rainfall tasks...
ID: 99621 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Brian Nixon

Send message
Joined: 12 Apr 20
Posts: 293
Credit: 8,432,366
RAC: 0
Message 99642 - Posted: 11 Nov 2020, 15:50:08 UTC - in response to Message 99602.  

Rosetta units do tend to stay pretty steadily in the lower hundreds of Megabytes.
At the moment. There have been times when most of them were around 1GB, with some around 3GB.
horns tasks in the current batch running around 4 GB each [example]
ID: 99642 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jim1348

Send message
Joined: 19 Jan 06
Posts: 881
Credit: 52,257,545
RAC: 0
Message 99643 - Posted: 11 Nov 2020, 16:02:40 UTC - in response to Message 99642.  

horns tasks in the current batch running around 4 GB each [example]

Yes, I have several. The biggest is 4.4 GB, the largest Rosetta WU I have ever seen.
That is great. I size my machines to handle them.
ID: 99643 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Brian Nixon

Send message
Joined: 12 Apr 20
Posts: 293
Credit: 8,432,366
RAC: 0
Message 99644 - Posted: 11 Nov 2020, 16:30:37 UTC - in response to Message 99643.  

That is great.
As long as we don’t get only tasks that big…

I wonder how smart the server is at bundling tasks of different types when sending work to individual hosts – does it actively avoid sending all tasks of the same type, or is it random chance that we (tend to) get a mix of different types?
ID: 99644 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jim1348

Send message
Joined: 19 Jan 06
Posts: 881
Credit: 52,257,545
RAC: 0
Message 99645 - Posted: 11 Nov 2020, 16:40:24 UTC - in response to Message 99644.  

I wonder how smart the server is at bundling tasks of different types when sending work to individual hosts – does it actively avoid sending all tasks of the same type, or is it random chance that we (tend to) get a mix of different types?

For the most part I get a good mix. That is probably just random, since they have a lot of sources.
If you have 1 GB/core, you are usually OK, though I prefer 2 GB.
ID: 99645 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1965
Credit: 38,162,917
RAC: 9,188
Message 99646 - Posted: 11 Nov 2020, 23:24:12 UTC - in response to Message 99642.  

Rosetta units do tend to stay pretty steadily in the lower hundreds of Megabytes.
At the moment. There have been times when most of them were around 1GB, with some around 3GB.
horns tasks in the current batch running around 4 GB each [example]

Blimey!
I'm not running any atm, but I can see 4 that'll run together at some point tomorrow morning. I'll look out for that.
5 of my 8 current-running tasks are consuming less than 100Mb each. Hopefully I'll have a few of those going through around the same time as the horns ones
ID: 99646 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Grant (SSSF)

Send message
Joined: 28 Mar 20
Posts: 1467
Credit: 14,316,187
RAC: 16,257
Message 99652 - Posted: 12 Nov 2020, 9:07:59 UTC - in response to Message 99643.  

horns tasks in the current batch running around 4 GB each [example]

Yes, I have several. The biggest is 4.4 GB, the largest Rosetta WU I have ever seen.
Just went through my list of Valids and found one over 5GB.
I think 2GB was the smallest.
Grant
Darwin NT
ID: 99652 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Falconet

Send message
Joined: 9 Mar 09
Posts: 350
Credit: 1,000,634
RAC: 0
Message 99655 - Posted: 12 Nov 2020, 19:06:23 UTC

Looks like I have one Horns task using just over 1 GB and another using around 3.3 GB.
Wonder what these are about.
ID: 99655 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jim1348

Send message
Joined: 19 Jan 06
Posts: 881
Credit: 52,257,545
RAC: 0
Message 99656 - Posted: 12 Nov 2020, 19:30:57 UTC - in response to Message 99655.  

Wonder what these are about.
That is what I miss on Rosetta. They could at least give a brief description when a new project comes out.
They have no problem with that on Folding or WCG, but it seems to be an insurmountable obstacle here.
ID: 99656 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Falconet

Send message
Joined: 9 Mar 09
Posts: 350
Credit: 1,000,634
RAC: 0
Message 99657 - Posted: 12 Nov 2020, 21:26:52 UTC - in response to Message 99656.  

Yeah, it would be nice to know.
Sometimes, you can tell what it is about.

COVID is COVID, MERS is coronavirus as well, VEGF seemed to have something to do with blood vessels, etc
CTLA4, I believe I saw COVID on a result log the other day.

But horns, I can't even attempt to guess what it may be lol
ID: 99657 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Breno

Send message
Joined: 8 Apr 20
Posts: 28
Credit: 10,835,843
RAC: 10,182
Message 99658 - Posted: 13 Nov 2020, 0:11:39 UTC - in response to Message 99656.  

I suggested this a while ago, maybe the solution would be adding a new column on the client manager.
If no String is provided, shows None. Otherwise, the folks who generate the work units could assign a content such as: COVID, FLU, AIDS, CANCER, ALL VIRAL, SOME VIRAL, BACTERIAL, PARASITE, OTHER.
But, we also know that just as adding GPU towards the WU, this is not a trivial task, and it requires some changes.
Meanwhile, I also hope they include something in these lines.
ID: 99658 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mrchips

Send message
Joined: 11 Nov 09
Posts: 8
Credit: 11,411,237
RAC: 13,356
Message 99661 - Posted: 13 Nov 2020, 0:57:38 UTC

Just got "error reported by file upload server: Server is out of disk space"
ID: 99661 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mikey
Avatar

Send message
Joined: 5 Jan 06
Posts: 1894
Credit: 8,536,986
RAC: 9,927
Message 99663 - Posted: 13 Nov 2020, 4:24:10 UTC - in response to Message 99658.  

I suggested this a while ago, maybe the solution would be adding a new column on the client manager.
If no String is provided, shows None. Otherwise, the folks who generate the work units could assign a content such as: COVID, FLU, AIDS, CANCER, ALL VIRAL, SOME VIRAL, BACTERIAL, PARASITE, OTHER.
But, we also know that just as adding GPU towards the WU, this is not a trivial task, and it requires some changes.
Meanwhile, I also hope they include something in these lines.


The problem has always been they want us to crunch ALL the tasks not just the ones we think we like, if they did something like WCG or PrimeGrid does with the different sub-projects then yes it would work but it's not like they have alot of tasks to begin with so splitting the project wouldn't be ideal right now.
ID: 99663 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1965
Credit: 38,162,917
RAC: 9,188
Message 99678 - Posted: 14 Nov 2020, 7:33:38 UTC - in response to Message 99661.  

Just got "error reported by file upload server: Server is out of disk space"

I finally reduced my runtime back to the default now we have a queue of tasks stretching into the millions. I'm not helping...
ID: 99678 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1965
Credit: 38,162,917
RAC: 9,188
Message 99679 - Posted: 14 Nov 2020, 9:15:59 UTC - in response to Message 99652.  

horns tasks in the current batch running around 4 GB each [example]

Yes, I have several. The biggest is 4.4 GB, the largest Rosetta WU I have ever seen.
Just went through my list of Valids and found one over 5GB.
I think 2GB was the smallest.

I've just checked mine and all went through fine on my desktop and laptop (3-4Gb), but on my mobile, even though I've set it to run only on 2 cores out of 8, I've had insufficient RAM to run 1 horns and 1 ordinary task, so I'm having to abort them on arrival unfortunately
ID: 99679 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Previous · 1 . . . 7 · 8 · 9 · 10 · 11 · 12 · 13 · Next

Message boards : News : Outage notice



©2024 University of Washington
https://www.bakerlab.org