climateprediction.net home page
New work Discussion

New work Discussion

Message boards : Number crunching : New work Discussion
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 86 · 87 · 88 · 89 · 90 · 91 · Next

AuthorMessage
Profile Iain Inglis
Volunteer moderator

Send message
Joined: 16 Jan 10
Posts: 1079
Credit: 6,904,049
RAC: 6,657
Message 65802 - Posted: 11 Aug 2022, 8:33:35 UTC - in response to Message 65801.  
Last modified: 11 Aug 2022, 8:37:49 UTC

Your axe to grind with Apple is well noted, and is increasingly irrelevant here. Are you going to rewrite the 32-bit MacOS stuff for 64-bit and ARM? No? Then quit complaining about how a company you don't buy anything from is doing things you don't like. Bitching about it here isn't going to change a single thing about either the tasks to run, or Apple's practices.
Might aswell point out to everyone using Apple products that they shouldn't be. Selling things which are obsolete after a short time is unethical, criminal, and environmentally unfriendly.

I run Windows and only Windows, because it's the easiest to use, the most commonly supported for programs, and doesn't ditch its users. I see no need to use anything else. Windows 11 on 7 PCs here.

This thread is indeed intended to be for the discussion of new work that appears on CPDN rather than platform comparisons, but in contradiction of that I would say in Apple’s defence that because they supply both hardware and software they also offer trade-ins, which Microsoft cannot in general — when there are no more 32-bit Mac models then I’ll trade in my ancient Mac mini, if they’ll still take it: I cannot do that with any of my PCs. On the other hand, I do have some important applications whose functionality erodes with each Windows version change. So I rather fancy the Microsoft/Apple comparison could just express a bias based on individual use.

I develop software for Windows, so I have a Windows machine; I have old software that won’t run under current Windows so I have an old Windows machine; I have some applications that are better on Mac so I have a Mac; until the power supply blew once too often, I had a Linux machine for yet more stuff. It’s horses for courses and I see no reason to suppose that we all have or should have exactly the same horse.
ID: 65802 · Report as offensive
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 65803 - Posted: 11 Aug 2022, 12:52:18 UTC

And there's now 3,500 Windows tasks on the server.
ID: 65803 · Report as offensive
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 65805 - Posted: 11 Aug 2022, 14:49:37 UTC

Not sure why the Linux batch #935 shows a submission date of the 21st of last month. I know I said I expected a further batch from that experiment though.
ID: 65805 · Report as offensive
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2167
Credit: 64,403,322
RAC: 5,085
Message 65806 - Posted: 11 Aug 2022, 16:02:43 UTC - in response to Message 65805.  

Not sure why the Linux batch #935 shows a submission date of the 21st of last month. I know I said I expected a further batch from that experiment though.

I think Sarah just made that comment on the Trello board for bookkeeping purposes. The batch had already been released in July, it just hadn't been properly placed on the Trello boards.
ID: 65806 · Report as offensive
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 65810 - Posted: 12 Aug 2022, 5:25:30 UTC

And all gone, both Linux and Windows tasks.
ID: 65810 · Report as offensive
Glenn Carver

Send message
Joined: 29 Oct 17
Posts: 774
Credit: 13,433,329
RAC: 7,110
Message 65977 - Posted: 24 Aug 2022, 17:16:08 UTC - in response to Message 65757.  

Woah.

Is there any way to make those opt-in? That's massive, even for CPDN, and I would wager not many people have enough RAM to run more than one of those at a time. One of my dedicated compute boxes could only run one of those, and the other might manage three...
Yes, OpenIFS, as it's a global model, and efficient computationally, it requires RAM. Only machines that have the required hardware will get workunits though, so noone should be compromised if they get one.

The point is though, as great an idea as BOINC is, unless the models in use are scientifically useful there's little point in applying them to the Boinc framework. Scientifically, higher resolution, and hence hardware requirements, are where we need to be to make the platform useful.
ID: 65977 · Report as offensive
Jim1348

Send message
Joined: 15 Jan 06
Posts: 637
Credit: 26,751,529
RAC: 653
Message 65978 - Posted: 24 Aug 2022, 17:44:31 UTC - in response to Message 65977.  

I expect that a lot of us have enough RAM, bandwidth, cores, etc.
The tendency of most BOINC projects is to cater to the minimalist user.

I see no reason to do that here. Use as many resources as you need and see what happens.
It is either enough to be scientifically useful, or it isn't. But you won't know if you limit yourself unnecessarily.
ID: 65978 · Report as offensive
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 65979 - Posted: 24 Aug 2022, 18:21:44 UTC

Yes, OpenIFS, as it's a global model, and efficient computationally, it requires RAM. Only machines that have the required hardware will get workunits though, so noone should be compromised if they get one.


It would be good to have the chance for the few people with metered connections to opt out of these. While not as fast as I might wish, I do have unlimited data. Those few with metered connections could quickly run into problems.
ID: 65979 · Report as offensive
Glenn Carver

Send message
Joined: 29 Oct 17
Posts: 774
Credit: 13,433,329
RAC: 7,110
Message 65982 - Posted: 24 Aug 2022, 20:00:03 UTC - in response to Message 65979.  

Yes, OpenIFS, as it's a global model, and efficient computationally, it requires RAM. Only machines that have the required hardware will get workunits though, so noone should be compromised if they get one.
It would be good to have the chance for the few people with metered connections to opt out of these. While not as fast as I might wish, I do have unlimited data. Those few with metered connections could quickly run into problems.
I don't know if the server checks the network capacity in deciding which computers get the tasks, I'll ask Andy, but the user can always just abort the task. Expected data volumes would be published on the forums.
ID: 65982 · Report as offensive
Mr. P Hucker

Send message
Joined: 9 Oct 20
Posts: 669
Credit: 4,391,754
RAC: 6,918
Message 65983 - Posted: 24 Aug 2022, 20:31:28 UTC - in response to Message 65982.  

I don't know if the server checks the network capacity in deciding which computers get the tasks, I'll ask Andy, but the user can always just abort the task. Expected data volumes would be published on the forums.
Perhaps the largest ones could be opt-in in preferences?
ID: 65983 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1045
Credit: 16,506,818
RAC: 5,001
Message 65984 - Posted: 24 Aug 2022, 21:17:43 UTC - in response to Message 65978.  

I expect that a lot of us have enough RAM, bandwidth, cores, etc.
The tendency of most BOINC projects is to cater to the minimalist user.

I see no reason to do that here. Use as many resources as you need and see what happens.
It is either enough to be scientifically useful, or it isn't. But you won't know if you limit yourself unnecessarily.


I agree.

I have 64 GBytes of RAM, 16896 KB cache, 75 Megabit/second up and down Internet bandwidth, 16 cores (I allow only 8 for BOINC).

Measured floating point speed 6.58 billion ops/sec
Measured integer speed 30.58 billion ops/sec
Average upload rate 181.06 KB/sec
Average download rate 6147.47 KB/sec

My machine is only lightly loaded at the moment because ClimatePrediction, WCG, and Rosetta all seem to be sending out few or no work units.
ID: 65984 · Report as offensive
Mr. P Hucker

Send message
Joined: 9 Oct 20
Posts: 669
Credit: 4,391,754
RAC: 6,918
Message 65985 - Posted: 24 Aug 2022, 21:35:51 UTC - in response to Message 65984.  

My machine is only lightly loaded at the moment because ClimatePrediction, WCG, and Rosetta all seem to be sending out few or no work units.
There are still lots of Virtualbox Rosettas to do. Unfortunately you need a modern processor, so only 2 of my 7 can do it (it needs AVX).
ID: 65985 · Report as offensive
Mr. P Hucker

Send message
Joined: 9 Oct 20
Posts: 669
Credit: 4,391,754
RAC: 6,918
Message 65988 - Posted: 24 Aug 2022, 23:12:16 UTC

WCG has loads to do, but apparently a 14 billion dollar company can't provide a server fast enough to cater for Boinc. I didn't think Canadians were quite that incompetant. Their server is so slow you can't even get onto their forum. I'm assuming it's a 486 in somebody's shed.
ID: 65988 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1045
Credit: 16,506,818
RAC: 5,001
Message 65989 - Posted: 25 Aug 2022, 5:38:21 UTC - in response to Message 65988.  

If I told you once, I told you a million times: do not exaggerate!
ID: 65989 · Report as offensive
KAMasud

Send message
Joined: 6 Oct 06
Posts: 204
Credit: 7,608,986
RAC: 0
Message 65990 - Posted: 25 Aug 2022, 7:21:26 UTC - in response to Message 65989.  

If I told you once, I told you a million times: do not exaggerate!


LOL. I think he is correct though. I have to sit and manually download WCG WU"s, otherwise it will be "project back off" for days.
ID: 65990 · Report as offensive
Richard Haselgrove

Send message
Joined: 1 Jan 07
Posts: 925
Credit: 34,100,818
RAC: 11,270
Message 65991 - Posted: 25 Aug 2022, 8:50:06 UTC - in response to Message 65990.  

That's for another forum. I've posted about it at WCG.
ID: 65991 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1045
Credit: 16,506,818
RAC: 5,001
Message 65992 - Posted: 25 Aug 2022, 14:52:02 UTC - in response to Message 65990.  

I think he is correct though. I have to sit and manually download WCG WU"s, otherwise it will be "project back off" for days.


Me too. Even if I baby-sit the connection and restart every time it backs off, it can still take many hours of frustration to download the files (about 20) for two ARP1 tasks. In three hours, I have downloaded only about half the needed files.

I do not think the problem is due to inadequate (486) cpu processing power, but one of hopeless networking, both hardware and possibly misconfigured software. Consider:

Thu 25 Aug 2022 10:29:29 AM EDT | World Community Grid | Started download of 5e71c2d062b7962bdb6bac1992fbeeea.
Thu 25 Aug 2022 10:29:33 AM EDT | World Community Grid | Temporarily failed download of 5e71c2d062b7962bdb6bac1992fbeeea.: transient HTTP error
Thu 25 Aug 2022 10:29:33 AM EDT | World Community Grid | Backing off 01:28:25 on download of 5e71c2d062b7962bdb6bac1992fbeeea.
Thu 25 Aug 2022 10:29:48 AM EDT | World Community Grid | Started download of 52e25b9ec8188295263a9b3cc85d65dc.7z
Thu 25 Aug 2022 10:29:51 AM EDT | World Community Grid | Temporarily failed download of 52e25b9ec8188295263a9b3cc85d65dc.7z: transient HTTP error
Thu 25 Aug 2022 10:29:51 AM EDT | World Community Grid | Backing off 01:12:58 on download of 52e25b9ec8188295263a9b3cc85d65dc.7z
Thu 25 Aug 2022 10:30:07 AM EDT | World Community Grid | Started download of 7fafa20f2f43dd6206726e20baabaa68.
Thu 25 Aug 2022 10:30:10 AM EDT | World Community Grid | Temporarily failed download of 7fafa20f2f43dd6206726e20baabaa68.: transient HTTP error
Thu 25 Aug 2022 10:30:10 AM EDT | World Community Grid | Backing off 01:13:24 on download of 7fafa20f2f43dd6206726e20baabaa68.
Thu 25 Aug 2022 10:30:20 AM EDT | World Community Grid | Started download of 44973705de8b4bb1c1113f8e96ecfbc8.
Thu 25 Aug 2022 10:30:24 AM EDT | World Community Grid | Temporarily failed download of 44973705de8b4bb1c1113f8e96ecfbc8.: transient HTTP error
Thu 25 Aug 2022 10:30:24 AM EDT | World Community Grid | Backing off 01:43:23 on download of 44973705de8b4bb1c1113f8e96ecfbc8.

ID: 65992 · Report as offensive
Glenn Carver

Send message
Joined: 29 Oct 17
Posts: 774
Credit: 13,433,329
RAC: 7,110
Message 65993 - Posted: 25 Aug 2022, 14:59:56 UTC - in response to Message 65988.  

WCG has loads to do, but apparently a 14 billion dollar company can't provide a server fast enough to cater for Boinc. I didn't think Canadians were quite that incompetant. Their server is so slow you can't even get onto their forum. I'm assuming it's a 486 in somebody's shed.
More likely a set of virtual machines or containers running on bare metal or cloud. It's not difficult to saturate a network connection.
ID: 65993 · Report as offensive
Richard Haselgrove

Send message
Joined: 1 Jan 07
Posts: 925
Credit: 34,100,818
RAC: 11,270
Message 65994 - Posted: 25 Aug 2022, 17:29:09 UTC

I can get through to their forums with no problems - just posted there again. My usual problem is dropped connections on short files - they can't handle the concurrency. But today it was a 102 MB shared file, that came at less than 350 KBps (on a 70 MB fibre line) - imagine that here.

Some projects have networking wizards, others don't.
ID: 65994 · Report as offensive
Glenn Carver

Send message
Joined: 29 Oct 17
Posts: 774
Credit: 13,433,329
RAC: 7,110
Message 65995 - Posted: 25 Aug 2022, 19:12:57 UTC - in response to Message 65994.  

I was surprised their WUs have so many individual (and often small) files to download, no wonder they have problems. Would be better if they were zipped into a smaller number of larger files surely.
ID: 65995 · Report as offensive
Previous · 1 . . . 86 · 87 · 88 · 89 · 90 · 91 · Next

Message boards : Number crunching : New work Discussion

©2024 climateprediction.net