Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 2
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 1878 times and has 1 reply Next Thread
NixChix
Veteran Cruncher
United States
Joined: Apr 29, 2007
Post Count: 1187
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
confused Please explain WU dependencies

I am a bit confused by the descriptions of the dependicies between the fortnights (cycle) and the areas of the simulation.
retsof:
AC@H was a bit different. Each area was processed in SERIES. ACH1_27_45 depended on the results from ACH1_26_45 ... depended on the results from ACH1_25_45 ... depended on the results from ACH1_24_45 ............. and everything depended on ACH1_1_45.

There were 82 area processing chains, and 10x redundancy for each area. Each area was something like 90000 hectares. The next fortnight could not get started for that area until the previous fortnight had a result for that area.

That description makes perfect sense. That is the vertical dependency.
Sekerob:
... The depencies are as what retsof describes. Before there was horizontal too i.e. cycle 14 would not start until 13 was complete for all 82 blocks, but that has effectively resulted in little project progress at all. We're now about 8 months in and still the 27x82x10 initial sets are not done.

Sekerob, I thought that you were implying that there was a change, that at one time an entire fortnight had to complete (horizontal dependency) before the next fortnight in the entire simulation could proceed.
NixChix:
Are you saying that the dependencies were modified to speed the project up? When did that occur?

All I am asking was did it change. If what Sekerob wrote is true; that there was a before; that there was a "horizontal dependency". It has been mentioned several times in postings.
retsof:
The dependencies were set by the scientists that wrote the program before the project even started.

<rant>resof, are you saying that there has been no change? That the dependencies have been the same from the beginning of the project. Perhaps there was never a full "horizontal dependency", because there have been postings about current WUs from different fortnights at roughly the same time. If there was a full "horizontal dependency" that could not be. All posters would be always refering to the same fortnight.
Well I don't want to ramble on. I just got suddenly curious about the dependencies after reading what seemed to be conflicting descriptions of the process.<rant/>
----------------------------------------

[Apr 5, 2008 8:56:51 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Please explain WU dependencies

Work is sent out in batches. So, in an ideal world, WCG would send out cycle 1, then a complete cycle 2, and so on. The vertical dependency means they can't send out cycle 1 and 2 at the same time.

As the project has continued, various work unit failures have caused the cycles to get mixed a little. To keep the project moving, incomplete cycles have been sent out.

This isn't ideal, and future projects are unlikely to be run the same way.
[Apr 5, 2008 9:08:58 PM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread