Message boards : Number crunching : Progress percentage and trickles
Message board moderation
Author | Message |
---|---|
Send message Joined: 14 Feb 06 Posts: 31 Credit: 4,507,116 RAC: 2,013 |
I'd always assumed that as there were 20 trickles for each task, the progress percentage would "match" that and increase by 5% for each trickle. Note that I've never thought to check that, so it could be completely wrong. But on batches 1020, 1022, 1023 and 1027 I'm definitely not seeing this. For example, there is a batch 1020 task at 79.80% that has 19 trickles, and a batch 1022 task that is at 38.76% that has has 10 trickles. So are there more trickles on these batches? Or is there something else going on? Thanks! |
Send message Joined: 15 May 09 Posts: 4540 Credit: 19,039,635 RAC: 18,944 |
I'd always assumed that as there were 20 trickles for each task, the progress percentage would "match" that and increase by 5% for each trickle. Note that I've never thought to check that, so it could be completely wrong.The 20 trickles isn't a constant. This completed task from 1022 wah2_eas25_n39o_201912_24_1022_012316411_0 certainly had 20 trickles. I haven't paid attention to the timing of them but I would expect them to be evenly spaced throughout a task's progress. The _24_ in the task name tells you the number of model months. I am pretty sure that in the past some batches have had the trickle up messages aligned with the production of the zips which would have given this task 24 trickle up messages. Glenn would probably be able to give you a more definitive answer. |
Send message Joined: 1 Jan 07 Posts: 1061 Credit: 36,742,261 RAC: 6,307 |
If you look at your current WaH2 task by name - WaH2 tasks for host 1465867, you'll see names like: wah2_eas25_3043_208912_24_1027_012338530_0 I've marked the planned number of trickles and the batch number. Other task types will have different trickle counts when complete. |
Send message Joined: 1 Jan 07 Posts: 1061 Credit: 36,742,261 RAC: 6,307 |
20 is the maximum number of trickles displayed on the result page. If you look at a task still in progress, like task 22496690, you'll see that the oldest trickle (at the bottom), is at timestep 11,819. But if you look at a completed WaH2 task like task 22430652, the oldest trickle is at timestep 57,899 - four have dropped off the bottom of the displayed list. |
Send message Joined: 29 Oct 17 Posts: 1049 Credit: 16,575,067 RAC: 15,735 |
The number of trickles is the same for all tasks in a batch and the same for all batches in a project like EAS25. It can happen that network delays cause trickles to arrive late. For the EAS25 batches, the output files with the model results are going to the scientists in South Korea, whereas the trickles are going back to Oxford/CPDN. --- CPDN Visiting Scientist |
Send message Joined: 14 Feb 06 Posts: 31 Credit: 4,507,116 RAC: 2,013 |
Thanks for all your comments, especially those from Richard showing the "secret" coding of the number of trickles. 🙂 It's always nice to learn something. |
©2024 cpdn.org