Message boards :
Number crunching :
The dreaded "not requesting new work, don't need message"
Message board moderation
Author | Message |
---|---|
Nuadormrac Send message Joined: 7 Apr 00 Posts: 136 Credit: 1,703,351 RAC: 0 |
When exactly is the core client going to get a fix to this nonsense? Older versions never had this problem. Right now, all projects except SETI are set to no new work, due to the wow signal event starting up, and wanting to build up pending credits. The GPU is crunching, but the CPU is completely inactive with no CPU tasks running. Preferences are set to get them, but the core client just spit this up while I was away today, and stopped asking for CPU tasks. The CPU cores are idle... Others have reported these problems with other projects to. Is the BOINC core client going to get a fix to this sometime, so it stops, not requesting work when the computer sits idle? I'd rather they did away with the whole concept of debts and whatever else, then have the core client stop asking for work. I think I like the way older versions of BOINC handled this one here, but need to run with a 7.0.4+ version as one project needs an app_config file. |
Khali Send message Joined: 21 Mar 14 Posts: 12 Credit: 2,395,589 RAC: 0 |
I'm having the same problem with my GPU. I am down to one task and its not even trying to get any more. The Boinc Dev Team is too much in love with what they call the "learning" aspect of the Boinc client, and what I call unneeded fuzzy logic that serves no useful purpose to us the users. |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Try updating to Boinc 7.2.42, those were the sort of issues reported and fixed when 7.0.64 was first released, rather than being backported into the 7.0.x line, those fixes eventually arrived in a new version. BOINC client software downloads Claggy |
Khali Send message Joined: 21 Mar 14 Posts: 12 Credit: 2,395,589 RAC: 0 |
Try updating to Boinc 7.2.42, those were the sort of issues reported and fixed when 7.0.64 was first released, rather than being backported into the 7.0.x line, those fixes eventually arrived in a new version. I am already running 7.42.2 so that isn't going to help my problem. Speaking of which I found why my client isn't asking for more work. It thinks it takes my GTX 780 Ti 11:57:25 to run one task when in reality it averages out to some where around 30 minutes. So it thinks I already have enough work for my minimum work/maximum additional work settings. Looks like I am going to have to play around with my Seti settings to get around the Boinc clients major run time estimation flaw. |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Try updating to Boinc 7.2.42, those were the sort of issues reported and fixed when 7.0.64 was first released, rather than being backported into the 7.0.x line, those fixes eventually arrived in a new version. You're changed to anonymous platform today, your host hasn't done it's 11 validations for each app_version yet, once it does then the task estimates will match reality: Application details for host 7245297 AstroPulse v6 (anonymous platform, NVIDIA GPU) Number of tasks completed 0 Max tasks per day 33 Number of tasks today 5 Consecutive valid tasks 0 Average turnaround time 0.00 days For Astropulse_v6, to be counted, the tasks need to be under 10% Blanked, and haven't exited early with 30 found single pulses, or 30 repetitive pulses. Claggy |
Khali Send message Joined: 21 Mar 14 Posts: 12 Credit: 2,395,589 RAC: 0 |
Not quite sure what it means by anonymous platform but I will keep chugging along and hope it gets sorted out. My apologies if I sound like I went off on a rant about Boinc. All I can say is some of the things the client does just do not make sense to me. Instead of following the logical path of doing things it seems to go off on a tangent and do the oddest things that make no sense to me. Bit of a pet peeve of mine. For example, it likes to jump all over the place in my task buffer and not run the tasks in a logical way by running the ones with the shortest deadlines first. Thus we end up with it going into panic mode or just not getting tasks done and having to abort them. I don't like taking tasks then not getting them run because of what the Devs call Boinc's "learning mode" when you complain about it over on the Boinc forums. A few of them get rather condescending over it as well. |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Not quite sure what it means by anonymous platform but I will keep chugging along and hope it gets sorted out. You're either run the Lunatics installer, and that has put third party apps and an app_info.xml in your project directory, then it's restarted Boinc, Or you're put third party apps and an app_info.xml in your project directory and restarted Boinc, hence your running Anonymous platform apps, as opposed to the Stock apps. Claggy |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
For example, it likes to jump all over the place in my task buffer and not run the tasks in a logical way by running the ones with the shortest deadlines first. Thus we end up with it going into panic mode or just not getting tasks done and having to abort them. I don't like taking tasks then not getting them run because of what the Devs call Boinc's "learning mode" when you complain about it over on the Boinc forums. A few of them get rather condescending over it as well. Boinc runs each projects tasks in FIFO order (First In, First Out), unless it thinks a task can't make the deadline, then it jumps that task into High Priority, If Boinc were to do tasks in Earliest deadline Order, Projects with tasks with really long deadlines like CPDN wouldn't get any runtime because earlier deadlines would keep trumping them. Claggy |
Khali Send message Joined: 21 Mar 14 Posts: 12 Credit: 2,395,589 RAC: 0 |
Ah, but in my case I run the same application on my CPU and each GPU. I don't mix and match several different CPU projects at the same time. For example, I am running AP on my 780 Ti, the only exception is when the odd Poem GPU task pops up and those are rare lately. Nothing but GPU Grid long runs on my 680, and my CPU is running only Poem tasks, aside from what CPU resources my GPU's need. Yes there is some variation in the estimated run times but not enough to mater when my task buffer is set to a minimum of 0.1 and with a max additional of 0.2. |
betreger Send message Joined: 29 Jun 99 Posts: 11361 Credit: 29,581,041 RAC: 66 |
Ah, but in my case I run the same application on my CPU and each GPU. I don't mix and match several different CPU projects at the same time. If I understand it correctly, Claggy is asking where did you get the Seti apps from? |
Darth Beaver Send message Joined: 20 Aug 99 Posts: 6728 Credit: 21,443,075 RAC: 3 |
sorry to jump in here but Claggy my 2 machines are showing the wrong times for estimates and have done heaps of ap the AMD is not to far out only 6hrs but the other 1 says 214hrs each time it downloads a AP but takes only 13hrs and will only download 3 units as that's how many cores i have going i'm wondering why the estimate time hasn't dropped yet how many units roughly will it take before it start to correct itself ? |
Khali Send message Joined: 21 Mar 14 Posts: 12 Credit: 2,395,589 RAC: 0 |
Ah, but in my case I run the same application on my CPU and each GPU. I don't mix and match several different CPU projects at the same time. Yes he did and I missed his post when I scrolled down and just saw the post after that one. My mistake. I did indeed run the Lunatic installer and did not make the connection that it might be why I'm having problems/issues since I have had similar issues on other Boinc projects where I had not installed any new software at all. I'm sorry if I am coming across argumentative or rude. That is not my intention at all. Lets just say I am dealing with a lot of stress due to my Father, who has Alzheimer's, being in the hospital after a bad fall late last night. Thank you every one for trying to help out. |
Donald L. Johnson Send message Joined: 5 Aug 02 Posts: 8240 Credit: 14,654,533 RAC: 20 |
I'm sorry if I am coming across argumentative or rude. That is not my intention at all. Lets just say I am dealing with a lot of stress due to my Father, who has Alzheimer's, being in the hospital after a bad fall late last night. If you'd like to talk about that, with people who have been through similar difficulties, drop imto the Cafe' Seti and visit the PTSD and PTSD 2 threads. Donald Infernal Optimist / Submariner, retired |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
sorry to jump in here but Claggy my 2 machines are showing the wrong times for estimates and have done heaps of ap the AMD is not to far out only 6hrs but the other 1 says 214hrs each time it downloads a AP but takes only 13hrs and will only download 3 units as that's how many cores i have going i'm wondering why the estimate time hasn't dropped yet how many units roughly will it take before it start to correct itself ? Two of your hosts haven't completed their 11 validations of non-outlier tasks yet, ie AP tasks with less than 10% Blanking and don't exit early with overflows, and for MB v7 tasks that don't exit early: (The 'Number of tasks completed' is the important one, it only increments on validation of non-outlier tasks) Application details for host 7326757 AstroPulse v6 (anonymous platform, CPU) Number of tasks completed 10 Max tasks per day 46 Number of tasks today 5 Consecutive valid tasks 13 Average processing rate 38.38 GFLOPS Average turnaround time 0.94 days Application details for host 7336776 SETI@home v7 (anonymous platform, CPU) Number of tasks completed 10 Max tasks per day 44 Number of tasks today 0 Consecutive valid tasks 11 Average processing rate 11.37 GFLOPS Average turnaround time 0.34 days AstroPulse v6 (anonymous platform, CPU) Number of tasks completed 2 Max tasks per day 35 Number of tasks today 2 Consecutive valid tasks 2 Average processing rate 17.20 GFLOPS Average turnaround time 1.36 days Claggy |
Darth Beaver Send message Joined: 20 Aug 99 Posts: 6728 Credit: 21,443,075 RAC: 3 |
Thanks Claggy i'm not to worry'd about the Lappy host 7336776 as it's only just started and i got a problem that i can't fix (hardware) it was the other the q8350 host 7326757 but now you explained it ok i just have to patent perhaps another day and it should start to go down from 214 hrs seeing as it takes aprox 13 hrs |
qbit Send message Joined: 19 Sep 04 Posts: 630 Credit: 6,868,528 RAC: 0 |
I have the same problem here. This morning I set the buffer to 9 days and when I just came home from work I found my lappy crunching the last 4 available tasks, all 4 APs, which means that I didn't get any tasks over the day. Now I set the buffer to 10 days+9.25 days additionally and I got 3 more APs, LOL. And yes, all my AP tasks show 300+ hours remaining. |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Don´t worry after 11 sucessfuly validated Wu your numbers will down to the real one, but that afect the WU DL after that the old ones remains unalterated. |
rob smith Send message Joined: 7 Mar 03 Posts: 22202 Credit: 416,307,556 RAC: 380 |
10+9.25 means you will not get any new work for several days, then you will grab a big chunk (up to 10 days worth, or the current processor limits, which ever is the less), then wait another 9 days and fill the tanks again. It is far better to have a smaller cache, say 5 days, and an even smaller "extra", something like 0.05 or less works pretty well, keeping your cache at the 5 days level (or the processor limits) Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Darth Beaver Send message Joined: 20 Aug 99 Posts: 6728 Credit: 21,443,075 RAC: 3 |
thanks Rob i did what Frankie did and up'd the extra buffer to 9 days and same thing no cpu units and only 2 gpu 1's lololol better change it back to 4 days hehehe |
Nuadormrac Send message Joined: 7 Apr 00 Posts: 136 Credit: 1,703,351 RAC: 0 |
Try updating to Boinc 7.2.42, those were the sort of issues reported and fixed when 7.0.64 was first released, rather than being backported into the 7.0.x line, those fixes eventually arrived in a new version. Actually, the reason I'm running the older version, was precisely to avoid that problem with bitcoin utopia, where one needs the app_config to restrict it to using just one core, to match up with one's ASIC which is attached to the computer. If I had to chose where I'd want it to not get work; it's clear where is payed better :d But still rather annoying in a challenge, or what will be a challenge in a couple weeks. As to what someone said about the thing learning, yeah there is such a thing as software trying to get too smart for it's own good, then imposing this on the user. I think many will agree that "Mr. Clippy", in Microsoft Office would be a perfect example of that. Nothing like a dancing paperclip, wanting to "auto-correct" one's spelling of proper names and the like. Speaking of which, I now have a newer cell phone that has that annoying little "feature". If I'm typing a text, post, or whatever, and go to type an acronym, it goes to "auto-correct" it to what it thinks the word should be. Problem, it's not a word, it's an acronym, and even in all caps, the stupid software insists, and even back spacing and retyping what it tried to correct, it won't give up ad nauseam... I could say where some who are more prone to act out their frustration then myself might be inclined to throw their phone at that point /rofl |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.