Posts by Keith Myers
|
1)
(Message 8945)
Posted 6 days ago by 
Keith Myers
Post: They don't have an abundance of already pre-formatted data in the pipeline. So they have to grab new sets of data and run them through the pre-formatter every time they run out of tasks. They don't work weekends. So it will be probably Tuesday at the earliest before we get new work. |
|
2)
(Message 8900)
Posted 17 days ago by
Keith Myers
Post: I see this even with running an update script every 2 minutes. I suspect the RTS buffer size is very small on the server and easily gets emptied with just one scheduler connection from a hungry host. And it then takes some time to refill it for the next host requesting work. So you just have to be patient and keeping asking for work. It will come down eventually. |
|
3)
(Message 8772)
Posted 6 Feb 2025 by
Keith Myers
Post: Server is out of disk space. No uploads happening. Feeder not running. |
|
4)
(Message 8760)
Posted 27 Jan 2025 by
Keith Myers
Post: I keep wondering if this issue of disk space is because they have just a single db that they never prune finished results to a separate science database that most other projects do. |
|
5)
(Message 8755)
Posted 18 Jan 2025 by
Keith Myers
Post: Good catch, I'm so used to seeing AMD as card definition that I forgot how ancient the client is with still labeling AMD cards as their predecessor ATI. |
|
6)
(Message 8753)
Posted 18 Jan 2025 by
Keith Myers
Post: I don't see anything wrong with your gpu app_version section. Rudimentary question. Is the app present in the asteroids directory? You normally have to run the stock client to download the app and tasks and get its download location installed into the client state file before converting everything over to an app_info installation. Common issues is that upon launch of an app_info installation is the client will delete an application it does not have a record of in the client state file. Best practice is to keep a copy of any executables outside of the Boinc project folders so you can replace them if they get deleted by the client somehow. So check for the presence of the gpu app first. Second, what kind of messages are present in the Event Log at startup. Does the client detect the gpu by finding the drivers for AMD? Should have an entry listing the gpu found and the OpenCL drivers loaded for it. |
|
7)
(Message 8751)
Posted 17 Jan 2025 by
Keith Myers
Post: You should read the first post in this thread https://www.boincstats.com/forum/1/12703 and temper your outrage. Show a little compassion. You've been told how to solve your problem, simply remove the BAM account manager and manage your accounts on your hosts yourself. There is little likelihood that BAM will exist in the future along with the BoincStats website. |
|
8)
(Message 8740)
Posted 5 Jan 2025 by
Keith Myers
Post: Not the same apps at Einstein compared to here. The apps here require minimum OpenCL version 2.1 AFAIK. |
|
9)
(Message 8737)
Posted 5 Jan 2025 by
Keith Myers
Post: This is the wrong place to report issues with Gridcoin. Projects have no control over how Gridcoin works. Post your issues at the Gridcoin Discord server. https://discord.gg/6Dyc8fJ4 |
|
10)
(Message 8734)
Posted 4 Jan 2025 by
Keith Myers
Post: Can't tell from your output which isn't much help. The first 30 lines of Boinc event log would tell me right away which OpenCL library you are using and version. Boinc does not like any version of OpenCL other than the one provided by the card manufacturers. So no Mesa or PoCL. I can't remember exactly what minimum version of OpenCL is necessary here. Your card may have too low capability. The mention about groups has me wondering in your OS did it ever add the user to the Boinc group. That is what is supposed to happen when installing Boinc. A groups output in the terminal would show that. |
|
11)
(Message 8731)
Posted 4 Jan 2025 by
Keith Myers
Post: Always check the server status before posting questions about lack of work. https://asteroidsathome.net/boinc/server_status.php Project has been out of work for over a day now. Nothing but resends available until the bucket gets refilled. |
|
12)
(Message 8714)
Posted 28 Dec 2024 by
Keith Myers
Post: I would assume so that here is the same case. Both projects use the highest SIMD function of a cpu to compute. |
|
13)
(Message 8707)
Posted 25 Dec 2024 by
Keith Myers
Post: There was a change in how long a finish file is allowed to be created and cleared back in 2019. Closed issue. I haven't seen any instance of that bug since then. |
|
14)
(Message 8696)
Posted 22 Dec 2024 by
Keith Myers
Post: Nothing nefarious going on here, just a quirk of the scheduler. Circumvent this issue simply by running an update script every 5-15 minute, your choice. |
|
15)
(Message 8688)
Posted 20 Dec 2024 by
Keith Myers
Post: Could be a permission issue. I don't know anything about AMD gpu drivers but it looks like access to or missing whatever is the AMD equivalent of a compiler like what nvcc does for Nvidia cards. |
|
16)
(Message 8673)
Posted 6 Dec 2024 by
Keith Myers
Post: Don't know. Never looked at Ralph. The issue is BoincStats and Willy's poor health. The website I'm pretty sure is running unmanaged now. Nobody around to fix anything. Hope that it will continue in some form after his demise. |
|
17)
(Message 8671)
Posted 6 Dec 2024 by
Keith Myers
Post: Should have been a BOINC wide Notice in the client about the project issues. Other projects manage to inform users of an impending planned outage or why servers aren't currently available. |
|
18)
(Message 8670)
Posted 6 Dec 2024 by
Keith Myers
Post: There is no problem with stats export here. The problem is with BoincStats which is broken. Stats have been available at FreeDC with no issues. |
|
19)
(Message 8663)
Posted 5 Dec 2024 by
Keith Myers
Post: +100 |
|
20)
(Message 8654)
Posted 5 Dec 2024 by
Keith Myers
Post: All my hundreds of uploads completed several days ago. But still waiting on the server to actually acknowledge those uploads, verify and credit them. |
Next 20