Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 
World Community Grid Forums
  Category: Active Research
  Forum: OpenPandemics - COVID-19 Project
  Thread: Work unit availability

Quick Go ยป
No member browsing this thread
Thread Status: Active
Total posts in this thread: 822
Posts: 822   Pages: 83   [ Previous Page | 9 10 11 12 13 14 15 16 17 18 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 860713 times and has 821 replies Next Thread
Andrew80431
Cruncher
Joined: Nov 25, 2005
Post Count: 36
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
A combination of low wu production, more folks interested now that it's not beta, and the script kiddies trying to suck up all that they can get. No way to stop the script kiddies as far as I know.
The kitties are not getting any here either.


I set the project to look for new WUs at an interval of approximately 5 mins.
So far I have had a few WUs every day. GPU Usage on the Nvidia under WinX is around 16%, while the slower Intel GPU under Linux is working 24% of the time.

----------------------------------------
[Apr 10, 2021 7:04:25 AM]   Link   Report threatening or abusive post: please login first  Go to top 
biscotto
Cruncher
Italy
Joined: Apr 11, 2020
Post Count: 27
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
Hello, I apparently keep receiving GPU tasks on my linux platform that are not being executed somehow. Boinc correctly finds my gpu (both with mesa clover and with amd opencl):
OpenCL: AMD/ATI GPU 0 (ignored by config): Radeon RX 560 Series (POLARIS11, DRM 3.40.0, 5.10.0-5-amd64, LLV (driver version 20.3.4, device version OpenCL 1.1 Mesa 20.3.4, 4096MB, 4096MB available, 1567 GFLOPS peak)
OpenCL: AMD/ATI GPU 1: Radeon RX 560 Series (driver version 3188.4, device version OpenCL 1.2 AMD-APP (3188.4), 4065MB, 4065MB available, 2507 GFLOPS peak)
But then reports:
[work_fetch] --- state for AMD/ATI GPU ---
[work_fetch] shortfall 17280.00 nidle 1.00 saturated 0.00 busy 0.00
[work_fetch] share 1.000
choose_project: scanning
can fetch CPU
can fetch AMD/ATI GPU
AMD/ATI GPU needs work - buffer low
[work_fetch] set_request() for AMD/ATI GPU: ninst 1 nused_total 0.00 nidle_now 1.00 fetch share 1.00 req_inst 1.00 req_secs 17280.00
AMD/ATI GPU set_request: 17280.000000
Requesting new tasks for CPU and AMD/ATI GPU
[sched_op] CPU work request: 683.75 seconds; 0.00 devices
[sched_op] AMD/ATI GPU work request: 17280.00 seconds; 1.00 devices
Scheduler request completed: got 1 new tasks
[sched_op] estimated total CPU task duration: 36631 seconds
[sched_op] estimated total AMD/ATI GPU task duration: 0 seconds
and project is backed off by scheduler. I have switched on all the gpu toggling options and enabled computation while computer is in use, but gpu never works. Am i doing something wrong? Should i stop altogether asking tasks? Thanks in advance
----------------------------------------
Papa Ryzen 5 3600 / Mama Radeon RX 560

----------------------------------------
[Edit 2 times, last edit by Biscotto at Apr 10, 2021 8:04:10 AM]
[Apr 10, 2021 7:59:27 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Andrew80431
Cruncher
Joined: Nov 25, 2005
Post Count: 36
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
Hello, I apparently keep receiving GPU tasks on my linux platform that are not being executed somehow. Boinc correctly finds my gpu (both with mesa clover and with amd opencl):
OpenCL: AMD/ATI GPU 0 (ignored by config): Radeon RX 560 Series (POLARIS11, DRM 3.40.0, 5.10.0-5-amd64, LLV (driver version 20.3.4, device version OpenCL 1.1 Mesa 20.3.4, 4096MB, 4096MB available, 1567 GFLOPS peak)
OpenCL: AMD/ATI GPU 1: Radeon RX 560 Series (driver version 3188.4, device version OpenCL 1.2 AMD-APP (3188.4), 4065MB, 4065MB available, 2507 GFLOPS peak)

The minimum requirements for the project is OpenCL 1.2. Your Mesa driver only supports 1.1. https://www.x.org/wiki/RadeonFeature/ states that OpenCL is supported via ROCm for recent AMD GPUS. I am not sure how well ROCm is supported by BOINC.
Maybe some one else here has found a guide on how to get the AMD GPUs working.
I have an old Radeon HD 6950 I'd like to get into service for WCG, but was also unable to get it working under Ubuntu 18.04.
----------------------------------------

----------------------------------------
[Edit 2 times, last edit by Andrew80431 at Apr 10, 2021 10:18:59 AM]
[Apr 10, 2021 10:00:12 AM]   Link   Report threatening or abusive post: please login first  Go to top 
biscotto
Cruncher
Italy
Joined: Apr 11, 2020
Post Count: 27
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
The minimum requirements for the project is OpenCL 1.2. Your Mesa driver only supports 1.1.
I know, this is why in also installed amd's opencl driver that has 1.2. If anyone experiences the same problem let me know, we might start a new thread to get amd gpus working.
----------------------------------------
Papa Ryzen 5 3600 / Mama Radeon RX 560
[Apr 10, 2021 11:40:20 AM]   Link   Report threatening or abusive post: please login first  Go to top 
TLD
Veteran Cruncher
USA
Joined: Jul 22, 2005
Post Count: 805
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
I have a ATI Radeon 5770 that is doing GPU WUs right now. While this card is old and slow it is opencl 1.2 capable and recognized by Boinc as so.
----------------------------------------
[Apr 10, 2021 11:52:06 AM]   Link   Report threatening or abusive post: please login first  Go to top 
squid
Advanced Cruncher
Germany
Joined: May 15, 2020
Post Count: 56
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
There is no work for NVIDIA GPUs.

10-Apr-2021 14:34:31 [World Community Grid] Sending scheduler request: To fetch work.
10-Apr-2021 14:34:31 [World Community Grid] Requesting new tasks for NVIDIA GPU
10-Apr-2021 14:34:32 [World Community Grid] Scheduler request completed: got 0 new tasks
10-Apr-2021 14:34:32 [World Community Grid] No tasks sent
10-Apr-2021 14:34:32 [World Community Grid] No tasks are available for OpenPandemics - COVID 19
10-Apr-2021 14:34:32 [World Community Grid] No tasks are available for OpenPandemics - COVID-19 - GPU
10-Apr-2021 14:34:32 [World Community Grid] Tasks for CPU are available, but your preferences are set to not accept them
10-Apr-2021 14:34:32 [World Community Grid] Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them
10-Apr-2021 14:34:32 [World Community Grid] Tasks for Intel GPU are available, but your preferences are set to not accept them
[Apr 10, 2021 12:52:23 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Dayle Diamond
Senior Cruncher
Joined: Jan 31, 2013
Post Count: 452
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
Squid, resends demand the same brand as the original.
Other than that, there are no separate queues.
[Apr 10, 2021 1:34:26 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Andrew80431
Cruncher
Joined: Nov 25, 2005
Post Count: 36
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
I have a ATI Radeon 5770 that is doing GPU WUs right now. While this card is old and slow it is opencl 1.2 capable and recognized by Boinc as so.


Are you running this card under Linux or Windows?
If it's Linux I'm interested in how you got it working. We could discuss this in a separate thread, if you like.
----------------------------------------
[Apr 10, 2021 1:59:28 PM]   Link   Report threatening or abusive post: please login first  Go to top 
floyd
Cruncher
Joined: May 28, 2016
Post Count: 47
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
The minimum requirements for the project is OpenCL 1.2. Your Mesa driver only supports 1.1.
While Mesa is not fully OpenCL 1.2 compliant it is good enough for OPNG. You only need to get past the bouncer scheduler.
Maybe some one else here has found a guide on how to get the AMD GPUs working.
I have an old Radeon HD 6950 I'd like to get into service for WCG, but was also unable to get it working under Ubuntu 18.04.
Are you at the point where BOINC detects an OpenCL 1.1 device? Set some debug options:
CLOVER_PLATFORM_VERSION_OVERRIDE=1.2
CLOVER_DEVICE_VERSION_OVERRIDE=1.2
CLOVER_DEVICE_CLC_VERSION_OVERRIDE=1.2
[Apr 10, 2021 2:00:39 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Andrew80431
Cruncher
Joined: Nov 25, 2005
Post Count: 36
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Work unit availability
The minimum requirements for the project is OpenCL 1.2. Your Mesa driver only supports 1.1.
While Mesa is not fully OpenCL 1.2 compliant it is good enough for OPNG. You only need to get past the bouncer scheduler.
Maybe some one else here has found a guide on how to get the AMD GPUs working.
I have an old Radeon HD 6950 I'd like to get into service for WCG, but was also unable to get it working under Ubuntu 18.04.
Are you at the point where BOINC detects an OpenCL 1.1 device? Set some debug options:
CLOVER_PLATFORM_VERSION_OVERRIDE=1.2
CLOVER_DEVICE_VERSION_OVERRIDE=1.2
CLOVER_DEVICE_CLC_VERSION_OVERRIDE=1.2


So I added 
Environment="CLOVER_PLATFORM_VERSION_OVERRIDE=1.2"
Environment="CLOVER_DEVICE_VERSION_OVERRIDE=1.2"
Environment="CLOVER_DEVICE_CLC_VERSION_OVERRIDE=1.2"

to "/etc/systemd/system/boinc-client.service"
and did a 
systemctl daemon-reload
systemctl restart boinc-client

Now BOINC's event log shows:
OpenCL: AMD/ATI GPU 0: AMD CAYMAN (DRM 2.50.0 / 5.4.0-67-generic, LLVM 10.0.0) (driver version 20.0.8, device version OpenCL 1.2 Mesa 20.0.8, 2048MB, 2048MB available, 768 GFLOPS peak)


I'm curious if this will work... I'll keep you posted.
----------------------------------------
[Apr 10, 2021 2:19:08 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 822   Pages: 83   [ Previous Page | 9 10 11 12 13 14 15 16 17 18 | Next Page ]
[ Jump to Last Post ]
Post new Thread