Hi!

Just wondering what could be the issue that I don't get any GPU task on my Kubuntu system. I've tested that other OpenCL program, like luxmark, they run fine. And all troubleshooting I've done indicates that Rocm is properly installed. Any help would be welcome.
Here's the log's top:

Thu 18 Jul 2019 11:51:12 AM EEST |  | Starting BOINC client version 7.14.2 for x86_64-pc-linux-gnu
Thu 18 Jul 2019 11:51:12 AM EEST |  | log flags: file_xfer, sched_ops, task
Thu 18 Jul 2019 11:51:12 AM EEST |  | Libraries: libcurl/7.64.0 OpenSSL/1.1.1b zlib/1.2.11 libidn2/2.0.5 libpsl/0.20.2 (+libidn2/2.0.5) libssh/0.8.6/openssl/zlib nghttp2/1.36.0 librtmp/2.3
Thu 18 Jul 2019 11:51:12 AM EEST |  | Data directory: /var/lib/boinc-client
Thu 18 Jul 2019 11:51:12 AM EEST |  | OpenCL: AMD/ATI GPU 0: Vega 10 XT [Radeon RX Vega 64] (driver version 2924.0 (HSA1.1,LC), device version OpenCL 2.0, 8176MB, 8176MB available, 11397 GFLOPS peak)
Thu 18 Jul 2019 11:51:13 AM EEST |  | [libc detection] gathered: 2.29, Ubuntu GLIBC 2.29-0ubuntu2
Thu 18 Jul 2019 11:51:13 AM EEST |  | Host name: TemeRyzen
Thu 18 Jul 2019 11:51:13 AM EEST |  | Processor: 16 AuthenticAMD AMD Ryzen 7 1700 Eight-Core Processor [Family 23 Model 1 Stepping 1]
Thu 18 Jul 2019 11:51:13 AM EEST |  | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb hw_pstate sme ssbd sev ibpb vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 xsaves clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif overflow_recov succor smca
Thu 18 Jul 2019 11:51:13 AM EEST |  | OS: Linux Ubuntu: Ubuntu 19.04 [5.0.0-20-generic|libc 2.29 (Ubuntu GLIBC 2.29-0ubuntu2)]
Thu 18 Jul 2019 11:51:13 AM EEST |  | Memory: 15.65 GB physical, 15.23 GB virtual
Thu 18 Jul 2019 11:51:13 AM EEST |  | Disk: 384.59 GB total, 382.59 GB free

---

Here's the log's top:

The following messages of WUs starting, scheduler request for work & response would be of use also.
It did get an AP WU on the 14th UTC.

I'd suggest making sure that "Use GPU based on preferences" is selected in BOINC Manager.

---

Grant
Darwin NT

Here's the log's top:

The following messages of WUs starting, scheduler request for work & response would be of use also.
It did get an AP WU on the 14th UTC.

I'd suggest making sure that "Use GPU based on preferences" is selected in BOINC Manager.

It asked WU’s for the gpu. Currently at work and can’t access the system remotely to double check it.

---

It asked WU’s for the gpu. Currently at work and can’t access the system remotely to double check it.

No sign of it asking for any more work- CPU or GPU, it's finished what it had & that's it.
If you have No New Tasks set, then that's what will happen.

---

Grant
Darwin NT

It asked WU’s for the gpu. Currently at work and can’t access the system remotely to double check it.

No sign of it asking for any more work- CPU or GPU, it's finished what it had & that's it.
If you have No New Tasks set, then that's what will happen.

Well I get this from time to time :

Fri 19 Jul 2019 11:13:05 AM EEST | SETI@home | Requesting new tasks for CPU and AMD/ATI GPU
Fri 19 Jul 2019 11:13:08 AM EEST | SETI@home | Scheduler request completed: got 100 new tasks

And no GPU WU's only CPU WU's. :/
clinfo -l gives this :

Platform #0: AMD Accelerated Parallel Processing
 `-- Device #0: gfx900

When looking around in the directories where the apps are I only see Astropulse app for AMD OpenCL and no app for Seti@Home for AMD OpenCL so that might cause issues. [/quote]

---

Well I get this from time to time :

The system has picked up work since I last saw it.

I suggest you check all your Web account & Manager settings for that system- it won't request work for the GPU if it isn't allowed to, and if it doesn't get any work for MB, then it won't download the application to process it.
Check the settings, web & Manager (better yet make sure the Manager is set to use the web settings). Update the Manager to get the web settings. If it still doesn't attempt to get any MB GPU work, then exit BOINC, restart, and post the first 40 or so lines of the log, making sure to include work requests & responses.

---

Grant
Darwin NT

Well after some time this is what I see in the settings.



So with the Rocm driver installed the gpu task button is grey...

---

That set of settings says nothing about the use of a GPU, just your CPU.
A couple of things that may help:
First you cache settings will not ensure you have a constant flow of work. The first is the size of the cache in days, 0.5 is a bit small, most folks settle on 2 or 3 as being a reasonable size. The second effectively says how often BOINC should make sure the cache is filled up, so you want this to be a smaller figure, and a small figure, say 0.1 will work better.

Next your CPU use figures - if you are aiming at using a GPU it is recommended that you set the "use % of CPUs" to between 75 and 90 - this way at least one CPU core will be available for other things, including supporting the GPU, at all times. At the same time set the "% time" back up to 100. The combination of these two actions will reduce the amount of work being done on the CPU, and reduce the task swapping that takes place when you limit the time - it's all to do with the way BOINC interacts with the operating system that conspires to a rather course control and inefficient and infective thermal management.

There are more options available from the "advanced" view, these include turning the use of GPUs on and off (which I thought was the default situation, but...)

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Since you appear to be running the "stock" applications, and haven't yet managed to get any "GPU" tasks you won't have an application - they are only delivered on demand.
Have a look at your web settings (*) just in case the use AMD/ATI flag isn't set - I think setting those flags work on the server to enable/disable the sending of work to specific processors.
(*) - https://setiathome.berkeley.edu/prefs.php?subset=project

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

You might want to send a P.M. to RueiKe and ask him for help, but to me your problem looks to be the driver that you're using not being SETI compatible.

[edit] The reason behind my thinking is that Rueike's list the driver as "OpenCL: 1.2" where as yours shows as "1 () driver: 2.00".

Cheers.

So with the Rocm driver installed the gpu task button is grey...

FYI:

I'm using the ROCm driver fine but this is on Gentoo.

The only 'problem' I had was with one update that didn't also update the libclc library to a later needed version. A good test is to see if the clinfo utility returns a full response for all the compute capabilities.

You may need to reinstall/juggle Mesa, LLVM, Clang, and libclc to get the right versions coordinated.

Check on the forums for your distro??


Happy fast crunchin'!
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)