Generic Multiple GPU discusssion

Tom M
Tom M
Joined: 2 Feb 06
Posts: 5586
Credit: 7673852828
RAC: 1779832

Keith Myers wrote:And I

Keith Myers wrote:

And I wonder why you were messing with the repository BOINC when you could have been running the team client.  That client gives you so much more with regard to the management of projects compared to any repository stock client, I have no clue as to why anyone with the option would want anything else.

The Team client doesn't allow you to combine Rx 5700's and Rx 580's. The Repository version and Windows 10 do.

==edit==

I am happy to report that I am now running 4 X Rx 5700's and a mix of World Community Grid with no "apparent" issues.  The other Rx 5700 is currently helping me experiment with my Intel motherboards.

==edit==

Tom M

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4700
Credit: 17544582799
RAC: 6406674

Well, you are going to have

Well, you are going to have to explain to me how the BOINC client has ANYTHING do with letting ANY hardware run.

That is is the job of the hardware drivers SOLELY.  Only the AMD video drivers have anything to do with reporting the capabilities of the hardware to the client. 

The client is just a receiver of information. It has nothing to do with determining the hardware's capabilities.

If the the drivers report incorrect information to the client, it isn't the clients fault, it is the drivers.

 

Ian&Steve C.
Ian&Steve C.
Joined: 19 Jan 20
Posts: 3681
Credit: 33828306909
RAC: 37780777

Keith Myers wrote:Well, you

Keith Myers wrote:

Well, you are going to have to explain to me how the BOINC client has ANYTHING do with letting ANY hardware run.

That is is the job of the hardware drivers SOLELY.  Only the AMD video drivers have anything to do with reporting the capabilities of the hardware to the client. 

The client is just a receiver of information. It has nothing to do with determining the hardware's capabilities.

If the the drivers report incorrect information to the client, it isn't the clients fault, it is the drivers.

 

the only thing I can think is that *something* in the code of the team client is malforming his coproc_info.xml file with regard to his AMD cards. most of us run Nvidia cards and maybe havent run into this issue. and as you know, a lot of the edits made to the client have to do with this file, for... reasons ;)

you're right that the hardware drivers are responsible for GPU operation, but BOINC needs a properly formed coproc_info.xml file to be able to use them.

 

Tom, please post the contents of the coproc_info.xml file from both the Team client and from the standard boinc client with any mix of RX5700 and RX 570 installed. I'm thinking you'll see the issue there.

_________________________________________________________________________

Tom M
Tom M
Joined: 2 Feb 06
Posts: 5586
Credit: 7673852828
RAC: 1779832

Ian&Steve C. wrote: you're

Ian&Steve C. wrote:

you're right that the hardware drivers are responsible for GPU operation, but BOINC needs a properly formed coproc_info.xml file to be able to use them.

Tom, please post the contents of the coproc_info.xml file from both the Team client and from the standard BOINC client with any mix of RX5700 and RX 570 installed. I'm thinking you'll see the issue there.

Let make sure I am understanding what exactly you are requesting before I start shuffling OS's and Hard disks around.

You want two setups with mixed Rx 5700 and Rx 460/570/580 GPUs under Linux 20 (Kernel 5.3)

1) The Tbar All-in-One/Team Boinc Manager install under Ubuntu.

2) The Repository Boing Manager install under Ubuntu.

You want the coproc_info.xml files from both of those setups.  And I will need to confirm that AIO setup is not recognizing both types of Rx gpus?  If I can't confirm my observed behavior then I have a moot question.

Would it be better to move this to the team area discussion for E@H or leave it out here?  Since we are dealing with a "private label" version of Boinc Manager?

This will take a little time to sort through and make sure I have clean Ubuntu 20 w/5.4 Kernels and drivers installed with both versions.

Tom M

 

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4700
Credit: 17544582799
RAC: 6406674

Tom, TBar's AIO is NOT

Tom, TBar's AIO is NOT Pandora.  We want you to install Pandora.

 

Ian&Steve C.
Ian&Steve C.
Joined: 19 Jan 20
Posts: 3681
Credit: 33828306909
RAC: 37780777

Keith Myers wrote:Tom,

Tom, I just want to see the coproc_info.xml from whatever setup "doesn't work" whether that be from TBar's AIO or from the pandora client. and without you changing the hardware or drivers, I want a coproc_info.xml from a setup that "does work" so I can compare them and see if there's a difference. i want the ONLY variable between the setups to be which version of BOINC you are running.

 

Keith Myers wrote:

Tom, TBar's AIO is NOT Pandora.  We want you to install Pandora.

 

yes. He can start with the normal AIO package, then replace the boinc executable with our Pandora version.

 

but since Tom seems confused about the what the team client is, maybe his issue is with the default AIO executable, and not with the Pandora version.

 

but all of this seems like exercises in futility. Tom has tried probably tons of configs and setups over a period of months and has yet to find a "solution" to the problem he himself is creating, probably partially due to poor bookkeeping and he's repeating bad configs without realizing it. the need to have both RX5700's and 570's in the same system seems asinine to me. why? if they don't play nice, just put them in separate systems and move on. the amount of downtime you're introducing and lost production from constantly doing this... is the juice worth the squeeze?

_________________________________________________________________________

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4700
Credit: 17544582799
RAC: 6406674

Quote:... is the juice worth

Quote:
... is the juice worth the squeeze?

My sentiment also.

But the modified coproc_info.xml is created by the Pandora client.  TBar's AIO would create a bone stock coproc_info.xml as would any repository version of BOINC.

I agree, I think there is some confusion as to what he is running in his attempts to run both gpu families.

 

Ian&Steve C.
Ian&Steve C.
Joined: 19 Jan 20
Posts: 3681
Credit: 33828306909
RAC: 37780777

Keith Myers wrote:But the

Keith Myers wrote:

But the modified coproc_info.xml is created by the Pandora client.  TBar's AIO would create a bone stock coproc_info.xml as would any repository version of BOINC

I agree it should, but it's worth verifying, could be an edge case that just hasnt been seen since no one else is trying to run this configuration. All of his Ubuntu systems in his profile list 7.16.6, and I don't recall if that's the version in the most recent AIO package or not (i think it's 7.16.5?). none of his clients show 7.17.0 either which the pandora client would show. unless he just overwrote them already with the repo install.

_________________________________________________________________________

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4700
Credit: 17544582799
RAC: 6406674

TBar's AIO stopped at

TBar's AIO stopped at 7.16.5.  The 7.16.6 is the last and latest release to get Android working again on BOINC properly. So he is running the official BOINC repo version if he reports 7.16.6.

I also think a lot of the distros have upgraded to that version also.

 

Tom M
Tom M
Joined: 2 Feb 06
Posts: 5586
Credit: 7673852828
RAC: 1779832

Lets review the biding

Lets review the biding here. 

Part I

I claim that when I install Ubuntu 20 with a 5.4 Kernel and the .45 gpu drivers I can "see" the cards via the command line.

tom@Intel-Testbed-02-AIO:~$ lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation UHD Graphics 630 (Desktop 9 Series) (rev 02)
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] (rev e7)
02:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Baffin [Radeon RX 460/560D / Pro 450/455/460/555/555X/560/560X] (rev e5)
03:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] (rev e7)
06:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Navi 10 [Radeon RX 5600 OEM/5600 XT / 5700/5700 XT] (rev c4)
tom@Intel-Testbed-02-AIO:~$

When I start up a copy of AIO (NOT pandora)

I claim I get the following log

Fri 26 Mar 2021 05:58:33 PM CDT |  | Starting BOINC client version 7.16.5 for x86_64-pc-linux-gnu
Fri 26 Mar 2021 05:58:33 PM CDT |  | log flags: file_xfer, sched_ops, task, sched_op_debug
Fri 26 Mar 2021 05:58:33 PM CDT |  | Libraries: libcurl/7.68.0 GnuTLS/3.6.13 zlib/1.2.11 brotli/1.0.7 libidn2/2.2.0 libpsl/0.21.0 (+libidn2/2.2.0) libssh/0.9.3/openssl/zlib nghttp2/1.40.0 librtmp/2.3
Fri 26 Mar 2021 05:58:33 PM CDT |  | Data directory: /home/tom/Desktop/BOINC
Fri 26 Mar 2021 05:58:39 PM CDT |  | OpenCL: AMD/ATI GPU 0: Radeon RX 580 Series (driver version 3188.4, device version OpenCL 1.2 AMD-APP (3188.4), 8178MB, 8178MB available, 5299 GFLOPS peak)
Fri 26 Mar 2021 05:58:39 PM CDT |  | OpenCL: AMD/ATI GPU 1: Radeon RX 560 Series (driver version 3188.4, device version OpenCL 1.2 AMD-APP (3188.4), 4082MB, 4082MB available, 2107 GFLOPS peak)
Fri 26 Mar 2021 05:58:39 PM CDT |  | OpenCL: AMD/ATI GPU 2: Radeon RX 580 Series (driver version 3188.4, device version OpenCL 1.2 AMD-APP (3188.4), 8178MB, 8178MB available, 5299 GFLOPS peak)
Fri 26 Mar 2021 05:58:39 PM CDT |  | Creating new client state file
Fri 26 Mar 2021 05:58:39 PM CDT |  | libc: Ubuntu GLIBC 2.31-0ubuntu9 version 2.31
Fri 26 Mar 2021 05:58:39 PM CDT |  | Host name: Intel-Testbed-02-AIO
Fri 26 Mar 2021 05:58:39 PM CDT |  | Processor: 16 GenuineIntel Intel(R) Core(TM) i9-9900 CPU @ 3.10GHz [Family 6 Model 158 Stepping 13]
Fri 26 Mar 2021 05:58:39 PM CDT |  | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp md_clear flush_l1d arch_capabilities
Fri 26 Mar 2021 05:58:39 PM CDT |  | OS: Linux Ubuntu: Ubuntu 20.04.1 LTS [5.4.0-42-generic|libc 2.31 (Ubuntu GLIBC 2.31-0ubuntu9)]
Fri 26 Mar 2021 05:58:39 PM CDT |  | Memory: 15.50 GB physical, 2.00 GB virtual
Fri 26 Mar 2021 05:58:39 PM CDT |  | Disk: 456.95 GB total, 423.08 GB free
Fri 26 Mar 2021 05:58:39 PM CDT |  | Local time is UTC -5 hours
Fri 26 Mar 2021 05:58:39 PM CDT |  | Config: use all coprocessors
Fri 26 Mar 2021 05:58:40 PM CDT |  | No general preferences found - using defaults
Fri 26 Mar 2021 05:58:40 PM CDT |  | Preferences:
Fri 26 Mar 2021 05:58:40 PM CDT |  | max memory usage when active: 7936.96 MB
Fri 26 Mar 2021 05:58:40 PM CDT |  | max memory usage when idle: 14286.52 MB
Fri 26 Mar 2021 05:58:40 PM CDT |  | max disk usage: 411.26 GB
Fri 26 Mar 2021 05:58:40 PM CDT |  | don't use GPU while active
Fri 26 Mar 2021 05:58:40 PM CDT |  | suspend work if non-BOINC CPU load exceeds 25%
Fri 26 Mar 2021 05:58:40 PM CDT |  | (to change preferences, visit a project web site or select Preferences in the Manager)
Fri 26 Mar 2021 05:58:40 PM CDT |  | Setting up project and slot directories
Fri 26 Mar 2021 05:58:40 PM CDT |  | Checking active tasks
Fri 26 Mar 2021 05:58:40 PM CDT |  | Setting up GUI RPC socket
Fri 26 Mar 2021 05:58:40 PM CDT |  | Checking presence of 0 project files
Fri 26 Mar 2021 05:58:40 PM CDT |  | This computer is not attached to any projects


So where did the Rx 5700 that Ubuntu see's go?

I will now shut down this system.  Unplug all the external gpus except the Rx 5700.

==========additional events to be added by editing===================

 

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.