Re: [DNG] Update successful but video card issue
Hi Ludovic,hi Tim, actually, I do have the same issues. I am running a Thinkstation P520Cwith a Radeon Firepro W4100 with a 4K Benq PD3200U. I literally spent weeks of searching information about this problem to get rid of the flickering,but without any success. I tried several distributions (Linux Mint, Devuan and actually MXLinux) without any difference. And I also tried the proprietary drivers, which didn't change anything,installing it after the open driver and by installing it as first driver. I even tried various connections (HDMI, DP). The problem appears irregularely and more frequently whilest watching videos.There is a slight difference between watching 4k/30fps and 4k/60fps, wheras 4k/60fpsshows less flickering of the whole screen than 4k/30fps. And after having watched a 4k/60fps video for some seconds,the screen won't flicker at all for several minutes. My guess is a timing problem of the AMD Firepro W4100. Might be, using a AMD Radeon Pro WX would change that? Sincerely Roland Am 30 Jan 2020 04:07:58 +0100 (CET) Von "Ludovic Bellière" : Hello Tim, First you need to know the chipset name of your card in order to knowit's family. To do that you need to run (or similar): ```lspci | grep VGA``` It's output on my desktop shows: ```01:00.0 VGA compatible controller: Advanced Micro Devices, Inc.[AMD/ATI] Ellesmere [Radeon RX 470/480] (rev ef)``` This tells me that the chipset name of my card is Ellesmere, which isthe POLARIS architecture. A quick look to [RadeonFeature][1] tells methat POLARIS is part of the [Volcanic Islands family][2]. The [gentoowiki][3] might help too. Knowing the family is important, as it tellsyou how to tell the kernel to load the amdgpu driver before radeon. Because I suspect FirePro W4100 to be quite old, the [archlinuxhowtos][4] might be helpful to you. That being said, you might prefer installing the [proprietarydrivers][5] if you plan to use a professional card. I suspect takingthat road would come with its own set of problems. I hope this helps,Ludovic [1]: https://www.x.org/wiki/RadeonFeature/[2]:https://www.x.org/wiki/RadeonFeature/#decoderringforengineeringvsmarketingnames[3]: https://wiki.gentoo.org/wiki/AMDGPU#Feature_support[4]:https://wiki.archlinux.org/index.php/AMDGPU#Enable_Southern_Islands_.28SI.29_and_Sea_Islands_.28CIK.29_support[5]:https://www.amd.com/en/support/professional-graphics/firepro/firepro-wx100-series/firepro-w4100 On 29/01/20 23:05, Tim Wallace via Dng wrote:> I have been happily running ascii but upgraded from my Intel built-in> graphics to an AMD FirePro W4100 because I do a lot of 4K video editing,> but no game-playing, and thought this 50W card would save energy and> work well with 4K. There is a nearly identical Nvidia card, but I> decided to support AMD, for easier Linux support, which has not been my> experience!> > The card booted OK under ascii but was using the Radeon Xorg driver> which gave flashing behavior from time to time, typical of a bad video> driver. I decided to try beowulf to get some newer versions of stuff.> > I installed a generic ascii on my other partition, then did the upgrade,> from an xterm, perhaps not the smartest idea but it worked. When I ran> synaptic, it added and removed hundreds of packages! Not exactly> efficient, but I was left with a working system. The only issue was> that my Samsung M.2 SSD only boots as efi, and install systems don't> seem to realize that, resulting in boot failure or boot install failure,> but I used the rEFInd program to recover everything, and all seems good> with both beowulf and ascii.> > The situation now is that I can't get the correct X video driver> working, which should be the amdgpu one, under either ascii or beowulf.> (Full screen 4K video plays great at 60 fps under mpv, with nothing> dropped and no flashes.) I did try disabling pretty much all the> acceleration options on the radeon driver, but that did not help with> the flashing.> > Details:> Motherboard ASRock Z97 Extreme6, i7-4790K, 16 GB, latest Bios 2.80> > added non-free repository, and installed firmware. On boot, both radeon> and amdgpu kernel modules are installed. When X starts, the radeon> driver is used. The xorg.conf output by Xorg -configure calls for the> amdgpu driver, but trying to run, get fatal error:> > amdgpu_device_initialize: DRM version is 2.50.0 but this kernel is only> compatible with 3.x.x> > > Under ascii, the radeon driver is called for by Xorg -configure, and the> error when forcing amdgpu is almost the same, except> > ... DRM version is 2.49.0...> > It seems that there are some incompatibilities built into both ascii and> beowulf when it comes to this card. I could compile a kernel (which I> used to do all the time in the 90's, and probably haven't done for 10> years) if that would help with the DRM version issue.> > I tried the Ubuntu version of the proprietary driver from AMD, and it> complained about the kernel version. I could compile a kernel to match> the
Re: [DNG] Update successful but video card issue
Hello Tim, First you need to know the chipset name of your card in order to know it's family. To do that you need to run (or similar): ``` lspci | grep VGA ``` It's output on my desktop shows: ``` 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon RX 470/480] (rev ef) ``` This tells me that the chipset name of my card is Ellesmere, which is the POLARIS architecture. A quick look to [RadeonFeature][1] tells me that POLARIS is part of the [Volcanic Islands family][2]. The [gentoo wiki][3] might help too. Knowing the family is important, as it tells you how to tell the kernel to load the amdgpu driver before radeon. Because I suspect FirePro W4100 to be quite old, the [archlinux howtos][4] might be helpful to you. That being said, you might prefer installing the [proprietary drivers][5] if you plan to use a professional card. I suspect taking that road would come with its own set of problems. I hope this helps, Ludovic [1]: https://www.x.org/wiki/RadeonFeature/ [2]: https://www.x.org/wiki/RadeonFeature/#decoderringforengineeringvsmarketingnames [3]: https://wiki.gentoo.org/wiki/AMDGPU#Feature_support [4]: https://wiki.archlinux.org/index.php/AMDGPU#Enable_Southern_Islands_.28SI.29_and_Sea_Islands_.28CIK.29_support [5]: https://www.amd.com/en/support/professional-graphics/firepro/firepro-wx100-series/firepro-w4100 On 29/01/20 23:05, Tim Wallace via Dng wrote: > I have been happily running ascii but upgraded from my Intel built-in > graphics to an AMD FirePro W4100 because I do a lot of 4K video editing, > but no game-playing, and thought this 50W card would save energy and > work well with 4K. There is a nearly identical Nvidia card, but I > decided to support AMD, for easier Linux support, which has not been my > experience! > > The card booted OK under ascii but was using the Radeon Xorg driver > which gave flashing behavior from time to time, typical of a bad video > driver. I decided to try beowulf to get some newer versions of stuff. > > I installed a generic ascii on my other partition, then did the upgrade, > from an xterm, perhaps not the smartest idea but it worked. When I ran > synaptic, it added and removed hundreds of packages! Not exactly > efficient, but I was left with a working system. The only issue was > that my Samsung M.2 SSD only boots as efi, and install systems don't > seem to realize that, resulting in boot failure or boot install failure, > but I used the rEFInd program to recover everything, and all seems good > with both beowulf and ascii. > > The situation now is that I can't get the correct X video driver > working, which should be the amdgpu one, under either ascii or beowulf. > (Full screen 4K video plays great at 60 fps under mpv, with nothing > dropped and no flashes.) I did try disabling pretty much all the > acceleration options on the radeon driver, but that did not help with > the flashing. > > Details: > Motherboard ASRock Z97 Extreme6, i7-4790K, 16 GB, latest Bios 2.80 > > added non-free repository, and installed firmware. On boot, both radeon > and amdgpu kernel modules are installed. When X starts, the radeon > driver is used. The xorg.conf output by Xorg -configure calls for the > amdgpu driver, but trying to run, get fatal error: > > amdgpu_device_initialize: DRM version is 2.50.0 but this kernel is only > compatible with 3.x.x > > > Under ascii, the radeon driver is called for by Xorg -configure, and the > error when forcing amdgpu is almost the same, except > > ... DRM version is 2.49.0... > > It seems that there are some incompatibilities built into both ascii and > beowulf when it comes to this card. I could compile a kernel (which I > used to do all the time in the 90's, and probably haven't done for 10 > years) if that would help with the DRM version issue. > > I tried the Ubuntu version of the proprietary driver from AMD, and it > complained about the kernel version. I could compile a kernel to match > the Ubuntu one, I suppose, and re-try the proprietary driver, if anyone > thinks that would help. I'm not looking for the fastest performance in > the world, though, just stability! > > Any advice? > > --Tim > > > > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng > signature.asc Description: OpenPGP digital signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] download image ISO of beowulf
On Wed, Jan 29, 2020 at 09:20:34AM -0500, Ismael L. Donis Garcia wrote: > - Original Message - From: "Hendrik Boom" > To: > Sent: Tuesday, January 28, 2020 5:57 PM > Subject: Re: [DNG] download image ISO of beowulf > > > > On Tue, Jan 28, 2020 at 05:23:46PM -0500, Haines Brown wrote: > > > On Tue, Jan 28, 2020 at 10:13:25AM -0500, Ismael L. Donis Garcia wrote: > > > > Is there any image ISO of beowulf that can be downloaded from some > > > > site? > > > > > > > > Best Regards > > > > -- > > > > Ismael > > > > > > It seems most folks, like myself, got BEOWULF by upgraded from > > > ASCII. > > > > > > I gather official ISOs are made available only for stabe releases. > > > > > > There were problems using the mini.iso installer. > > > > I suspect beowulf is going to be stable soon. I think we should get > > started debugging the official ISO so it will be stable too. But first we'd need a draft of the official ISO. -- hendrik > > ___ > > Dng mailing list > > Dng@lists.dyne.org > > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng > > > > What I wanted is to upgrade to CHIMAERA from BEOWULF > In a virtual machine to test it, but starting from BEOWULF not from ASCII > > Thanks > -- > Ismael > > > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] Update successful but video card issue
I have been happily running ascii but upgraded from my Intel built-in graphics to an AMD FirePro W4100 because I do a lot of 4K video editing, but no game-playing, and thought this 50W card would save energy and work well with 4K. There is a nearly identical Nvidia card, but I decided to support AMD, for easier Linux support, which has not been my experience! The card booted OK under ascii but was using the Radeon Xorg driver which gave flashing behavior from time to time, typical of a bad video driver. I decided to try beowulf to get some newer versions of stuff. I installed a generic ascii on my other partition, then did the upgrade, from an xterm, perhaps not the smartest idea but it worked. When I ran synaptic, it added and removed hundreds of packages! Not exactly efficient, but I was left with a working system. The only issue was that my Samsung M.2 SSD only boots as efi, and install systems don't seem to realize that, resulting in boot failure or boot install failure, but I used the rEFInd program to recover everything, and all seems good with both beowulf and ascii. The situation now is that I can't get the correct X video driver working, which should be the amdgpu one, under either ascii or beowulf. (Full screen 4K video plays great at 60 fps under mpv, with nothing dropped and no flashes.) I did try disabling pretty much all the acceleration options on the radeon driver, but that did not help with the flashing. Details: Motherboard ASRock Z97 Extreme6, i7-4790K, 16 GB, latest Bios 2.80 added non-free repository, and installed firmware. On boot, both radeon and amdgpu kernel modules are installed. When X starts, the radeon driver is used. The xorg.conf output by Xorg -configure calls for the amdgpu driver, but trying to run, get fatal error: amdgpu_device_initialize: DRM version is 2.50.0 but this kernel is only compatible with 3.x.x Under ascii, the radeon driver is called for by Xorg -configure, and the error when forcing amdgpu is almost the same, except ... DRM version is 2.49.0... It seems that there are some incompatibilities built into both ascii and beowulf when it comes to this card. I could compile a kernel (which I used to do all the time in the 90's, and probably haven't done for 10 years) if that would help with the DRM version issue. I tried the Ubuntu version of the proprietary driver from AMD, and it complained about the kernel version. I could compile a kernel to match the Ubuntu one, I suppose, and re-try the proprietary driver, if anyone thinks that would help. I'm not looking for the fastest performance in the world, though, just stability! Any advice? --Tim ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] systemd breaking linuxcnc ...
Le 29/01/2020 à 11:58, Rainer Weikusat via Dng a écrit : Dimitris via Dng writes: On 1/28/20 6:46 PM, Rainer Weikusat via Dng wrote: run ntpdate from i've witnessed cases in the past where ntpd wasn't time-syncing correctly, and ntpdate saved the date/time. so, IMHO, there's nothing wrong with ntpdate. There's nothing wrong with ntpdate provided there's nothing wrong with having the wallclock make arbitrary backwards or forward jumps in unpredictable intervals, IOW, if the only user is a human who doesn't really care, everything's fine. In pre-systemd Debian and derivatives, the isssue was mitigated by running ntpdate only once, just before starting ntpd. Didier ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] systemd breaking linuxcnc ...
Anno domini 2020 Wed, 29 Jan 12:46:35 -0500 Steve Litt scripsit: > On Wed, 29 Jan 2020 10:58:57 + > Rainer Weikusat via Dng wrote: > > > Dimitris via Dng writes: > > > On 1/28/20 6:46 PM, Rainer Weikusat via Dng wrote: > > >> run ntpdate from > > > > > > i've witnessed cases in the past where ntpd wasn't time-syncing > > > correctly, and ntpdate saved the date/time. > > > so, IMHO, there's nothing wrong with ntpdate. > > > > There's nothing wrong with ntpdate provided there's nothing wrong with > > having the wallclock make arbitrary backwards or forward jumps in > > unpredictable intervals, IOW, if the only user is a human who doesn't > > really care, everything's fine. > > Don't all time clients have an option to choose between all at once > catch-up and slightly redefining a second so the catchup is more slow? > Maybe ntpdate and that systemd thing also have such an option? "ntpdate -B" should do it that way. > > SteveT > > Steve Litt > January 2020 featured book: Troubleshooting: Just the Facts > http://www.troubleshooters.com/tjust > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng > -- Please do not email me anything that you are not comfortable also sharing with the NSA, CIA ... ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] systemd breaking linuxcnc ...
On Wed, 29 Jan 2020 10:58:57 + Rainer Weikusat via Dng wrote: > Dimitris via Dng writes: > > On 1/28/20 6:46 PM, Rainer Weikusat via Dng wrote: > >> run ntpdate from > > > > i've witnessed cases in the past where ntpd wasn't time-syncing > > correctly, and ntpdate saved the date/time. > > so, IMHO, there's nothing wrong with ntpdate. > > There's nothing wrong with ntpdate provided there's nothing wrong with > having the wallclock make arbitrary backwards or forward jumps in > unpredictable intervals, IOW, if the only user is a human who doesn't > really care, everything's fine. Don't all time clients have an option to choose between all at once catch-up and slightly redefining a second so the catchup is more slow? Maybe ntpdate and that systemd thing also have such an option? SteveT Steve Litt January 2020 featured book: Troubleshooting: Just the Facts http://www.troubleshooters.com/tjust ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] download image ISO of beowulf
- Original Message - From: "Hendrik Boom" To: Sent: Tuesday, January 28, 2020 5:57 PM Subject: Re: [DNG] download image ISO of beowulf On Tue, Jan 28, 2020 at 05:23:46PM -0500, Haines Brown wrote: On Tue, Jan 28, 2020 at 10:13:25AM -0500, Ismael L. Donis Garcia wrote: > Is there any image ISO of beowulf that can be downloaded from some > site? > > Best Regards > -- > Ismael It seems most folks, like myself, got BEOWULF by upgraded from ASCII. I gather official ISOs are made available only for stabe releases. There were problems using the mini.iso installer. I suspect beowulf is going to be stable soon. I think we should get started debugging the official ISO so it will be stable too. -- hendrik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng What I wanted is to upgrade to CHIMAERA from BEOWULF In a virtual machine to test it, but starting from BEOWULF not from ASCII Thanks -- Ismael ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] systemd breaking linuxcnc ...
Dimitris via Dng writes: > On 1/28/20 6:46 PM, Rainer Weikusat via Dng wrote: >> run ntpdate from > > i've witnessed cases in the past where ntpd wasn't time-syncing > correctly, and ntpdate saved the date/time. > so, IMHO, there's nothing wrong with ntpdate. There's nothing wrong with ntpdate provided there's nothing wrong with having the wallclock make arbitrary backwards or forward jumps in unpredictable intervals, IOW, if the only user is a human who doesn't really care, everything's fine. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] systemd breaking linuxcnc ...
On 1/28/20 6:46 PM, Rainer Weikusat via Dng wrote: > run ntpdate from i've witnessed cases in the past where ntpd wasn't time-syncing correctly, and ntpdate saved the date/time. so, IMHO, there's nothing wrong with ntpdate. just 2c, d. signature.asc Description: OpenPGP digital signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
