<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body bgcolor="white" lang="EN-US" link="blue" vlink="purple"> <div class="WordSection1"> +Alex,<o:p></o:p> <o:p> </o:p> Hi Yanik,<o:p></o:p> <o:p> </o:p> Did you ever run the following command to let your OD settings take effect (before running games)? Otherwise, they did not take effect actually.<o:p></o:p> echo "c" > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_od_clk_voltage<o:p></o:p> <o:p> </o:p> Regards,<o:p></o:p> Evan<o:p></o:p> <div style="border:none;border-left:solid blue 1.5pt;padding:0in 0in 0in 4.0pt"> <div> <div style="border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0in 0in 0in"> From: Yanik Yiannakis <yanik@yiannakis.de> Sent: Monday, April 29, 2019 7:44 AM To: rex.zhu@amd.com; Quan, Evan <Evan.Quan@amd.com>; amd-gfx@lists.freedesktop.org Subject: Bug Report: [PowerPlay] MCLK can't be set above 1107MHz on Vega 64<o:p></o:p> </div> </div> <o:p> </o:p> Hello,<o:p></o:p> I experience a bug that prevents me from setting the MCLK of my Vega 64 LC above 1107MHz. <o:p></o:p> I am using Unigine Superposition 1.1 in "Game"-mode to check the performance by watching the FPS.<o:p></o:p> <o:p> </o:p> Behaviour with a single monitor:<o:p></o:p> First I set the MCLK to a known stable value below 1108MHz:<o:p></o:p> $ echo "m 3 1100 950" > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_od_clk_voltage <o:p></o:p> In Unigine Superposition the FPS increase as expected. <o:p></o:p> pp_dpm_mclk also confirms the change. <o:p></o:p> $ watch cat /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_dpm_mclk<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> 0: 167Mhz 1: 500Mhz 2: 800Mhz 3: 1100Mhz *<o:p></o:p> </blockquote> <o:p> </o:p> After that I set the MCLK to a stable value above 1107MHz:<o:p></o:p> $ echo "m 3 1200 950" > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_od_clk_voltage <o:p></o:p> In Unigine Superposition the FPS drop drastically.<o:p></o:p> pp_dpm_mclk indicates that the MCLK is stuck in state 0 (167MHz):<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> 0: 167Mhz * 1: 500Mhz 2: 800Mhz 3: 1200Mhz<o:p></o:p> </blockquote> <o:p> </o:p> Behaviour with multiple monitors that have different refresh rates:<o:p></o:p> My monitors have different refresh rates. This causes the MCLK to stay in state 3 (945MHz stock) which is the expected behaviour as I understand it.<o:p></o:p> <o:p> </o:p> Now I try to set the MCLK to a value above 1107MHz:<o:p></o:p> $ echo "m 3 1200 950" > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_od_clk_voltage <o:p></o:p> The FPS in Unigine Superposition remain the same as they were with 945MHz.<o:p></o:p> pp_dpm_mclk shows however that the value was set:<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> 0: 167Mhz 1: 500Mhz 2: 800Mhz 3: 1200Mhz *<o:p></o:p> </blockquote> <o:p> </o:p> Then I set the MCLK to a value of 1107MHz or lower:<o:p></o:p> $ echo "m 3 1100 950" > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_od_clk_voltage <o:p></o:p> The FPS in Unigine Superposition increase.<o:p></o:p> pp_dpm_mclk again confirms the set value:<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> 0: 167Mhz 1: 500Mhz 2: 800Mhz 3: 1100Mhz *<o:p></o:p> </blockquote> <o:p> </o:p> Finally I increase MCLK to a known unstable value:<o:p></o:p> $ echo "m 3 1300 950" > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/pp_od_clk_voltage <o:p></o:p> The FPS in Unigine Superposition remain the same. I therefore believe the value was not actually applied.<o:p></o:p> However pp_dpm_mclk shows that it was:<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> 0: 167Mhz 1: 500Mhz 2: 800Mhz 3: 1300Mhz *<o:p></o:p> </blockquote> <o:p> </o:p> amdgpu_pm_info also claims that the value was set:<o:p></o:p> $ sudo watch cat /sys/kernel/debug/dri/1/amdgpu_pm_info<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> GFX Clocks and Power: 1300 MHz (MCLK) 27 MHz (SCLK) 1348 MHz (PSTATE_SCLK) 800 MHz (PSTATE_MCLK) 825 mV (VDDGFX) 4.0 W (average GPU)<o:p></o:p> </blockquote> Again, I think the displayed MCLK is false and the memory still runs at 1100MHz because the performance in Unigine Superposition indicates this and 1300MHz would cause a crash immediately.<o:p></o:p> A stable value (e.g. 1200MHz) causes the same behaviour. I just chose 1300MHz to be sure.<o:p></o:p> <o:p> </o:p> <o:p> </o:p> Tested on these Kernels:<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> Arch-Linux 5.0.9 (Arch)<o:p></o:p> Linux 5.1-rc6 (Ubuntu)<o:p></o:p> Linux 5.0 with amd-staging-drm-next (Ubuntu) (<a href="https://github.com/M-Bab/linux-kernel-amdgpu-binaries">https://github.com/M-Bab/linux-kernel-amdgpu-binaries</a>)<o:p></o:p> </blockquote> (Same behaviour on every kernel.)<o:p></o:p> <o:p> </o:p> Tested on this hardware:<o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> CPU: Intel i7-8700k<o:p></o:p> Motherboard: MSI Z370 Gaming Pro Carbon<o:p></o:p> GPU: Powercolor Vega 64 Liquid Cooled (Memory stable below 1220MHz, tested on Windows 10 with Wattman and Unigine Superposition)<o:p></o:p> </blockquote> <o:p> </o:p> Unigine Superposition "Game"-Mode settings: <o:p></o:p> <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt"> Preset: Custom<o:p></o:p> Fullscreen: Disabled<o:p></o:p> Resolution: 3840x2160 (4K UHD)<o:p></o:p> Shaders Quality: Extreme<o:p></o:p> Textures Quality: High<o:p></o:p> Vsync: Off<o:p></o:p> Depth of Field: On<o:p></o:p> Motion Blur: On<o:p></o:p> </blockquote> <o:p> </o:p> I hope this helps.<o:p></o:p> Yanik Yiannakis<o:p></o:p> </div> </div> </body> </html>