role-ify

roles/tuned_amdgpu/templates/amdgpu-clock.sh.j2

@@ -0,0 +1,72 @@
+#!/bin/bash
+# script for tuned AMDGPU clock control
+# configures GPU power/clock characteristics
+# clocks/power in 3D are dynamic based on need/usage
+#
+# for 'amdgpu-default' tuned profiles, this will reset the characteristics to default
+# for others this will apply overclocking settings -- leaving clock choices to the associated power profile (eg: VR)
+#
+# rendered by Ansible with environment-appropriate values:
+#   card #, eg: card0
+#   path to discovered sysfs device files (power/clock/voltage control)
+#
+# AMDGPU driver/sysfs references:
+#   https://01.org/linuxgraphics/gfx-docs/drm/gpu/amdgpu.html
+#   https://docs.kernel.org/gpu/amdgpu/thermal.html
+
+{# done this way to avoid issues with the card number possibly shifting after playbook run #}
+# dynamically determine the connected GPU using the DRM subsystem
+CARD=$(/usr/bin/grep -ls ^connected /sys/class/drm/*/status | /usr/bin/grep -o 'card[0-9]' | /usr/bin/sort | /usr/bin/uniq | /usr/bin/sort -h | /usr/bin/tail -1)
+
+{# begin the templated script for 'default' profiles to reset state #}
+{% if 'default' in profile_name %}
+# set power state transition heuristics to default                                               
+echo '{{ item.0.value.pwrmode }}' | tee /sys/class/drm/"${CARD}"/device/pp_power_profile_mode
+
+# set control mode back to auto
+#   attempts to dynamically set optimal power profile for (load) conditions
+echo 'auto' | tee /sys/class/drm/"${CARD}"/device/power_dpm_force_performance_level
+
+# reset any existing profile clock changes
+echo 'r' | tee /sys/class/drm/"${CARD}"/device/pp_od_clk_voltage
+
+# give '{{ profile_name }}' profile ~{{ profile_percentage }}% (rounded) of the max power capability
+#   {{ profile_watts }} Watts of {{ board_watts }} total
+echo '{{ profile_microwatts | int }}' | tee '{{ powercap_set.files.0.path }}'
+{% else %}
+{# begin the templated script for non-default AMD GPU profiles, eg: 'VR' or '3D_FULL_SCREEN' #}
+# set manual control mode
+#   allows control via 'pp_dpm_mclk', 'pp_dpm_sclk', 'pp_dpm_pcie', 'pp_dpm_fclk', and 'pp_power_profile_mode' files
+#   only interested in 'pp_power_profile_mode' for power and 'pp_dpm_mclk' for memory clock (flickering).
+#   GPU clocks are dynamic based on (load) condition
+echo 'manual' | tee /sys/class/drm/"${CARD}"/device/power_dpm_force_performance_level
+
+# set power state transition heuristics to '{{ profile_name }}' profile
+echo '{{ item.0.value.pwrmode }}' | tee /sys/class/drm/"${CARD}"/device/pp_power_profile_mode
+
+# give '{{ profile_name }}' profile ~{{ profile_percentage }}% (rounded) of the max power capability
+#   {{ profile_watts }} Watts of {{ board_watts }} total
+echo '{{ profile_microwatts | int }}' | tee '{{ powercap_set.files.0.path }}'
+
+# set the minimum GPU clock
+echo 's 0 {{ gpu_clock_min }}' | tee /sys/class/drm/"${CARD}"/device/pp_od_clk_voltage
+
+# set the maximum GPU clock
+echo 's 1 {{ gpu_clock_max }}' | tee /sys/class/drm/"${CARD}"/device/pp_od_clk_voltage
+
+# set the maximum GPU *memory* clock
+echo 'm 1 {{ gpumem_clock_max }}' | tee /sys/class/drm/"${CARD}"/device/pp_od_clk_voltage
+{% if gpu_mv_offset is defined %}
+
+# offset GPU voltage {{ gpu_mv_offset }}mV
+echo 'vo {{ gpu_mv_offset }}' | tee /sys/class/drm/"${CARD}"/device/pp_od_clk_voltage
+{% endif %}
+
+# commit the changes
+echo 'c' | tee /sys/class/drm/"${CARD}"/device/pp_od_clk_voltage
+
+# force GPU memory into highest clock (fix flickering)
+#   pp_dpm_*clk settings are unintuitive, giving profiles that may be used
+#   opt not to set the others (eg: sclk/fclk) - those should remain for benefits from the curve
+echo '3' | tee /sys/class/drm/"${CARD}"/device/pp_dpm_mclk
+{% endif %}

roles/tuned_amdgpu/templates/tuned.conf.j2

@@ -0,0 +1,22 @@
+[main]
+include={{ item.1 }}
+summary={{ item.1 }} + TCP/RAID tweaks + AMDGPU pp_power_profile_mode = {{ item.0.value.pwrmode }} ({{ item.0.key }})
+
+[sysctl]
+net.core.default_qdisc=fq
+# 'bbr2' requires a [modified] supporting kernel - stock Fedora kernels do *not* support it (currently)
+#    eg: 'kernel-xanmode-edge' from COPR 'rmnscnce/kernel-xanmod'
+net.ipv4.tcp_congestion_control=bbr2
+net.core.rmem_max=33554432
+net.core.wmem_max=33554432
+dev.raid.speed_limit_min=600000
+dev.raid.speed_limit_max=9000000
+# allow some games to run (eg: DayZ)
+vm.max_map_count=1048576
+
+# reference/execute AMDGPU control script
+# used because some sysfs interfaces are transactional
+# cannot be set by a single param in 'tuned'/[sysfs]
+[gpuclockscript]
+type=script
+script=${i:PROFILE_DIR}/amdgpu-clock.sh