You are not logged in.
Did you see the other two posts suggesting "amdgpu.ppfeaturemask=0xfffd3fff" and "amd_iommu=off amdgpu.mcbp=0 "?
Online
I now have
initcall_blacklist=simpledrm_platform_driver_init amdgpu.runpm=0 amdgpu.bapm=0 amdgpu.aspm=0 amdgpu.ppfeaturemask=0xfffd3fff amd_iommu=off amdgpu.mcbp=0 initcall_blacklist=simpledrm_platform_driver_init
I also have VT-d disabled in the BIOS
Boot with the intel GPU into linux-drm-tip-git and the amdgpu: [gfxhub] page fault errors have gone.
xrandr --listproviders
Providers: number : 2
Provider 0: id: 0x47 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 4 outputs: 5 associated providers: 1 name:modesetting
Provider 1: id: 0xa0 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 4 outputs: 4 associated providers: 1 name:modesetting
[dik@archlinux ~]$ xrandr --prop
Screen 0: minimum 320 x 200, current 5120 x 1440, maximum 16384 x 16384
HDMI-1 connected primary 5120x1440+0+0 (normal left inverted right x axis y axis) 1200mm x 340mm
_MUTTER_PRESENTATION_OUTPUT: 0
EDID:
00ffffffffffff001e6d859e250b0600
0721010380782278e819c1ae5044af26
0e5054210900d1c081c0614045400101
01010101010174d600a0f03840403020
3a00ae514100001a000000fd0030901e
e176000a202020202020000000fc004c
4720554c545241574944450a000000ff
003330374e544b46424e3036390a03cb
f0027000000000000000000000000000
00000000000000000000000000000000
00000000000000000000000000000000
00000000000000000000000000000000
00000000000000000000000000000000
00000000000000000000000000000000
00000000000000000000000000000000
0000000000000000000000000000009e
02034570230907078301000050766160
5f5e5d3f22201f1312100403016d030c
002000b83c20006001020367d85dc401
788063e30f0700e2006ae305c000e606
0501606028f4b000a0f038354030203a
00ae514100001a1a6800a0f0381f4030
203a00ae514100001a6fc200a0a0a055
5030203500ae514100001a000000008a
701279030001000ccd2e2a0d0014a005
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
max bpc: 12
range: (8, 12)
content type: No Data
supported: No Data, Graphics, Photo, Cinema, Game
Colorspace: Default
supported: Default, SMPTE_170M_YCC, BT709_YCC, XVYCC_601, XVYCC_709, SYCC_601, opYCC_601, opRGB, BT2020_CYCC, BT2020_RGB, BT2020_YCC, DCI-P3_RGB_D65, DCI-P3_RGB_Theater
aspect ratio: Automatic
supported: Automatic, 4:3, 16:9
Broadcast RGB: Automatic
supported: Automatic, Full, Limited 16:235
audio: auto
supported: force-dvi, off, auto, on
link-status: Good
supported: Good, Bad
CONNECTOR_ID: 236
supported: 236
non-desktop: 0
range: (0, 1)
5120x1440 71.98*+
3840x2160 60.00 50.00 59.94 30.00 25.00 24.00 29.97 23.98
3840x1080 119.97 99.96 59.97
2560x1440 120.00
1920x1080 120.00 119.88 60.00 60.00 50.00 59.94 30.00 24.00 29.97 23.98
1280x1024 75.02
1280x720 60.00 50.00 59.94
1024x768 60.00
800x600 60.32
720x576 50.00
720x480 60.00 59.94
640x480 60.00 59.94
DP-1 disconnected (normal left inverted right x axis y axis)
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
vrr_capable: 0
range: (0, 1)
Colorspace: Default
supported: Default, BT709_YCC, XVYCC_601, XVYCC_709, SYCC_601, opYCC_601, opRGB, BT2020_CYCC, BT2020_RGB, BT2020_YCC, DCI-P3_RGB_D65, RGB_WIDE_FIXED, RGB_WIDE_FLOAT, BT601_YCC
max bpc: 12
range: (6, 12)
Broadcast RGB: Automatic
supported: Automatic, Full, Limited 16:235
audio: auto
supported: force-dvi, off, auto, on
subconnector: Unknown
supported: Unknown, VGA, DVI-D, HDMI, DP, Wireless, Native
link-status: Good
supported: Good, Bad
CONNECTOR_ID: 246
supported: 246
non-desktop: 0
range: (0, 1)
HDMI-4 disconnected (normal left inverted right x axis y axis)
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
max bpc: 12
range: (8, 12)
content type: No Data
supported: No Data, Graphics, Photo, Cinema, Game
Colorspace: Default
supported: Default, SMPTE_170M_YCC, BT709_YCC, XVYCC_601, XVYCC_709, SYCC_601, opYCC_601, opRGB, BT2020_CYCC, BT2020_RGB, BT2020_YCC, DCI-P3_RGB_D65, DCI-P3_RGB_Theater
aspect ratio: Automatic
supported: Automatic, 4:3, 16:9
Broadcast RGB: Automatic
supported: Automatic, Full, Limited 16:235
audio: auto
supported: force-dvi, off, auto, on
link-status: Good
supported: Good, Bad
CONNECTOR_ID: 255
supported: 255
non-desktop: 0
range: (0, 1)
DP-4 disconnected (normal left inverted right x axis y axis)
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
vrr_capable: 0
range: (0, 1)
Colorspace: Default
supported: Default, BT709_YCC, XVYCC_601, XVYCC_709, SYCC_601, opYCC_601, opRGB, BT2020_CYCC, BT2020_RGB, BT2020_YCC, DCI-P3_RGB_D65, RGB_WIDE_FIXED, RGB_WIDE_FLOAT, BT601_YCC
max bpc: 12
range: (6, 12)
Broadcast RGB: Automatic
supported: Automatic, Full, Limited 16:235
audio: auto
supported: force-dvi, off, auto, on
subconnector: Unknown
supported: Unknown, VGA, DVI-D, HDMI, DP, Wireless, Native
link-status: Good
supported: Good, Bad
CONNECTOR_ID: 259
supported: 259
non-desktop: 0
range: (0, 1)
HDMI-5 disconnected (normal left inverted right x axis y axis)
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
max bpc: 12
range: (8, 12)
content type: No Data
supported: No Data, Graphics, Photo, Cinema, Game
Colorspace: Default
supported: Default, SMPTE_170M_YCC, BT709_YCC, XVYCC_601, XVYCC_709, SYCC_601, opYCC_601, opRGB, BT2020_CYCC, BT2020_RGB, BT2020_YCC, DCI-P3_RGB_D65, DCI-P3_RGB_Theater
aspect ratio: Automatic
supported: Automatic, 4:3, 16:9
Broadcast RGB: Automatic
supported: Automatic, Full, Limited 16:235
audio: auto
supported: force-dvi, off, auto, on
link-status: Good
supported: Good, Bad
CONNECTOR_ID: 267
supported: 267
non-desktop: 0
range: (0, 1)
DP-1-2 disconnected (normal left inverted right x axis y axis)
PRIME Synchronization: 1
supported: 0, 1
subconnector: Unknown
supported: Unknown, VGA, DVI-D, HDMI, DP, Wireless, Native
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
vrr_capable: 0
range: (0, 1)
Colorspace: Default
supported: Default, BT709_YCC, opRGB, BT2020_RGB, BT2020_YCC
max bpc: 16
range: (8, 16)
underscan vborder: 0
range: (0, 128)
underscan hborder: 0
range: (0, 128)
underscan: off
supported: off, on, auto
scaling mode: None
supported: None, Full, Center, Full aspect
link-status: Good
supported: Good, Bad
CTM: 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0
0 1
CONNECTOR_ID: 93
supported: 93
non-desktop: 0
range: (0, 1)
DP-1-3 disconnected (normal left inverted right x axis y axis)
PRIME Synchronization: 1
supported: 0, 1
subconnector: Unknown
supported: Unknown, VGA, DVI-D, HDMI, DP, Wireless, Native
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
vrr_capable: 0
range: (0, 1)
Colorspace: Default
supported: Default, BT709_YCC, opRGB, BT2020_RGB, BT2020_YCC
max bpc: 16
range: (8, 16)
underscan vborder: 0
range: (0, 128)
underscan hborder: 0
range: (0, 128)
underscan: off
supported: off, on, auto
scaling mode: None
supported: None, Full, Center, Full aspect
link-status: Good
supported: Good, Bad
CTM: 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0
0 1
CONNECTOR_ID: 101
supported: 101
non-desktop: 0
range: (0, 1)
HDMI-1-2 disconnected (normal left inverted right x axis y axis)
PRIME Synchronization: 1
supported: 0, 1
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
vrr_capable: 0
range: (0, 1)
Colorspace: Default
supported: Default, BT709_YCC, opRGB, BT2020_RGB, BT2020_YCC
content type: No Data
supported: No Data, Graphics, Photo, Cinema, Game
max bpc: 16
range: (8, 16)
underscan vborder: 0
range: (0, 128)
underscan hborder: 0
range: (0, 128)
underscan: off
supported: off, on, auto
scaling mode: None
supported: None, Full, Center, Full aspect
link-status: Good
supported: Good, Bad
CTM: 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0
0 1
CONNECTOR_ID: 106
supported: 106
non-desktop: 0
range: (0, 1)
HDMI-1-3 disconnected (normal left inverted right x axis y axis)
PRIME Synchronization: 1
supported: 0, 1
HDCP Content Type: HDCP Type0
supported: HDCP Type0, HDCP Type1
Content Protection: Undesired
supported: Undesired, Desired, Enabled
vrr_capable: 0
range: (0, 1)
Colorspace: Default
supported: Default, BT709_YCC, opRGB, BT2020_RGB, BT2020_YCC
content type: No Data
supported: No Data, Graphics, Photo, Cinema, Game
max bpc: 16
range: (8, 16)
underscan vborder: 0
range: (0, 128)
underscan hborder: 0
range: (0, 128)
underscan: off
supported: off, on, auto
scaling mode: None
supported: None, Full, Center, Full aspect
link-status: Good
supported: Good, Bad
CTM: 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0
0 1
CONNECTOR_ID: 112
supported: 112
non-desktop: 0
range: (0, 1)
xrandr --output HDMI-1 --set "PRIME Synchronization" 1
X Error of failed request: BadName (named color or font does not exist)
Major opcode of failed request: 140 (RANDR)
Minor opcode of failed request: 11 (RRQueryOutputProperty)
Serial number of failed request: 59
Current serial number in output stream: 59
Last edited by dIQi (2023-11-16 18:59:03)
Offline
Just realised my copy / paste must have got a bit excited and I have initcall_blacklist=simpledrm_platform_driver_init twice. Don't think that matters, please correct me if I'm wrong
Offline
You can add this as many times as you want. To be sure you can check the system journal whether the simpledrm device still shows up, but it won't.
Still no help on up-clocking the GPU? I guess.
Only the disconnected DP-1-2, DP-1-3, HDMI-1-2 and HDMI-1-3 have prime sync attributes - it seems it's only available for outputs that are not attached to the main GPU (for regular reverse prime)
So that's a dead end either
Online
DP-1-2, DP-1-3, HDMI-1-2 and HDMI-1-3 could well be the usb-c connections, what do you think?
Spec is here : https://www.intel.com/content/dam/suppo … odSpec.pdf
Offline
Just found this - some guy is running a NVIDIA 4090 under Rocky Linux 8.5. Going to try Rocky Linux USB, I've not tried any RPM based Linux yet. Not holding out a lot of hope
Offline
Tried Rocky 9 live but it came with 5.14 which doesn't have 7800 XT support. Installed on the thumb drive and updated to 6.6 but it was full of amdgpu errors so gave up with it
Decided to try Fedora 39 on the thumb drive. That came with 6.5.11-300.fc39.x86_64 and gave me the same results. Updated to 6.7.0-0.rc1.20231119gt037266a5.320.vanilla.fc39.x86_64, same results
Also tried Tumbleweed. Same results with both 6.6.1-1-default and 6.7~rc1-2.1.g86e46c2
Last edited by dIQi (2023-11-19 20:30:00)
Offline
Looks like I might be able to borrow a RX 580 and a GTX 1060 but not for a couple of days
Offline
Remembered that overclocking with AMD Navi 31/32/33 will not work until kernel 6.7 at the earliest
Installed 6.7.0-rc3-1-mainline from miffe
Added amdgpu.ppfeaturemask=0xffffffff
CoreCtrl works!
If I set Performance mode to Advanced and change nothing else the GPU memory finally rises above 96Mhz in the benchmark
The result isn't great but it's a step in the right direction
Offline
Nearly the same situation here (NUC13RNGi7 + RX 7900 XT)! Thus, waiting for Kernel 6.7.* is the only solution for now?
Offline
Nearly the same situation here (NUC13RNGi7 + RX 7900 XT)! Thus, waiting for Kernel 6.7.* is the only solution for now?
It does look that way
Offline
Regarding your The i7-13700K. On Linux there is no prioritization for the p-cores and e-cores. This messed up my game performance. Disable all e-cores in BIOS.
Offline
6.7.0-rc4-1-mainline seemed to regress but 6.7.0-rc5-1-mainline is working again, although I'm still not getting anywhere near what I was getting on my old 4790k
Offline
Regarding your The i7-13700K. On Linux there is no prioritization for the p-cores and e-cores. This messed up my game performance. Disable all e-cores in BIOS.
Thanks for the information. I've just tried with all e cores disabled and all e and p cores enabled. As you can see I'm getting very slightly better results with all cores enabled
6.7.0-rc5-1-mainline is giving better results than 6.7.0-rc3-1-mainline so it looks to be moving in the right direction but it's still not fixed
Offline
I seemed to have missed the 6.7.0-rc6-1-mainline release but 6.7.0-rc7-1-mainline gives me this which is another step backwards
Offline
That's a really informative benchmark result /sarcasm.
From the looks I'd say it actually ran on the IGP ("Driver: i915")?
Online
nvtop shows me it's running on the amd gpu
The benchmark may not be the best but it's all I have to gauge any progress made in 6.7. If there's something better I'm open to suggestions. I'm gradually losing hope for 6.7 getting anywhere near fixing this
Offline
https://wiki.archlinux.org/title/Benchmarking#Graphics
Are the results w/ the previous rc reproducible (to rule out flukes)?
If so it'd be rather easy to bisect the offending commit and file a bug, fixing by hope it will probably not.
Online
All of the rcs are giving results in the same range and the differences could easily be flukes
Not one single rc has yet given a result more than 20% of what I obtained by moving the drive and gpu to my old 4790k or from running Windows on this machine. At the same time there's a definite improvement over 6.6
I've just tried rc8 with almost the exact same results as rc7
Still the gpu is working similar to 6.6 with the mem stuck at 97Mhz (depending on monitor refresh rate). With the 6.7rcs this is the same until I increase the min mem MHz in CoreCtrl from 97 to 98 (or anything). After that I am able to reset CoreCtrl to default and the mem in nvtop increases above 97Mhz when the benchmark is running
I already have a bug report open here but it's not going anywhere so I'm left with documenting my results
I have no experience bisecting
Offline
https://wiki.archlinux.org/title/Bisect … s_with_Git - but if it's not a regression and/or there's no reliable test of the impact, that's moot.
If I set Performance mode to Advanced and change nothing else the GPU memory finally rises above 96Mhz in the benchmark
slightly spins up mclk $ cat /sys/class/drm/card1/device/pp_dpm_mclk 0: 96Mhz 1: 456Mhz * 2: 772Mhz 3: 1249Mhz $ cat /sys/class/drm/card1/device/pp_dpm_sclk 0: 500Mhz 1: 10Mhz * 2: 2025Mhz
but the performance still sucks badly
Is the change likelwise (30%/50%) limited for you?
https://wiki.archlinux.org/title/AMDGPU#Power_profiles
Is https://wiki.archlinux.org/title/AMDGPU … _(default) now possible?
Online
Looks like mainline has gone stable now - 6.7-1
Will test as soon as I can, time poor at the moment
Offline
Is the change likelwise (30%/50%) limited for you?
https://wiki.archlinux.org/title/AMDGPU#Power_profiles
Is https://wiki.archlinux.org/title/AMDGPU … _(default) now possible?
Now using 6.7.0-arch3-1
I'm stuck at around 20% of what I'd expect from testing with Windows and my old 4th gen i7
ls /sys/kernel/debug/dri/2/amdgpu_pm_info
/sys/kernel/debug/dri/2/amdgpu_pm_info
cd /sys/class/drm/card2/device/
cat power_dpm_force_performance_level
auto
echo "manual" > power_dpm_force_performance_level
cat power_dpm_force_performance_level
manual
cat pp_power_profile_mode
0 BOOTUP_DEFAULT*:
0( GFXCLK) 0 1 0 4 800 4587520 -65536 0
1( FCLK) 0 3 0 1 0 3276800 -65536 -6553
1 3D_FULL_SCREEN :
0( GFXCLK) 0 0 1200 4 650 3932160 -3276 -65536
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
2 POWER_SAVING :
0( GFXCLK) 0 1 0 3 0 5898240 -65536 0
1( FCLK) 0 1 0 1 0 3407872 -65536 -6553
3 VIDEO :
0( GFXCLK) 0 1 0 4 500 4587520 -65536 0
1( FCLK) 0 3 0 3 0 3473408 -65536 -6553
4 VR :
0( GFXCLK) 0 2 1000 1 0 3276800 0 0
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
5 COMPUTE :
0( GFXCLK) 0 2 1000 1 0 3932160 0 0
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
6 CUSTOM :
0( GFXCLK) 0 0 1200 4 0 655360 -3276 -65536
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
7 WINDOW_3D :
0( GFXCLK) 0 0 1200 4 650 3932160 -3276 -65536
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
At this point mem is still stuck at 96MHz in the benchmark
echo "1" > pp_power_profile_mode
cat pp_power_profile_mode
0 BOOTUP_DEFAULT :
0( GFXCLK) 0 1 0 4 800 4587520 -65536 0
1( FCLK) 0 3 0 1 0 3276800 -65536 -6553
1 3D_FULL_SCREEN*:
0( GFXCLK) 0 0 1200 4 650 3932160 -3276 -65536
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
2 POWER_SAVING :
0( GFXCLK) 0 1 0 3 0 5898240 -65536 0
1( FCLK) 0 1 0 1 0 3407872 -65536 -6553
3 VIDEO :
0( GFXCLK) 0 1 0 4 500 4587520 -65536 0
1( FCLK) 0 3 0 3 0 3473408 -65536 -6553
4 VR :
0( GFXCLK) 0 2 1000 1 0 3276800 0 0
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
5 COMPUTE :
0( GFXCLK) 0 2 1000 1 0 3932160 0 0
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
6 CUSTOM :
0( GFXCLK) 0 0 1200 4 0 655360 -3276 -65536
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
7 WINDOW_3D :
0( GFXCLK) 0 0 1200 4 650 3932160 -3276 -65536
1( FCLK) 0 3 0 3 0 1310720 -6553 -6553
Mem still stuck at 96MHz
cat pp_dpm_mclk
0: 96Mhz *
1: 456Mhz
2: 772Mhz
3: 1218Mhz
echo "3" > pp_dpm_mclk
cat pp_dpm_mclk
0: 96Mhz *
1: 456Mhz
2: 772Mhz
3: 1218Mhz
Still stuck at 96Mhz in the benchmark
corectrl 1.3.9-1
Set performance mode to Advanced
Increase Memory minimum from 97 to 98 and apply (not sure why that's 97 not 96)
Memory is now no longer stuck at 96Mhz in the benchmark. However performance is ~20% of expected
In corectrl choose Reset - "Do you want to load the default settings" - Yes
Performance mode is now set to Automatic again and the mem is not stuck at 96Mhz in the benchmark. Performance is ~20% of expected
Last edited by dIQi (2024-01-21 19:28:07)
Offline
This is what I'm getting in the benchmark now, nowhere near what I got on my 4th gen i7
Offline
Your 4th gen i7 uses a hybrid graphics setup where superposition runs on the intel i915 driver.
The current setup is not a hybrid setup (either the cpu has no internal intel graphics or they are disabled) and runs superpostion on the amdgpu driver.
Please stop comparing oranges to pears.
Last edited by Lone_Wolf (2024-01-22 11:24:01)
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Offline
I had a very similar, although slightly higher, result running Windows on my current machine to what the 4th Gen i7 achieved
Both had results in the same region as the phoronix article so I presume the GPU works as it should in those situations
Last edited by dIQi (2024-01-22 19:07:06)
Offline