Arch Linux

cncz

Member

Registered: 2021-05-07

Posts: 11

[SOLVED] 7900xt sudden performance drop

After months of stability my 7900xt suddenly dropped in performance with reported clocks never exceeding 500mhz and power usage hovering at around 80w. It started mid-session in KDE Plasma Wayland when I discord screenshared a minecraft window, then minimized it. I Tried every relevant fix on https://wiki.archlinux.org/title/AMDGPU as well as reinstalling linux kernel firmware and mesa drivers. It seems that the power modes don't affect the gpu clocks at all anymore, and due to https://gitlab.freedesktop.org/drm/amd/-/issues/2356 I can't manually set the power limit to stably overclock it to stock, I also attempted switching the vbios to the backup one provided on the card to no effect. Power modes in /sys/class/drm/card0/device/pp_power_profile_mode are also showing a clock of 1200mhz, which mangohud shows as the gpu clock constantly. This issue does not appear on windows, although I have yet to try another Linux environment. Dmesg did report a firmware "gc_11_0_0_mes_2.bin" failing to load but manually re-adding that firmware to mkinitcpio failed to change anything. Comparing journalctl from before and after the incident shows no difference in messages during boot.

output of dmesg | grep amd:

 [    0.000000] Command line: initrd=\amd-ucode.img initrd=\initramfs-linux-zen.img cryptdevice=PARTUUID=0d57f95f-3ecc-4ebb-927e-9646a810b90f:luksdev root=/dev/mapper/luksdev zswap.enabled=0 rootflags=subvol=@ rw rootfstype=btrfs quiet splash amdgpu.ppfeaturemask=0xfff7ffff
[    0.046595] Kernel command line: initrd=\amd-ucode.img initrd=\initramfs-linux-zen.img cryptdevice=PARTUUID=0d57f95f-3ecc-4ebb-927e-9646a810b90f:luksdev root=/dev/mapper/luksdev zswap.enabled=0 rootflags=subvol=@ rw rootfstype=btrfs quiet splash amdgpu.ppfeaturemask=0xfff7ffff
[    0.485714] perf/amd_iommu: Detected AMD IOMMU #0 (2 banks, 4 counters/bank).
[    0.522440] amd_pstate: driver load is disabled, boot with specific mode to enable this
[   25.439018] kvm_amd: TSC scaling supported
[   25.439022] kvm_amd: Nested Virtualization enabled
[   25.439023] kvm_amd: Nested Paging enabled
[   25.439025] kvm_amd: SEV supported: 509 ASIDs
[   25.439067] kvm_amd: Virtual VMLOAD VMSAVE supported
[   25.439068] kvm_amd: Virtual GIF supported
[   25.439068] kvm_amd: LBR virtualization supported
[   27.036609] [drm] amdgpu kernel modesetting enabled.
[   27.042134] amdgpu: Ignoring ACPI CRAT on non-APU system
[   27.042137] amdgpu: Virtual CRAT table created for CPU
[   27.042145] amdgpu: Topology: Add CPU node
[   27.056199] amdgpu 0000:0c:00.0: No more image in the PCI ROM
[   27.056218] amdgpu 0000:0c:00.0: amdgpu: Fetched VBIOS from ROM BAR
[   27.056220] amdgpu: ATOM BIOS: 113-31TESHBS1-L01
[   27.089422] amdgpu 0000:0c:00.0: amdgpu: CP RS64 enable
[   27.114835] amdgpu 0000:0c:00.0: [drm:jpeg_v4_0_early_init [amdgpu]] JPEG decode is enabled in VM mode
[   27.115228] amdgpu 0000:0c:00.0: Direct firmware load for amdgpu/gc_11_0_0_mes_2.bin failed with error -2
[   27.115230] [drm] try to fall back to amdgpu/gc_11_0_0_mes.bin
[   27.121205] amdgpu 0000:0c:00.0: vgaarb: deactivate vga console
[   27.121209] amdgpu 0000:0c:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
[   27.121304] amdgpu 0000:0c:00.0: amdgpu: MEM ECC is not presented.
[   27.121306] amdgpu 0000:0c:00.0: amdgpu: SRAM ECC is not presented.
[   27.121323] amdgpu 0000:0c:00.0: amdgpu: VRAM: 20464M 0x0000008000000000 - 0x00000084FEFFFFFF (20464M used)
[   27.121325] amdgpu 0000:0c:00.0: amdgpu: GART: 512M 0x0000000000000000 - 0x000000001FFFFFFF
[   27.121326] amdgpu 0000:0c:00.0: amdgpu: AGP: 267878400M 0x0000008800000000 - 0x0000FFFFFFFFFFFF
[   27.121452] [drm] amdgpu: 20464M of VRAM memory ready
[   27.121455] [drm] amdgpu: 15999M of GTT memory ready.
[   27.126463] amdgpu 0000:0c:00.0: amdgpu: Will use PSP to load VCN firmware
[   27.376993] amdgpu 0000:0c:00.0: amdgpu: RAP: optional rap ta ucode is not available
[   27.377000] amdgpu 0000:0c:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
[   27.377057] amdgpu 0000:0c:00.0: amdgpu: smu driver if version = 0x00000037, smu fw if version = 0x00000034, smu fw program = 0, smu fw version = 0x004e4b00 (78.75.0)
[   27.377060] amdgpu 0000:0c:00.0: amdgpu: SMU driver if version not matched
[   27.528742] amdgpu 0000:0c:00.0: amdgpu: SMU is initialized successfully!
[   27.554931] snd_hda_intel 0000:0c:00.1: bound 0000:0c:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])
[   27.883267] amdgpu 0000:0c:00.0: [drm:jpeg_v4_0_hw_init [amdgpu]] JPEG decode initialized successfully.
[   27.885228] kfd kfd: amdgpu: Allocated 3969056 bytes on gart
[   27.885261] amdgpu: sdma_bitmap: fff0
[   27.941989] amdgpu: HMM registered 20464MB device memory
[   27.942000] amdgpu: SRAT table not found
[   27.942001] amdgpu: Virtual CRAT table created for GPU
[   27.942149] amdgpu: Topology: Add dGPU node [0x744c:0x1002]
[   27.942151] kfd kfd: amdgpu: added device 1002:744c
[   27.942163] amdgpu 0000:0c:00.0: amdgpu: SE 6, SH per SE 2, CU per SH 8, active_cu_number 84
[   27.942227] amdgpu 0000:0c:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[   27.942229] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[   27.942230] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[   27.942231] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[   27.942232] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[   27.942233] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[   27.942233] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[   27.942234] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[   27.942235] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[   27.942236] amdgpu 0000:0c:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
[   27.942237] amdgpu 0000:0c:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0
[   27.942238] amdgpu 0000:0c:00.0: amdgpu: ring vcn_unified_0 uses VM inv eng 0 on hub 1
[   27.942239] amdgpu 0000:0c:00.0: amdgpu: ring vcn_unified_1 uses VM inv eng 1 on hub 1
[   27.942239] amdgpu 0000:0c:00.0: amdgpu: ring jpeg_dec uses VM inv eng 4 on hub 1
[   27.942240] amdgpu 0000:0c:00.0: amdgpu: ring mes_kiq_3.1.0 uses VM inv eng 14 on hub 0
[   27.950504] amdgpu 0000:0c:00.0: amdgpu: Using BACO for runtime pm
[   27.950789] [drm] Initialized amdgpu 3.52.0 20150101 for 0000:0c:00.0 on minor 0
[   27.961314] fbcon: amdgpudrmfb (fb0) is primary device
[   27.961318] amdgpu 0000:0c:00.0: [drm] fb0: amdgpudrmfb frame buffer device
[   28.152494] WARNING: CPU: 4 PID: 364 at drivers/gpu/drm/amd/amdgpu/../display/dc/dcn32/dcn32_resource_helpers.c:329 dcn32_determine_det_override+0x143/0xb00 [amdgpu]
[   28.152755] Modules linked in: vfat fat amdgpu snd_hda_codec_realtek snd_hda_codec_generic intel_rapl_msr intel_rapl_common snd_hda_codec_hdmi edac_mce_amd snd_usb_audio snd_hda_intel btusb snd_intel_dspcfg snd_usbmidi_lib btrtl snd_intel_sdw_acpi kvm_amd drm_buddy snd_rawmidi btbcm snd_hda_codec eeepc_wmi snd_seq_device gpu_sched snd_hda_core btintel asus_wmi mc drm_ttm_helper btmtk snd_hwdep kvm ledtrig_audio ttm hid_playstation snd_pcm sparse_keymap r8169 platform_profile drm_display_helper irqbypass mxm_wmi video asus_ec_sensors wmi_bmof bluetooth snd_timer led_class_multicolor cec ff_memless realtek sp5100_tco igb snd rapl pcspkr mdio_devres ecdh_generic soundcore i2c_piix4 k10temp i2c_algo_bit rfkill crc16 libphy dca wmi mousedev joydev acpi_cpufreq mac_hid ip6t_REJECT xt_hl ip6t_rt ipt_REJECT xt_LOG nf_log_syslog xt_limit xt_addrtype xt_tcpudp xt_conntrack ip6table_filter ip6_tables iptable_filter nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_masq nft_ct nft_chain_nat nf_nat nf_conntrack
[   28.152816] RIP: 0010:dcn32_determine_det_override+0x143/0xb00 [amdgpu]
[   28.153076]  ? dcn32_determine_det_override+0x143/0xb00 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.153338]  ? dcn32_determine_det_override+0x143/0xb00 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.153577]  ? dcn32_predict_pipe_split+0x187/0x230 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.153841]  dcn32_populate_dml_pipes_from_context+0x591/0x900 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.154087]  dcn32_internal_validate_bw+0x12c/0x2cb0 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.154356]  dcn32_validate_bandwidth+0x213/0x320 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.154598]  dc_validate_global_state+0x3db/0x580 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.154833]  amdgpu_dm_atomic_check+0x1183/0x1400 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]
[   28.155114]  amdgpu_drm_ioctl+0x4e/0x90 [amdgpu 4a0cb193d9a6e95ff7a4a07c722a059ca0ca5eed]

Last edited by cncz (2023-07-20 17:24:18)

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 25,194

Re: [SOLVED] 7900xt sudden performance drop

If it literally "suddenly dropped" without you updating/changing linux-firmware or the kernel then you might have to worry about HW troubles. Since I'd consider that less likely consider downgrading the kernel (... and linux-firmware) to a known working version. FWIW since you mention Windows, make sure it's fast boot setting is disabled as that can lead to all sorts of weirdness: https://wiki.archlinux.org/title/Dual_b … ibernation

MikeAc

Member

Registered: 2023-06-21

Posts: 2

Re: [SOLVED] 7900xt sudden performance drop

I have the same issue with my gpu and it is totally new. I have these errors with kernel 6.3 for some reasons. This issue is gone in kernel 6.1. I  think it might be releated to the firmware that leads to some error when the driver is probing but not pretty sure.

Kernel 6.3(error)

  
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: [drm] fb0: amdgpudrmfb frame buffer device
Jun 21 09:45:17 SuperTux kernel:  ? __pfx_init_module+0x10/0x10 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  amdgpu_pci_probe+0x37b/0x3f0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  amdgpu_dm_atomic_check+0x10e5/0x1370 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  dc_validate_global_state+0x3d4/0x580 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  dcn32_validate_bandwidth+0x1f5/0x320 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  ? dcn32_validate_bandwidth+0x6f/0x320 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  dcn32_internal_validate_bw+0x109/0x2bf0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  dcn32_populate_dml_pipes_from_context+0x403/0x7d0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  ? dcn32_predict_pipe_split+0x168/0x210 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  ? dcn32_determine_det_override+0x13f/0xab0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel:  ? dcn32_determine_det_override+0x13f/0xab0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel: RIP: 0010:dcn32_determine_det_override+0x13f/0xab0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel: Modules linked in: ext4 nls_iso8859_1 mbcache vfat squashfs jbd2 fat intel_rapl_msr amdgpu(+) intel_rapl_common snd_sof_pci_intel_tgl snd_sof_intel_hda_common soundwire_intel soundwire_generic_allocation soundwire_cadence snd_sof_intel_hda snd_sof_pci snd_sof_xtensa_dsp snd_sof snd_sof_utils snd_soc_hdac_hda x86_pkg_temp_thermal snd_hda_ext_core snd_soc_acpi_intel_match snd_soc_acpi intel_powerclamp snd_soc_core iwlmvm snd_compress kvm_intel snd_pcm_dmaengine snd_hda_codec_realtek ac97_bus soundwire_bus mac80211 kvm snd_hda_codec_generic snd_hda_codec_hdmi btusb libarc4 snd_hda_intel btrtl iommu_v2 btbcm snd_intel_dspcfg drm_buddy irqbypass snd_usb_audio btintel snd_intel_sdw_acpi gpu_sched uvcvideo snd_hda_codec i2c_algo_bit rapl i2c_designware_platform snd_usbmidi_lib drm_ttm_helper uvc i2c_designware_core snd_hda_core intel_cstate bluetooth ttm snd_rawmidi videobuf2_vmalloc iwlwifi snd_hwdep snd_seq_device videobuf2_memops drm_display_helper asus_nb_wmi videobuf2_v4l2 eeepc_wmi snd_pcm asus_wmi
Jun 21 09:45:17 SuperTux kernel: WARNING: CPU: 8 PID: 1296 at drivers/gpu/drm/amd/amdgpu/../display/dc/dcn32/dcn32_resource_helpers.c:329 dcn32_determine_det_override+0x13f/0xab0 [amdgpu]
Jun 21 09:45:17 SuperTux kernel: fbcon: amdgpudrmfb (fb0) is primary device
Jun 21 09:45:17 SuperTux kernel: [drm] Initialized amdgpu 3.52.0 20150101 for 0000:03:00.0 on minor 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Using BACO for runtime pm
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring mes_kiq_3.1.0 uses VM inv eng 14 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring jpeg_dec uses VM inv eng 4 on hub 1
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring vcn_unified_1 uses VM inv eng 1 on hub 1
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring vcn_unified_0 uses VM inv eng 0 on hub 1
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 11 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 10 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 9 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 8 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 7 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 6 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SE 6, SH per SE 2, CU per SH 8, active_cu_number 96
Jun 21 09:45:17 SuperTux kernel: kfd kfd: amdgpu: added device 1002:744c
Jun 21 09:45:17 SuperTux kernel: amdgpu: Topology: Add dGPU node [0x744c:0x1002]
Jun 21 09:45:17 SuperTux kernel: amdgpu: Virtual CRAT table created for GPU
Jun 21 09:45:17 SuperTux kernel: amdgpu: SRAT table not found
Jun 21 09:45:17 SuperTux kernel: amdgpu: HMM registered 24560MB device memory
Jun 21 09:45:17 SuperTux kernel: amdgpu: sdma_bitmap: fff0
Jun 21 09:45:17 SuperTux kernel: kfd kfd: amdgpu: Allocated 3969056 bytes on gart
Jun 21 09:45:17 SuperTux kernel: amdgpu 0000:03:00.0: [drm:jpeg_v4_0_hw_init [amdgpu]] JPEG decode initialized successfully.
Jun 21 09:45:16 SuperTux kernel: snd_hda_intel 0000:03:00.1: bound 0000:03:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SMU is initialized successfully!
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SMU driver if version not matched
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: smu driver if version = 0x00000037, smu fw if version = 0x00000034, smu fw program = 0, smu fw version = 0x004e4b00 (78.75.0)
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: RAP: optional rap ta ucode is not available
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Will use PSP to load VCN firmware
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/sdma_6_0_0.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_imu.bin
Jun 21 09:45:16 SuperTux kernel: [drm] amdgpu: 32019M of GTT memory ready.
Jun 21 09:45:16 SuperTux kernel: [drm] amdgpu: 24560M of VRAM memory ready
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: AGP: 267878400M 0x0000008800000000 - 0x0000FFFFFFFFFFFF
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: GART: 512M 0x0000000000000000 - 0x000000001FFFFFFF
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: VRAM: 24560M 0x0000008000000000 - 0x00000085FEFFFFFF (24560M used)
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SRAM ECC is not presented.
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: MEM ECC is not presented.
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: vgaarb: deactivate vga console
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mes1.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mes.bin
Jun 21 09:45:16 SuperTux kernel: [drm] try to fall back to amdgpu/gc_11_0_0_mes.bin
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: Direct firmware load for amdgpu/gc_11_0_0_mes_2.bin failed with error -2
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mes_2.bin
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: [drm:jpeg_v4_0_early_init [amdgpu]] JPEG decode is enabled in VM mode
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/vcn_4_0_0.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mec.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_rlc.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_me.bin
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: CP RS64 enable
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_pfp.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/dcn_3_2_0_dmcub.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/smu_13_0_0.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/psp_13_0_0_ta.bin
Jun 21 09:45:16 SuperTux kernel: Loading firmware: amdgpu/psp_13_0_0_sos.bin
Jun 21 09:45:16 SuperTux kernel: amdgpu: ATOM BIOS: 113-4E4710U-T4S
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Fetched VBIOS from VFCT
Jun 21 09:45:16 SuperTux kernel: amdgpu 0000:03:00.0: enabling device (0006 -> 0007)
Jun 21 09:45:16 SuperTux kernel: amdgpu: Topology: Add CPU node
Jun 21 09:45:16 SuperTux kernel: amdgpu: Virtual CRAT table created for CPU
Jun 21 09:45:16 SuperTux kernel: amdgpu: CRAT table not found
Jun 21 09:45:16 SuperTux kernel: [drm] amdgpu kernel modesetting enabled.

kernel 6.1

  
Jun 21 09:50:29 SuperTux kernel: amdgpu 0000:03:00.0: [drm] fb0: amdgpudrmfb frame buffer device
Jun 21 09:50:29 SuperTux kernel: fbcon: amdgpudrmfb (fb0) is primary device
Jun 21 09:50:28 SuperTux kernel: [drm] Initialized amdgpu 3.49.0 20150101 for 0000:03:00.0 on minor 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Using BACO for runtime pm
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring mes_kiq_3.1.0 uses VM inv eng 14 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring jpeg_dec uses VM inv eng 4 on hub 1
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring vcn_unified_1 uses VM inv eng 1 on hub 1
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring vcn_unified_0 uses VM inv eng 0 on hub 1
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 11 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 10 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 9 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 8 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 7 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 6 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SE 6, SH per SE 2, CU per SH 8, active_cu_number 96
Jun 21 09:50:28 SuperTux kernel: kfd kfd: amdgpu: added device 1002:744c
Jun 21 09:50:28 SuperTux kernel: amdgpu: Topology: Add dGPU node [0x744c:0x1002]
Jun 21 09:50:28 SuperTux kernel: amdgpu: Virtual CRAT table created for GPU
Jun 21 09:50:28 SuperTux kernel: amdgpu: SRAT table not found
Jun 21 09:50:28 SuperTux kernel: amdgpu: HMM registered 24560MB device memory
Jun 21 09:50:28 SuperTux kernel: amdgpu: sdma_bitmap: fff0
Jun 21 09:50:28 SuperTux kernel: kfd kfd: amdgpu: Allocated 3969056 bytes on gart
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: [drm:jpeg_v4_0_hw_init [amdgpu]] JPEG decode initialized successfully.
Jun 21 09:50:28 SuperTux kernel: snd_hda_intel 0000:03:00.1: bound 0000:03:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SMU is initialized successfully!
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SMU driver if version not matched
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: smu driver if version = 0x00000037, smu fw if version = 0x00000034, smu fw program = 0, smu fw version = 0x004e4b00 (78.75.0)
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: RAP: optional rap ta ucode is not available
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mes1.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mes.bin
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Will use PSP to load VCN firmware
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/vcn_4_0_0.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/sdma_6_0_0.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_mec.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_rlc.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_me.bin
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: CP RS64 enable
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_pfp.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/gc_11_0_0_imu.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/dcn_3_2_0_dmcub.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/smu_13_0_0.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/psp_13_0_0_ta.bin
Jun 21 09:50:28 SuperTux kernel: Loading firmware: amdgpu/psp_13_0_0_sos.bin
Jun 21 09:50:28 SuperTux kernel: [drm] amdgpu: 32021M of GTT memory ready.
Jun 21 09:50:28 SuperTux kernel: [drm] amdgpu: 24560M of VRAM memory ready
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: GART: 512M 0x0000000000000000 - 0x000000001FFFFFFF
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: VRAM: 24560M 0x0000008000000000 - 0x00000085FEFFFFFF (24560M used)
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: SRAM ECC is not presented.
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: MEM ECC is not presented.
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: vgaarb: deactivate vga console
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: [drm:jpeg_v4_0_early_init [amdgpu]] JPEG decode is enabled in VM mode
Jun 21 09:50:28 SuperTux kernel: amdgpu: ATOM BIOS: 113-4E4710U-T4S
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: amdgpu: Fetched VBIOS from VFCT
Jun 21 09:50:28 SuperTux kernel: amdgpu 0000:03:00.0: enabling device (0006 -> 0007)
Jun 21 09:50:28 SuperTux kernel: amdgpu: Topology: Add CPU node
Jun 21 09:50:28 SuperTux kernel: amdgpu: Virtual CRAT table created for CPU
Jun 21 09:50:28 SuperTux kernel: amdgpu: CRAT table not found
Jun 21 09:50:28 SuperTux kernel: [drm] amdgpu kernel modesetting enabled.

Last edited by MikeAc (2023-06-21 16:46:05)

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

neither linux-lts (5.1) nor linux-firmware-git off aur was able to fix the performance downgrade, nor did downgrading all packages back to june 1. Might just reinstall at this point although what manner of hardware issue would cause this to happen exclusively on linux? Upon further investigation corectrl is able to get the clock running past 500mhz in some scenarios, but the gpu won't go past a roughly 80w power limit which still can't be changed due to the complete lack of the power limit setting mentioned in the gitlab issue. What kind of hardware issue would even cause power limit to change on one OS? (also fast boot was off well before the error).

Last edited by cncz (2023-06-21 18:04:30)

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

Weirder still, corectrl only lets me change the max clock when amdgpu.ppfeaturemask=0xffffffff isn't on the kernel options, otherwise it shows no power limit change and no clock change

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 25,194

Re: [SOLVED] 7900xt sudden performance drop

Is fast boot disabled in Windows? Double check this and reboot a few times, this is absolutely necessary and can explain all sorts of strange behaviours in linux in comparison to Windows.

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

fast boot is 100% off, I disable it first thing when I have to install windows so it's also been off for a few months now, I'm going to try linux-mainline as a last-ditch effort since 6.4 has a few RDNA3 fixes, otherwise I guess I'll just do a full reinstall. Any decent guides on how to reinstall while preserving a btrfs home subvolume?

Last edited by cncz (2023-06-21 18:40:13)

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

I guess I'll also try to run a few reboots out of windows, maybe enable fast boot and disable it again for good measure.

MikeAc

Member

Registered: 2023-06-21

Posts: 2

Re: [SOLVED] 7900xt sudden performance drop

I made a mistake I meant 6.1. did you tried kernel 6.1?

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

I tried linux-lts, which uses kernel 6.1, do you install it some other way?

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

Fixed it by installing linux-amd-drm-next from aur, so I guess whatever bug this is will be fixed when those fixes get merged in 6.5

Last edited by cncz (2023-06-22 00:33:53)

kostjanix

Member

Registered: 2023-07-03

Posts: 5

Re: [SOLVED] 7900xt sudden performance drop

Thank you for posting this. I had a similar issue.

Basically all 3d-accelerated applications (so, games basically) that can utilize the GPU to 100% are suddenly capped to around 20-25 FPS (according to mangohud and/or ingame benchmarks). i tested e.g. shadow of the tombraider (steam/flatpak) and ELEX2 (dx12-version, steam/flatpak). One thing that caught my attention, was, that it seems like, that the GPU is not utilized by 100%, but there seems to be a powercap of 175W in place somehow (7900xtx has ~300W or so  as cap usually). the bug must have been introduced after Kernel `6.3.8.arch1-1 ` which ran fine until yesterday. currently, on 6.4.1.arch1-1 it is broken.

my setup:
ryzen 9 7900x
amd rx 7900xtx

i am currently  installing the suggested kernel, and will report back the result.

Last edited by kostjanix (2023-07-03 18:17:18)

kostjanix

Member

Registered: 2023-07-03

Posts: 5

Re: [SOLVED] 7900xt sudden performance drop

quick update:
linux-amd-drm-next did NOT fixed it for me. however, linux-lts does, so this is indeed something related to the kernel itself.

confirmed working:
6.1.37-1-lts
6.3.8.arch1-1

confirmed not working:
6.4.1.arch1-1
6.4.0-rc4-1-amd-drm-next-g663b930e2484

...so i guess i'll use the LTS kernel then until 6.4 is fixed....

cubethethird

Member

Registered: 2016-01-25

Posts: 99

Website

Re: [SOLVED] 7900xt sudden performance drop

Hey, I just upgraded to an AMD gpu only to discover this issue as well!

I did find a sort of workaround though? For me, when I suspend and wakeup my system, the VRAM speed gets set to the max 1000MHz. Unfortunately it seems to stay at that speed, but at least the performance is better.

ewaller

Administrator

From: Pasadena, CA

Registered: 2009-07-13

Posts: 20,634

Re: [SOLVED] 7900xt sudden performance drop

<Soapbox>
Grrr. 

500mhz and power usage hovering at around 80w.

Capitalization matters.  It is Hz and W.  And, I am reasonably sure you meant MHz, not mHz.  It is a six order of magnitude difference..
</Soapbox>

---

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
The shortest way to ruin a country is to give power to demagogues.— Dionysius of Halicarnassus
---
How to Ask Questions the Smart Way

kostjanix

Member

Registered: 2023-07-03

Posts: 5

Re: [SOLVED] 7900xt sudden performance drop

This issue is still not fixed within 6.4.3-arch1-1 as it seems. This is a very annoying bug.

kostjanix

Member

Registered: 2023-07-03

Posts: 5

Re: [SOLVED] 7900xt sudden performance drop

it very much seems, as if this here is the root-cause:
https://gitlab.freedesktop.org/drm/amd/-/issues/2657

loqs

Member

Registered: 2014-03-06

Posts: 18,884

Re: [SOLVED] 7900xt sudden performance drop

it very much seems, as if this here is the root-cause:
https://gitlab.freedesktop.org/drm/amd/-/issues/2657

You can test it is the same issue with the kernel from https://bbs.archlinux.org/viewtopic.php … 6#p2108566

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

Update, the issue no longer appears on default kernels, although according to corectrl the memory is stuck at 1249mhz on both kernels but I've yet to determine if that's just a reporting issue considering I don't notice a significant performance downgrade.

Last edited by cncz (2023-07-20 16:06:56)

loqs

Member

Registered: 2014-03-06

Posts: 18,884

Re: [SOLVED] 7900xt sudden performance drop

If you consider the issue solved please edit your first post and prepend [SOLVED] to the title.

cncz

Member

Registered: 2021-05-07

Posts: 11

Re: [SOLVED] 7900xt sudden performance drop

nevermind, 1249mhz seems to be normal since linux reports half the advertised vram clocks for everything usually. Still not sure if I should mark this solved yet considering other people in the thread still have this issue.

loqs

Member

Registered: 2014-03-06

Posts: 18,884

Re: [SOLVED] 7900xt sudden performance drop

Still not sure if I should mark this solved yet considering other people in the thread still have this issue.

If other people have the same issue it will be resolved by the same fix.  If they still have an issue that needs a different fix then it should have its own thread,  upstream bug report e.t.c..