github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-24 06:31:50 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #12708] failure during GPU discovery (AMD 7900 XT) using systemd, but OK from command line #70493

New Issue
Closed
opened 2026-05-04 21:44:14 -05:00 by GiteaMirror · 7 comments
GiteaMirror commented 2026-05-04 21:44:14 -05:00
Owner
Copy Link

Originally created by @papoteur-mga on GitHub (Oct 20, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/12708

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

The journal reports a coredump when starting ollama.service with systemd, but launching ollama serve from command line is OK.
The service is still running, but not responding.

Relevant log output

● ollama.service - Ollama Service
     Loaded: loaded (/etc/systemd/system/ollama.service; enabled; preset: disabled)
     Active: active (running) since Mon 2025-10-20 15:10:19 CEST; 3h 23min ago
 Invocation: 0632f28338c84128b899813cbfc86927
   Main PID: 1666 (ollama)
      Tasks: 12 (limit: 36986)
     Memory: 512.2M (peak: 1.4G)
        CPU: 4.757s
     CGroup: /system.slice/ollama.service
             └─1666 /usr/local/bin/ollama serve

oct. 20 15:10:19 centre-ai systemd[1]: Started Ollama Service.
oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.022+02:00 level=INFO source=routes.go:1511 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://192.168.1.14:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/usr/share/ollama/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[* http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.024+02:00 level=INFO source=images.go:522 msg="total blobs: 5"
oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.024+02:00 level=INFO source=images.go:529 msg="total unused blobs removed: 0"
oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.024+02:00 level=INFO source=routes.go:1564 msg="Listening on 192.168.1.14:11434 (version 0.12.6)"
oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.040+02:00 level=INFO source=runner.go:80 msg="discovering available GPUs..."
oct. 20 15:10:23 centre-ai systemd-coredump[2386]: Process 2168 (ollama) of user 983 dumped core.
                                                   
                                                   Stack trace of thread 2168:
                                                   #0  0x00007fbaf90a2ecc n/a (n/a + 0x0)
                                                   ELF object binary architecture: AMD x86-64
oct. 20 15:10:23 centre-ai ollama[1666]: time=2025-10-20T15:10:23.366+02:00 level=INFO source=runner.go:545 msg="failure during GPU discovery" OLLAMA_LIBRARY_PATH="[/usr/local/lib/ollama /usr/local/lib/ollama/rocm]" extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=1]" error="runner crashed"
oct. 20 15:10:24 centre-ai ollama[1666]: time=2025-10-20T15:10:24.071+02:00 level=INFO source=types.go:112 msg="inference compute" id=GPU-d4a094a34a2037e3 library=ROCm compute=gfx1100 name=ROCm0 description="AMD Radeon Graphics" libdirs=ollama,rocm driver=60342.13 pci_id=03:00.0 type=discrete total="20.0 GiB" available="19.9 GiB"
oct. 20 15:10:24 centre-ai ollama[1666]: time=2025-10-20T15:10:24.071+02:00 level=INFO source=routes.go:1605 msg="entering low vram mode" "total vram"="20.0 GiB" threshold="20.0 GiB"


User `ollama` belongs to groups `video` and `render`

OS

Linux

GPU

AMD

CPU

AMD

Ollama version

0.12.6

Originally created by @papoteur-mga on GitHub (Oct 20, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/12708 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? The journal reports a coredump when starting ollama.service with systemd, but launching ollama serve from command line is OK. The service is still running, but not responding. ### Relevant log output ```shell ● ollama.service - Ollama Service Loaded: loaded (/etc/systemd/system/ollama.service; enabled; preset: disabled) Active: active (running) since Mon 2025-10-20 15:10:19 CEST; 3h 23min ago Invocation: 0632f28338c84128b899813cbfc86927 Main PID: 1666 (ollama) Tasks: 12 (limit: 36986) Memory: 512.2M (peak: 1.4G) CPU: 4.757s CGroup: /system.slice/ollama.service └─1666 /usr/local/bin/ollama serve oct. 20 15:10:19 centre-ai systemd[1]: Started Ollama Service. oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.022+02:00 level=INFO source=routes.go:1511 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://192.168.1.14:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/usr/share/ollama/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[* http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]" oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.024+02:00 level=INFO source=images.go:522 msg="total blobs: 5" oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.024+02:00 level=INFO source=images.go:529 msg="total unused blobs removed: 0" oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.024+02:00 level=INFO source=routes.go:1564 msg="Listening on 192.168.1.14:11434 (version 0.12.6)" oct. 20 15:10:20 centre-ai ollama[1666]: time=2025-10-20T15:10:20.040+02:00 level=INFO source=runner.go:80 msg="discovering available GPUs..." oct. 20 15:10:23 centre-ai systemd-coredump[2386]: Process 2168 (ollama) of user 983 dumped core. Stack trace of thread 2168: #0 0x00007fbaf90a2ecc n/a (n/a + 0x0) ELF object binary architecture: AMD x86-64 oct. 20 15:10:23 centre-ai ollama[1666]: time=2025-10-20T15:10:23.366+02:00 level=INFO source=runner.go:545 msg="failure during GPU discovery" OLLAMA_LIBRARY_PATH="[/usr/local/lib/ollama /usr/local/lib/ollama/rocm]" extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=1]" error="runner crashed" oct. 20 15:10:24 centre-ai ollama[1666]: time=2025-10-20T15:10:24.071+02:00 level=INFO source=types.go:112 msg="inference compute" id=GPU-d4a094a34a2037e3 library=ROCm compute=gfx1100 name=ROCm0 description="AMD Radeon Graphics" libdirs=ollama,rocm driver=60342.13 pci_id=03:00.0 type=discrete total="20.0 GiB" available="19.9 GiB" oct. 20 15:10:24 centre-ai ollama[1666]: time=2025-10-20T15:10:24.071+02:00 level=INFO source=routes.go:1605 msg="entering low vram mode" "total vram"="20.0 GiB" threshold="20.0 GiB" User `ollama` belongs to groups `video` and `render` ``` ### OS Linux ### GPU AMD ### CPU AMD ### Ollama version 0.12.6
GiteaMirror added the amdbuggpu labels 2026-05-04 21:44:15 -05:00
GiteaMirror commented 2026-05-04 21:44:23 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Oct 21, 2025):

Can you run with OLLAMA_DEBUG=2 and share the logs of the startup up to the point of "inference compute" being reported?

<!-- gh-comment-id:3424193725 --> @dhiltgen commented on GitHub (Oct 21, 2025): Can you run with OLLAMA_DEBUG=2 and share the logs of the startup up to the point of "inference compute" being reported?
GiteaMirror commented 2026-05-04 21:44:28 -05:00
Author
Owner
Copy Link

@gaurav712 commented on GitHub (Oct 21, 2025):

I think it's a similar case for me @dhiltgen

OS: Artix Linux
GPU: AMD 6700xt
Ollama: 0.12.6 (Using ollama-rocm from extra repo)

Here's the logs:

time=2025-10-21T15:33:45.612+05:30 level=INFO source=routes.go:1511 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:DEBUG-4 OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/home/gaurav/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
time=2025-10-21T15:33:45.612+05:30 level=INFO source=images.go:522 msg="total blobs: 0"
time=2025-10-21T15:33:45.612+05:30 level=INFO source=images.go:529 msg="total unused blobs removed: 0"
time=2025-10-21T15:33:45.612+05:30 level=INFO source=routes.go:1564 msg="Listening on 127.0.0.1:11434 (version 0.12.6)"
time=2025-10-21T15:33:45.612+05:30 level=DEBUG source=sched.go:123 msg="starting llm scheduler"
time=2025-10-21T15:33:45.612+05:30 level=INFO source=runner.go:80 msg="discovering available GPUs..."
time=2025-10-21T15:33:45.612+05:30 level=DEBUG source=runner.go:448 msg="spawning runner with" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs=[]
time=2025-10-21T15:33:45.613+05:30 level=TRACE source=runner.go:529 msg="starting runner for device discovery" env="[OLLAMA_DEBUG=2 SHELL=/bin/bash SESSION_MANAGER=local/artix:@/tmp/.ICE-unix/1549,unix/artix:/tmp/.ICE-unix/1549 WINDOWID=93839753915440 COLORTERM=truecolor XDG_CONFIG_DIRS=/home/gaurav/.config/kdedefaults:/etc/xdg XDG_SESSION_PATH=/org/freedesktop/DisplayManager/Session1 XDG_MENU_PREFIX=plasma- TERM_PROGRAM_VERSION=3.5a TMUX=/tmp/tmux-1000/default,2016,0 ICEAUTHORITY=/run/user/1000/iceauth_EFeUCY DESKTOP_SESSION=plasma GTK_RC_FILES=/etc/gtk/gtkrc:/home/gaurav/.gtkrc:/home/gaurav/.config/gtkrc EDITOR=nvim is_vim=ps -o state= -o comm= -t '#{pane_tty}'     | grep -iqE '^[^TXZ ]+ +(\\S+\\/)?g?(view|n?vim?x?)(diff)?$' XDG_SEAT=seat0 PWD=/home/gaurav LOGNAME=gaurav XDG_SESSION_TYPE=wayland XAUTHORITY=/run/user/1000/xauth_CkPlWl MOTD_SHOWN=pam GTK2_RC_FILES=/etc/gtk-2.0/gtkrc:/home/gaurav/.gtkrc-2.0:/home/gaurav/.config/gtkrc-2.0 HOME=/home/gaurav LANG=C.UTF-8 XDG_CURRENT_DESKTOP=KDE WAYLAND_DISPLAY=wayland-0 XDG_SEAT_PATH=/org/freedesktop/DisplayManager/Seat0 KDE_SESSION_UID=1000 ALACRITTY_SOCKET=/run/user/1000/Alacritty-wayland-0-1990.sock XDG_SESSION_CLASS=user TERM=tmux-256color USER=gaurav TMUX_PANE=%0 QT_WAYLAND_RECONNECT=1 KDE_SESSION_VERSION=6 PAM_KWALLET5_LOGIN=/run/user/1000/kwallet5.socket DISPLAY=:1 SHLVL=1 XDG_VTNR=8 XDG_SESSION_ID=1 ROCM_PATH=/opt/rocm XDG_RUNTIME_DIR=/run/user/1000 DEBUGINFOD_URLS=https://debuginfod.artixlinux.org  XDG_DATA_DIRS=/home/gaurav/.local/share/flatpak/exports/share:/var/lib/flatpak/exports/share:/usr/local/share:/usr/share KDE_FULL_SESSION=true PATH=/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/bin/site_perl:/usr/bin/vendor_perl:/usr/bin/core_perl:/opt/rocm/bin:/usr/lib/rustup/bin ALACRITTY_LOG=/tmp/Alacritty-1990.log DBUS_SESSION_BUS_ADDRESS=unix:path=/tmp/dbus-ASfGKgwaBG,guid=dc85cac501ae6645feed705468f75642 KDE_APPLICATIONS_AS_SCOPE=1 tmux_version=$(tmux -V | sed -En \"s/^tmux ([0-9]+(.[0-9]+)?).*/\\1/p\") MAIL=/var/spool/mail/gaurav ALACRITTY_WINDOW_ID=93839753915440 TERM_PROGRAM=tmux _=/usr/bin/ollama LD_LIBRARY_PATH=/usr/lib/ollama:/usr/lib/ollama OLLAMA_LIBRARY_PATH=/usr/lib/ollama]" cmd="/usr/bin/ollama runner --ollama-engine --port 37907"
time=2025-10-21T15:33:45.619+05:30 level=INFO source=runner.go:1332 msg="starting ollama engine"
time=2025-10-21T15:33:45.619+05:30 level=INFO source=runner.go:1367 msg="Server listening on 127.0.0.1:37907"
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=gguf.go:578 msg=general.architecture type=string
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=gguf.go:578 msg=tokenizer.ggml.model type=string
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.file_type default=0
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.name default=""
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.description default=""
time=2025-10-21T15:33:45.623+05:30 level=INFO source=ggml.go:134 msg="" architecture=llama file_type=unknown name="" description="" num_tensors=0 num_key_values=3
time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:94 msg="ggml backend load all from path" path=/usr/lib/ollama
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
  Device 0: AMD Radeon RX 6700 XT, gfx1031 (0x1031), VMM: no, Wave Size: 32, ID: 0
load_backend: loaded ROCm backend from /usr/lib/ollama/libggml-hip.so
load_backend: loaded CPU backend from /usr/lib/ollama/libggml-cpu-haswell.so
time=2025-10-21T15:33:51.681+05:30 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 ROCm.0.NO_VMM=1 ROCm.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc)
time=2025-10-21T15:33:51.681+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.block_count default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.pooling_type default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.expert_count default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.tokens default="&{size:0 values:[]}"
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.scores default="&{size:0 values:[]}"
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.token_type default="&{size:0 values:[]}"
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.merges default="&{size:0 values:[]}"
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.add_bos_token default=true
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.bos_token_id default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.add_eos_token default=false
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.eos_token_id default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.eos_token_ids default="&{size:0 values:[]}"
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.pre default=""
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.block_count default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.embedding_length default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.head_count default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.head_count_kv default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.key_length default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.rope.dimension_count default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.layer_norm_rms_epsilon default=0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.rope.freq_base default=100000
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.rope.scaling.factor default=1
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:1307 msg="dummy model load took" duration=6.058623955s
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:1312 msg="gathering device infos took" duration=20.19µs
time=2025-10-21T15:33:51.682+05:30 level=TRACE source=runner.go:548 msg="runner enumerated devices" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] devices="[{DeviceID:{ID:0 Library:ROCm} Name:ROCm0 Description:AMD Radeon RX 6700 XT FilteredID: Integrated:false PCIID:0d:00.0 TotalMemory:12868124672 FreeMemory:12826181632 ComputeMajor:16 ComputeMinor:49 DriverMajor:60443 DriverMinor:48 LibraryPath:[/usr/lib/ollama]}]"
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:451 msg="bootstrap discovery took" duration=6.069713727s OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs=[]
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:118 msg="filtering out unsupported or overlapping GPU library combinations" count=1
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:130 msg="verifying GPU is supported" library=/usr/lib/ollama description="AMD Radeon RX 6700 XT" compute=gfx1031 pci_id=0d:00.0
time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:448 msg="spawning runner with" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]"
time=2025-10-21T15:33:51.682+05:30 level=TRACE source=runner.go:529 msg="starting runner for device discovery" env="[OLLAMA_DEBUG=2 SHELL=/bin/bash SESSION_MANAGER=local/artix:@/tmp/.ICE-unix/1549,unix/artix:/tmp/.ICE-unix/1549 WINDOWID=93839753915440 COLORTERM=truecolor XDG_CONFIG_DIRS=/home/gaurav/.config/kdedefaults:/etc/xdg XDG_SESSION_PATH=/org/freedesktop/DisplayManager/Session1 XDG_MENU_PREFIX=plasma- TERM_PROGRAM_VERSION=3.5a TMUX=/tmp/tmux-1000/default,2016,0 ICEAUTHORITY=/run/user/1000/iceauth_EFeUCY DESKTOP_SESSION=plasma GTK_RC_FILES=/etc/gtk/gtkrc:/home/gaurav/.gtkrc:/home/gaurav/.config/gtkrc EDITOR=nvim is_vim=ps -o state= -o comm= -t '#{pane_tty}'     | grep -iqE '^[^TXZ ]+ +(\\S+\\/)?g?(view|n?vim?x?)(diff)?$' XDG_SEAT=seat0 PWD=/home/gaurav LOGNAME=gaurav XDG_SESSION_TYPE=wayland XAUTHORITY=/run/user/1000/xauth_CkPlWl MOTD_SHOWN=pam GTK2_RC_FILES=/etc/gtk-2.0/gtkrc:/home/gaurav/.gtkrc-2.0:/home/gaurav/.config/gtkrc-2.0 HOME=/home/gaurav LANG=C.UTF-8 XDG_CURRENT_DESKTOP=KDE WAYLAND_DISPLAY=wayland-0 XDG_SEAT_PATH=/org/freedesktop/DisplayManager/Seat0 KDE_SESSION_UID=1000 ALACRITTY_SOCKET=/run/user/1000/Alacritty-wayland-0-1990.sock XDG_SESSION_CLASS=user TERM=tmux-256color USER=gaurav TMUX_PANE=%0 QT_WAYLAND_RECONNECT=1 KDE_SESSION_VERSION=6 PAM_KWALLET5_LOGIN=/run/user/1000/kwallet5.socket DISPLAY=:1 SHLVL=1 XDG_VTNR=8 XDG_SESSION_ID=1 ROCM_PATH=/opt/rocm XDG_RUNTIME_DIR=/run/user/1000 DEBUGINFOD_URLS=https://debuginfod.artixlinux.org  XDG_DATA_DIRS=/home/gaurav/.local/share/flatpak/exports/share:/var/lib/flatpak/exports/share:/usr/local/share:/usr/share KDE_FULL_SESSION=true PATH=/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/bin/site_perl:/usr/bin/vendor_perl:/usr/bin/core_perl:/opt/rocm/bin:/usr/lib/rustup/bin ALACRITTY_LOG=/tmp/Alacritty-1990.log DBUS_SESSION_BUS_ADDRESS=unix:path=/tmp/dbus-ASfGKgwaBG,guid=dc85cac501ae6645feed705468f75642 KDE_APPLICATIONS_AS_SCOPE=1 tmux_version=$(tmux -V | sed -En \"s/^tmux ([0-9]+(.[0-9]+)?).*/\\1/p\") MAIL=/var/spool/mail/gaurav ALACRITTY_WINDOW_ID=93839753915440 TERM_PROGRAM=tmux _=/usr/bin/ollama LD_LIBRARY_PATH=/usr/lib/ollama:/usr/lib/ollama OLLAMA_LIBRARY_PATH=/usr/lib/ollama GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]" cmd="/usr/bin/ollama runner --ollama-engine --port 42403"
time=2025-10-21T15:33:51.690+05:30 level=INFO source=runner.go:1332 msg="starting ollama engine"
time=2025-10-21T15:33:51.690+05:30 level=INFO source=runner.go:1367 msg="Server listening on 127.0.0.1:42403"
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=gguf.go:578 msg=general.architecture type=string
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=gguf.go:578 msg=tokenizer.ggml.model type=string
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.file_type default=0
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.name default=""
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.description default=""
time=2025-10-21T15:33:51.693+05:30 level=INFO source=ggml.go:134 msg="" architecture=llama file_type=unknown name="" description="" num_tensors=0 num_key_values=3
time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:94 msg="ggml backend load all from path" path=/usr/lib/ollama
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
ggml_cuda_init: initializing rocBLAS on device 0

rocBLAS error: Cannot read /opt/rocm/lib/rocblas/library/TensileLibrary.dat: Illegal seek for GPU arch : gfx1031
 List of available TensileLibrary Files :
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1010.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1012.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1030.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1101.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1100.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1150.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1151.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1200.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1201.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx908.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx906.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx90a.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1102.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx900.dat"
"/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx942.dat"
time=2025-10-21T15:33:57.731+05:30 level=INFO source=runner.go:545 msg="failure during GPU discovery" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]" error="runner crashed"
time=2025-10-21T15:33:57.731+05:30 level=TRACE source=runner.go:548 msg="runner enumerated devices" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] devices=[]
time=2025-10-21T15:33:57.731+05:30 level=DEBUG source=runner.go:451 msg="bootstrap discovery took" duration=6.048545334s OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]"
time=2025-10-21T15:33:57.731+05:30 level=TRACE source=runner.go:171 msg="supported GPU library combinations" supported=map[]
time=2025-10-21T15:33:57.731+05:30 level=TRACE source=runner.go:182 msg="removing unsupported or overlapping GPU combination" libDir=/usr/lib/ollama description="AMD Radeon RX 6700 XT" compute=gfx1031 pci_id=0d:00.0
time=2025-10-21T15:33:57.731+05:30 level=DEBUG source=runner.go:45 msg="GPU bootstrap discovery took" duration=12.118452761s
time=2025-10-21T15:33:57.731+05:30 level=INFO source=types.go:129 msg="inference compute" id=cpu library=cpu compute="" name=cpu description=cpu libdirs=ollama driver="" pci_id="" type="" total="15.5 GiB" available="12.5 GiB"
time=2025-10-21T15:33:57.731+05:30 level=INFO source=routes.go:1605 msg="entering low vram mode" "total vram"="0 B" threshold="20.0 GiB"
<!-- gh-comment-id:3425773753 --> @gaurav712 commented on GitHub (Oct 21, 2025): I think it's a similar case for me @dhiltgen OS: Artix Linux GPU: AMD 6700xt Ollama: 0.12.6 (Using `ollama-rocm` from `extra` repo) Here's the logs: ``` time=2025-10-21T15:33:45.612+05:30 level=INFO source=routes.go:1511 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:DEBUG-4 OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/home/gaurav/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]" time=2025-10-21T15:33:45.612+05:30 level=INFO source=images.go:522 msg="total blobs: 0" time=2025-10-21T15:33:45.612+05:30 level=INFO source=images.go:529 msg="total unused blobs removed: 0" time=2025-10-21T15:33:45.612+05:30 level=INFO source=routes.go:1564 msg="Listening on 127.0.0.1:11434 (version 0.12.6)" time=2025-10-21T15:33:45.612+05:30 level=DEBUG source=sched.go:123 msg="starting llm scheduler" time=2025-10-21T15:33:45.612+05:30 level=INFO source=runner.go:80 msg="discovering available GPUs..." time=2025-10-21T15:33:45.612+05:30 level=DEBUG source=runner.go:448 msg="spawning runner with" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs=[] time=2025-10-21T15:33:45.613+05:30 level=TRACE source=runner.go:529 msg="starting runner for device discovery" env="[OLLAMA_DEBUG=2 SHELL=/bin/bash SESSION_MANAGER=local/artix:@/tmp/.ICE-unix/1549,unix/artix:/tmp/.ICE-unix/1549 WINDOWID=93839753915440 COLORTERM=truecolor XDG_CONFIG_DIRS=/home/gaurav/.config/kdedefaults:/etc/xdg XDG_SESSION_PATH=/org/freedesktop/DisplayManager/Session1 XDG_MENU_PREFIX=plasma- TERM_PROGRAM_VERSION=3.5a TMUX=/tmp/tmux-1000/default,2016,0 ICEAUTHORITY=/run/user/1000/iceauth_EFeUCY DESKTOP_SESSION=plasma GTK_RC_FILES=/etc/gtk/gtkrc:/home/gaurav/.gtkrc:/home/gaurav/.config/gtkrc EDITOR=nvim is_vim=ps -o state= -o comm= -t '#{pane_tty}' | grep -iqE '^[^TXZ ]+ +(\\S+\\/)?g?(view|n?vim?x?)(diff)?$' XDG_SEAT=seat0 PWD=/home/gaurav LOGNAME=gaurav XDG_SESSION_TYPE=wayland XAUTHORITY=/run/user/1000/xauth_CkPlWl MOTD_SHOWN=pam GTK2_RC_FILES=/etc/gtk-2.0/gtkrc:/home/gaurav/.gtkrc-2.0:/home/gaurav/.config/gtkrc-2.0 HOME=/home/gaurav LANG=C.UTF-8 XDG_CURRENT_DESKTOP=KDE WAYLAND_DISPLAY=wayland-0 XDG_SEAT_PATH=/org/freedesktop/DisplayManager/Seat0 KDE_SESSION_UID=1000 ALACRITTY_SOCKET=/run/user/1000/Alacritty-wayland-0-1990.sock XDG_SESSION_CLASS=user TERM=tmux-256color USER=gaurav TMUX_PANE=%0 QT_WAYLAND_RECONNECT=1 KDE_SESSION_VERSION=6 PAM_KWALLET5_LOGIN=/run/user/1000/kwallet5.socket DISPLAY=:1 SHLVL=1 XDG_VTNR=8 XDG_SESSION_ID=1 ROCM_PATH=/opt/rocm XDG_RUNTIME_DIR=/run/user/1000 DEBUGINFOD_URLS=https://debuginfod.artixlinux.org XDG_DATA_DIRS=/home/gaurav/.local/share/flatpak/exports/share:/var/lib/flatpak/exports/share:/usr/local/share:/usr/share KDE_FULL_SESSION=true PATH=/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/bin/site_perl:/usr/bin/vendor_perl:/usr/bin/core_perl:/opt/rocm/bin:/usr/lib/rustup/bin ALACRITTY_LOG=/tmp/Alacritty-1990.log DBUS_SESSION_BUS_ADDRESS=unix:path=/tmp/dbus-ASfGKgwaBG,guid=dc85cac501ae6645feed705468f75642 KDE_APPLICATIONS_AS_SCOPE=1 tmux_version=$(tmux -V | sed -En \"s/^tmux ([0-9]+(.[0-9]+)?).*/\\1/p\") MAIL=/var/spool/mail/gaurav ALACRITTY_WINDOW_ID=93839753915440 TERM_PROGRAM=tmux _=/usr/bin/ollama LD_LIBRARY_PATH=/usr/lib/ollama:/usr/lib/ollama OLLAMA_LIBRARY_PATH=/usr/lib/ollama]" cmd="/usr/bin/ollama runner --ollama-engine --port 37907" time=2025-10-21T15:33:45.619+05:30 level=INFO source=runner.go:1332 msg="starting ollama engine" time=2025-10-21T15:33:45.619+05:30 level=INFO source=runner.go:1367 msg="Server listening on 127.0.0.1:37907" time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32 time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=gguf.go:578 msg=general.architecture type=string time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=gguf.go:578 msg=tokenizer.ggml.model type=string time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32 time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.file_type default=0 time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.name default="" time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.description default="" time=2025-10-21T15:33:45.623+05:30 level=INFO source=ggml.go:134 msg="" architecture=llama file_type=unknown name="" description="" num_tensors=0 num_key_values=3 time=2025-10-21T15:33:45.623+05:30 level=DEBUG source=ggml.go:94 msg="ggml backend load all from path" path=/usr/lib/ollama ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 ROCm devices: Device 0: AMD Radeon RX 6700 XT, gfx1031 (0x1031), VMM: no, Wave Size: 32, ID: 0 load_backend: loaded ROCm backend from /usr/lib/ollama/libggml-hip.so load_backend: loaded CPU backend from /usr/lib/ollama/libggml-cpu-haswell.so time=2025-10-21T15:33:51.681+05:30 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 ROCm.0.NO_VMM=1 ROCm.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc) time=2025-10-21T15:33:51.681+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.block_count default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.pooling_type default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.expert_count default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.tokens default="&{size:0 values:[]}" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.scores default="&{size:0 values:[]}" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.token_type default="&{size:0 values:[]}" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.merges default="&{size:0 values:[]}" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.add_bos_token default=true time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.bos_token_id default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.add_eos_token default=false time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.eos_token_id default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.eos_token_ids default="&{size:0 values:[]}" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=tokenizer.ggml.pre default="" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.block_count default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.embedding_length default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.head_count default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.head_count_kv default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.key_length default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.rope.dimension_count default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.attention.layer_norm_rms_epsilon default=0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.rope.freq_base default=100000 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=llama.rope.scaling.factor default=1 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:1307 msg="dummy model load took" duration=6.058623955s time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:1312 msg="gathering device infos took" duration=20.19µs time=2025-10-21T15:33:51.682+05:30 level=TRACE source=runner.go:548 msg="runner enumerated devices" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] devices="[{DeviceID:{ID:0 Library:ROCm} Name:ROCm0 Description:AMD Radeon RX 6700 XT FilteredID: Integrated:false PCIID:0d:00.0 TotalMemory:12868124672 FreeMemory:12826181632 ComputeMajor:16 ComputeMinor:49 DriverMajor:60443 DriverMinor:48 LibraryPath:[/usr/lib/ollama]}]" time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:451 msg="bootstrap discovery took" duration=6.069713727s OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs=[] time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:118 msg="filtering out unsupported or overlapping GPU library combinations" count=1 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:130 msg="verifying GPU is supported" library=/usr/lib/ollama description="AMD Radeon RX 6700 XT" compute=gfx1031 pci_id=0d:00.0 time=2025-10-21T15:33:51.682+05:30 level=DEBUG source=runner.go:448 msg="spawning runner with" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]" time=2025-10-21T15:33:51.682+05:30 level=TRACE source=runner.go:529 msg="starting runner for device discovery" env="[OLLAMA_DEBUG=2 SHELL=/bin/bash SESSION_MANAGER=local/artix:@/tmp/.ICE-unix/1549,unix/artix:/tmp/.ICE-unix/1549 WINDOWID=93839753915440 COLORTERM=truecolor XDG_CONFIG_DIRS=/home/gaurav/.config/kdedefaults:/etc/xdg XDG_SESSION_PATH=/org/freedesktop/DisplayManager/Session1 XDG_MENU_PREFIX=plasma- TERM_PROGRAM_VERSION=3.5a TMUX=/tmp/tmux-1000/default,2016,0 ICEAUTHORITY=/run/user/1000/iceauth_EFeUCY DESKTOP_SESSION=plasma GTK_RC_FILES=/etc/gtk/gtkrc:/home/gaurav/.gtkrc:/home/gaurav/.config/gtkrc EDITOR=nvim is_vim=ps -o state= -o comm= -t '#{pane_tty}' | grep -iqE '^[^TXZ ]+ +(\\S+\\/)?g?(view|n?vim?x?)(diff)?$' XDG_SEAT=seat0 PWD=/home/gaurav LOGNAME=gaurav XDG_SESSION_TYPE=wayland XAUTHORITY=/run/user/1000/xauth_CkPlWl MOTD_SHOWN=pam GTK2_RC_FILES=/etc/gtk-2.0/gtkrc:/home/gaurav/.gtkrc-2.0:/home/gaurav/.config/gtkrc-2.0 HOME=/home/gaurav LANG=C.UTF-8 XDG_CURRENT_DESKTOP=KDE WAYLAND_DISPLAY=wayland-0 XDG_SEAT_PATH=/org/freedesktop/DisplayManager/Seat0 KDE_SESSION_UID=1000 ALACRITTY_SOCKET=/run/user/1000/Alacritty-wayland-0-1990.sock XDG_SESSION_CLASS=user TERM=tmux-256color USER=gaurav TMUX_PANE=%0 QT_WAYLAND_RECONNECT=1 KDE_SESSION_VERSION=6 PAM_KWALLET5_LOGIN=/run/user/1000/kwallet5.socket DISPLAY=:1 SHLVL=1 XDG_VTNR=8 XDG_SESSION_ID=1 ROCM_PATH=/opt/rocm XDG_RUNTIME_DIR=/run/user/1000 DEBUGINFOD_URLS=https://debuginfod.artixlinux.org XDG_DATA_DIRS=/home/gaurav/.local/share/flatpak/exports/share:/var/lib/flatpak/exports/share:/usr/local/share:/usr/share KDE_FULL_SESSION=true PATH=/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/home/gaurav/Android/Sdk/platform-tools:/home/gaurav/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/bin/site_perl:/usr/bin/vendor_perl:/usr/bin/core_perl:/opt/rocm/bin:/usr/lib/rustup/bin ALACRITTY_LOG=/tmp/Alacritty-1990.log DBUS_SESSION_BUS_ADDRESS=unix:path=/tmp/dbus-ASfGKgwaBG,guid=dc85cac501ae6645feed705468f75642 KDE_APPLICATIONS_AS_SCOPE=1 tmux_version=$(tmux -V | sed -En \"s/^tmux ([0-9]+(.[0-9]+)?).*/\\1/p\") MAIL=/var/spool/mail/gaurav ALACRITTY_WINDOW_ID=93839753915440 TERM_PROGRAM=tmux _=/usr/bin/ollama LD_LIBRARY_PATH=/usr/lib/ollama:/usr/lib/ollama OLLAMA_LIBRARY_PATH=/usr/lib/ollama GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]" cmd="/usr/bin/ollama runner --ollama-engine --port 42403" time=2025-10-21T15:33:51.690+05:30 level=INFO source=runner.go:1332 msg="starting ollama engine" time=2025-10-21T15:33:51.690+05:30 level=INFO source=runner.go:1367 msg="Server listening on 127.0.0.1:42403" time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32 time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=gguf.go:578 msg=general.architecture type=string time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=gguf.go:578 msg=tokenizer.ggml.model type=string time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.alignment default=32 time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.file_type default=0 time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.name default="" time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:276 msg="key with type not found" key=general.description default="" time=2025-10-21T15:33:51.693+05:30 level=INFO source=ggml.go:134 msg="" architecture=llama file_type=unknown name="" description="" num_tensors=0 num_key_values=3 time=2025-10-21T15:33:51.693+05:30 level=DEBUG source=ggml.go:94 msg="ggml backend load all from path" path=/usr/lib/ollama ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 ROCm devices: ggml_cuda_init: initializing rocBLAS on device 0 rocBLAS error: Cannot read /opt/rocm/lib/rocblas/library/TensileLibrary.dat: Illegal seek for GPU arch : gfx1031 List of available TensileLibrary Files : "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1010.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1012.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1030.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1101.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1100.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1150.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1151.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1200.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1201.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx908.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx906.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx90a.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1102.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx900.dat" "/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx942.dat" time=2025-10-21T15:33:57.731+05:30 level=INFO source=runner.go:545 msg="failure during GPU discovery" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]" error="runner crashed" time=2025-10-21T15:33:57.731+05:30 level=TRACE source=runner.go:548 msg="runner enumerated devices" OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] devices=[] time=2025-10-21T15:33:57.731+05:30 level=DEBUG source=runner.go:451 msg="bootstrap discovery took" duration=6.048545334s OLLAMA_LIBRARY_PATH=[/usr/lib/ollama] extra_envs="[GGML_CUDA_INIT=1 ROCR_VISIBLE_DEVICES=0]" time=2025-10-21T15:33:57.731+05:30 level=TRACE source=runner.go:171 msg="supported GPU library combinations" supported=map[] time=2025-10-21T15:33:57.731+05:30 level=TRACE source=runner.go:182 msg="removing unsupported or overlapping GPU combination" libDir=/usr/lib/ollama description="AMD Radeon RX 6700 XT" compute=gfx1031 pci_id=0d:00.0 time=2025-10-21T15:33:57.731+05:30 level=DEBUG source=runner.go:45 msg="GPU bootstrap discovery took" duration=12.118452761s time=2025-10-21T15:33:57.731+05:30 level=INFO source=types.go:129 msg="inference compute" id=cpu library=cpu compute="" name=cpu description=cpu libdirs=ollama driver="" pci_id="" type="" total="15.5 GiB" available="12.5 GiB" time=2025-10-21T15:33:57.731+05:30 level=INFO source=routes.go:1605 msg="entering low vram mode" "total vram"="0 B" threshold="20.0 GiB" ```
GiteaMirror commented 2026-05-04 21:44:32 -05:00
Author
Owner
Copy Link

@papoteur-mga commented on GitHub (Oct 21, 2025):

I have added the DEBUG variable. There is a first attempt without DEBUG, then the restart. At restart, the service was running.

ollama.txt

<!-- gh-comment-id:3427663096 --> @papoteur-mga commented on GitHub (Oct 21, 2025): I have added the DEBUG variable. There is a first attempt without DEBUG, then the restart. At restart, the service was running. [ollama.txt](https://github.com/user-attachments/files/23027073/ollama.txt)
GiteaMirror commented 2026-05-04 21:44:37 -05:00
Author
Owner
Copy Link

@papoteur-mga commented on GitHub (Oct 21, 2025):

I must add that the system has 2 AMD GPUs, the one integrated with the CPU, and the discrete card intended to be used for ollama.

The coredump seems linked to the presence of the basic GPU:

oct. 21 18:20:06 ollama[23862]: rocBLAS error: Cannot read /usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary.dat: Illegal seek for GPU arch : gfx1036
oct. 21 18:20:06 ollama[23862]:  List of available TensileLibrary Files :
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1030.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1151.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1101.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1100.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1012.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx90a.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1102.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1201.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx942.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1010.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1200.dat"
oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx908.dat"
oct. 21 18:20:06 systemd-coredump[23889]: Process 23862 (ollama) of user 983 terminated abnormally with signal 6/ABRT, processing...
oct. 21 18:20:06 systemd[1]: Started Process Core Dump (PID 23889/UID 0).
<!-- gh-comment-id:3427705264 --> @papoteur-mga commented on GitHub (Oct 21, 2025): I must add that the system has 2 AMD GPUs, the one integrated with the CPU, and the discrete card intended to be used for ollama. The coredump seems linked to the presence of the basic GPU: ``` oct. 21 18:20:06 ollama[23862]: rocBLAS error: Cannot read /usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary.dat: Illegal seek for GPU arch : gfx1036 oct. 21 18:20:06 ollama[23862]: List of available TensileLibrary Files : oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1030.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1151.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1101.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1100.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1012.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx90a.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1102.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1201.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx942.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1010.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx1200.dat" oct. 21 18:20:06 ollama[23862]: "/usr/local/lib/ollama/rocm/rocblas/library/TensileLibrary_lazy_gfx908.dat" oct. 21 18:20:06 systemd-coredump[23889]: Process 23862 (ollama) of user 983 terminated abnormally with signal 6/ABRT, processing... oct. 21 18:20:06 systemd[1]: Started Process Core Dump (PID 23889/UID 0). ```
GiteaMirror commented 2026-05-04 21:44:40 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Oct 21, 2025):

@gaurav712 your GPU is a gfx1031 which is not officially supported by ROCm, however gfx1030 is. You'll have to use the HSA_OVERRIDE_GFX_VERSION environment variable to get ROCm to support the GPU on linux - see our docs at https://github.com/ollama/ollama/blob/main/docs/gpu.md#overrides-on-linux

<!-- gh-comment-id:3428694949 --> @dhiltgen commented on GitHub (Oct 21, 2025): @gaurav712 your GPU is a gfx1031 which is not officially supported by ROCm, however gfx1030 is. You'll have to use the HSA_OVERRIDE_GFX_VERSION environment variable to get ROCm to support the GPU on linux - see our docs at https://github.com/ollama/ollama/blob/main/docs/gpu.md#overrides-on-linux
GiteaMirror commented 2026-05-04 21:44:41 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Oct 21, 2025):

@papoteur-mga with our new GPU discovery code, to improve reliability, what we've implemented is an algorithm that forces ROCm to fully initialize on the GPU so we can discover if the GPU is supported or not. Unfortunately, the ROCm APIs abort the program instead of returning an error code on unsupported devices, so the end result is we spawn multiple child processes, and some of them may crash if there are unsupported GPUs in the system. Based on those crashes, we can eliminate GPUs from the set we leverage for inference. It looks like you have 2 AMD GPUs: a gfx1100 and a gfx1036. The gfx1036 is an iGPU.

I believe we do ultimately get the GPU discovered correctly on your system

oct. 21 18:20:07 ollama[23787]: time=2025-10-21T18:20:07.236+02:00 level=INFO source=types.go:112 msg="inference compute" id=GPU-d4a094a34a2037e3 library=ROCm compute=gfx1100 name=ROCm0 description="AMD Radeon Graphics" libdirs=ollama,rocm driver=60342.13 pci_id=03:00.0 type=discrete total="20.0 GiB" available="19.9 GiB"

If you want to avoid the failed discovery of the iGPU, you should be able to set ROCR_VISIBLE_DEVICES=GPU-d4a094a34a2037e3 and that will hide the iGPU so we never probe it to trigger the crash.

<!-- gh-comment-id:3428850722 --> @dhiltgen commented on GitHub (Oct 21, 2025): @papoteur-mga with our new GPU discovery code, to improve reliability, what we've implemented is an algorithm that forces ROCm to fully initialize on the GPU so we can discover if the GPU is supported or not. Unfortunately, the ROCm APIs abort the program instead of returning an error code on unsupported devices, so the end result is we spawn multiple child processes, and some of them may crash if there are unsupported GPUs in the system. Based on those crashes, we can eliminate GPUs from the set we leverage for inference. It looks like you have 2 AMD GPUs: a gfx1100 and a gfx1036. The gfx1036 is an iGPU. I believe we do ultimately get the GPU discovered correctly on your system ``` oct. 21 18:20:07 ollama[23787]: time=2025-10-21T18:20:07.236+02:00 level=INFO source=types.go:112 msg="inference compute" id=GPU-d4a094a34a2037e3 library=ROCm compute=gfx1100 name=ROCm0 description="AMD Radeon Graphics" libdirs=ollama,rocm driver=60342.13 pci_id=03:00.0 type=discrete total="20.0 GiB" available="19.9 GiB" ``` If you want to avoid the failed discovery of the iGPU, you should be able to set `ROCR_VISIBLE_DEVICES=GPU-d4a094a34a2037e3` and that will hide the iGPU so we never probe it to trigger the crash.
GiteaMirror commented 2026-05-04 21:44:41 -05:00
Author
Owner
Copy Link

@papoteur-mga commented on GitHub (Nov 8, 2025):

OK, the reported crash is not a problem for running ollama. This was hiding the real problem.
I just put the last env variable to see if avoid the crash report.
We can close the report

<!-- gh-comment-id:3506312340 --> @papoteur-mga commented on GitHub (Nov 8, 2025): OK, the reported crash is not a problem for running ollama. This was hiding the real problem. I just put the last env variable to see if avoid the crash report. We can close the report
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label amd bug gpu
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#70493
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?