Tuxedo 17 (en Ubuntu 24) + GIGABYTE AORUS RTX 5060 Ti AI Box Carte Graphique - 16GB GDDR7.
-
Nouveau build : ( avec DCMAKE_CUDA_ARCHITECTURES=86 )
# apt install -y nvidia-cuda-toolkit # cd llama.cpp # export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64 # export PATH=$PATH:$CUDA_HOME/bin # cmake -B build -DGGML_CUDA=ON -DCMAKE_CUDA_COMPILER=`which nvcc` -DLLAMA_CURL=ON -DCMAKE_CUDA_ARCHITECTURES=86 # cmake --build build -j$(nproc) -
Info chouchou : https://github.com/NVIDIA/nvbench/discussions/129
So for your example of a 3090, the compute capability is 8.6. So to build both PTX and SASS for that GPU you would specify -DCMAKE_CUDA_ARCHITECTURES=86.
-
Ajout de nvidia-dkms ( pour noyau personnalisé )
# apt-get install nvidia-dkms-595A suivre.
-
Je recommance :
# nvidia-smi Thu Jul 16 15:09:01 2026 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 580.159.04 Driver Version: 580.159.04 CUDA Version: 13.0 | +-----------------------------------------+------------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 NVIDIA GeForce RTX 3060 ... Off | 00000000:01:00.0 Off | N/A | | N/A 40C P0 25W / 115W | 1MiB / 6144MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ +-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | No running processes found | +-----------------------------------------------------------------------------------------+ # apt-get install nvidia-dkms-580 -
Ce message a été supprimé !
-
Uname & dkms
# uname -a Linux tuxedo-17 6.17.0-122035-tuxedo #35~24.04.1tux1 SMP PREEMPT_DYNAMIC Mon Jun 15 14:27:53 UTC 2026 x86_64 x86_64 x86_64 GNU/Linux # dkms status evdi/1.14.14+dfsg, 6.17.0-122035-tuxedo, x86_64: installed nvidia/580.95.05, 6.17.0-122035-tuxedo, x86_64: installed r8125/9.011.00, 6.17.0-122035-tuxedo, x86_64: installed tuxedo-drivers/4.22.3, 6.17.0-122035-tuxedo, x86_64: installed -
Installation de chronium :
# apt-get install chromium-browser -
Sans l’egpu :
# llama-bench -m /models/gemma-3-1b-it-q4_k_m.gguf ggml_cuda_init: found 1 CUDA devices (Total VRAM: 5806 MiB): Device 0: NVIDIA GeForce RTX 3060 Laptop GPU, compute capability 8.6, VMM: yes, VRAM: 5806 MiB | model | size | params | backend | ngl | test | t/s | | ------------------------------ | ---------: | ---------: | ---------- | --: | --------------: | -------------------: | | gemma3 1B Q4_K - Medium | 762.49 MiB | 999.89 M | CUDA | -1 | pp512 | 11293.53 ± 542.50 | | gemma3 1B Q4_K - Medium | 762.49 MiB | 999.89 M | CUDA | -1 | tg128 | 224.54 ± 0.68 | build: c3d47e696 (10030) -
Toujours des crash …
$ dpkg --get-selections | grep -i nvidia libnvidia-cfg1-580:amd64 install libnvidia-common install libnvidia-compute:amd64 install libnvidia-compute:i386 install libnvidia-compute-595:amd64 deinstall libnvidia-decode:amd64 install libnvidia-decode:i386 install libnvidia-egl-wayland1:amd64 install libnvidia-egl-wayland1:i386 install libnvidia-encode:amd64 install libnvidia-encode:i386 install libnvidia-extra:amd64 install libnvidia-fbc1:amd64 install libnvidia-fbc1:i386 install libnvidia-gl:amd64 install libnvidia-gl:i386 install libnvidia-ml-dev:amd64 install nvidia-compute-utils install nvidia-cuda-dev:amd64 install nvidia-cuda-gdb install nvidia-cuda-toolkit install nvidia-cuda-toolkit-doc install nvidia-dkms-580 deinstall nvidia-dkms-open deinstall nvidia-firmware-580 install nvidia-firmware-595-595.71.05 install nvidia-kernel-common deinstall nvidia-kernel-common-580 install nvidia-kernel-source-580 install nvidia-modprobe install nvidia-opencl-dev:amd64 install nvidia-prime install nvidia-profiler install nvidia-settings install nvidia-utils install nvidia-visual-profiler install tuxedo-fix-nvidia-preserve-vram-suspend install tuxedo-nvidia-driver-default-pinning install -
lspci :
03:00.0 PCI bridge: Intel Corporation JHL9480 Thunderbolt 5 80/120G Bridge [Barlow Ridge Hub 80G 2023] (rev 85) (prog-if 00 [Normal decode]) Subsystem: Device 2222:1111 Physical Slot: 0 Control: I/O- Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx- Interrupt: pin A routed to IRQ 16 IOMMU group: 1 Bus: primary=00, secondary=00, subordinate=00, sec-latency=0 I/O behind bridge: 7000-afff [size=16K] [16-bit] Memory behind bridge: 54000000-5bffffff [size=128M] [32-bit] Prefetchable memory behind bridge: 4100000000-413fffffff [size=1G] [32-bit] Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- <SERR- <PERR- BridgeCtl: Parity- SERR- NoISA- VGA- VGA16- MAbort- >Reset- FastB2B- PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn- Capabilities: <access denied> Kernel driver in use: pcieport 04:00.0 PCI bridge: Intel Corporation JHL9480 Thunderbolt 5 80/120G Bridge [Barlow Ridge Hub 80G 2023] (rev 85) (prog-if 00 [Normal decode]) Subsystem: Device 2222:1111 !!! Unknown header type 7f Interrupt: pin ? routed to IRQ 128 IOMMU group: 2 I/O behind bridge: 00007000-00007fff [size=4K] [32-bit] Memory behind bridge: 54000000-59ffffff [size=96M] [32-bit] Prefetchable memory behind bridge: 0000004100000000-000000412fffffff [size=768M] [64-bit] Kernel driver in use: pcieport 04:01.0 PCI bridge: Intel Corporation JHL9480 Thunderbolt 5 80/120G Bridge [Barlow Ridge Hub 80G 2023] (rev 85) (prog-if 00 [Normal decode]) Subsystem: Device 2222:1111 !!! Unknown header type 7f Interrupt: pin ? routed to IRQ 17 IOMMU group: 3 I/O behind bridge: 00008000-00008fff [size=4K] [32-bit] Memory behind bridge: 5a000000-5a1fffff [size=2M] [32-bit] Prefetchable memory behind bridge: 0000004130000000-00000041301fffff [size=2M] [64-bit] 04:02.0 PCI bridge: Intel Corporation JHL9480 Thunderbolt 5 80/120G Bridge [Barlow Ridge Hub 80G 2023] (rev 85) (prog-if 00 [Normal decode]) Subsystem: Device 2222:1111 !!! Unknown header type 7f Interrupt: pin ? routed to IRQ 18 IOMMU group: 4 I/O behind bridge: 00009000-00009fff [size=4K] [32-bit] Memory behind bridge: 5a200000-5a3fffff [size=2M] [32-bit] Prefetchable memory behind bridge: 0000004130200000-00000041303fffff [size=2M] [64-bit] 04:03.0 PCI bridge: Intel Corporation JHL9480 Thunderbolt 5 80/120G Bridge [Barlow Ridge Hub 80G 2023] (rev 85) (prog-if 00 [Normal decode]) Subsystem: Device 2222:1111 !!! Unknown header type 7f Interrupt: pin ? routed to IRQ 19 IOMMU group: 5 I/O behind bridge: 0000a000-0000afff [size=4K] [32-bit] Memory behind bridge: 5a400000-5a5fffff [size=2M] [32-bit] Prefetchable memory behind bridge: 0000004130400000-00000041305fffff [size=2M] [64-bit] 05:00.0 VGA compatible controller: NVIDIA Corporation GB206 [GeForce RTX 5060 Ti] (rev a1) (prog-if 00 [VGA controller]) Subsystem: Gigabyte Technology Co., Ltd Device 41cd !!! Unknown header type 7f Interrupt: pin ? routed to IRQ 16 IOMMU group: 6 Region 0: Memory at 54000000 (32-bit, non-prefetchable) [size=64M] Region 1: Memory at 4100000000 (64-bit, prefetchable) [size=256M] Region 3: Memory at 4120000000 (64-bit, prefetchable) [size=32M] Region 5: I/O ports at 7000 [size=128] Expansion ROM at 58000000 [disabled] [size=512K] Kernel modules: nvidiafb, nouveau, nvidia_drm, nvidia 05:00.1 Audio device: NVIDIA Corporation GB206 High Definition Audio Controller (rev a1) (prog-if 00 [HDA compatible]) Subsystem: NVIDIA Corporation Device 0000 !!! Unknown header type 7f Interrupt: pin ? routed to IRQ 17 IOMMU group: 6 Region 0: Memory at 58080000 (32-bit, non-prefetchable) [size=16K] Kernel modules: snd_hda_intel 2e:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8125 2.5GbE Controller (rev 05) Subsystem: AIstone Global Limited RTL8125 2.5GbE Controller Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+ Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx- Latency: 0, Cache Line Size: 64 bytes Interrupt: pin A routed to IRQ 16 IOMMU group: 24 Region 0: I/O ports at 3000 [size=256] Region 2: Memory at 62200000 (64-bit, non-prefetchable) [size=64K] Region 4: Memory at 62210000 (64-bit, non-prefetchable) [size=16K] Capabilities: <access denied> Kernel driver in use: r8169 Kernel modules: r8169, r8125 -
Nouveau test :
# cat /etc/modprobe.d/thunderbolt.conf options thunderbolt host_reset=false # cat /etc/modprobe.d/nvidia nvidia.conf nvidia-graphics-drivers.conf nvidia-graphics-drivers-kms.conf # cat /etc/modprobe.d/nvidia.conf options nvidia-drm modeset=1 # cat /etc/modprobe.d/nvidia-graphics-drivers.conf options nvidia NVreg_EnableGpuFirmware=0 # cat /etc/modprobe.d/nvidia-graphics-drivers-kms.conf # Nvidia modesetting support. Set to 0 or comment to disable kernel modesetting # and framebuffer console support. This must be disabled in case of Mosaic or SLI. options nvidia-drm modeset=1 -
Lien : https://github.com/NVIDIA/open-gpu-kernel-modules/issues/974#issuecomment-4311518502
# lspci |grep "RTX 5060" 05:00.0 VGA compatible controller: NVIDIA Corporation GB206 [GeForce RTX 5060 Ti] (rev a1) -
Link : https://www.nvidia.com/fr-fr/drivers/details/274213/
Linux x86_64/AMD64/EM64T
Dernière version des pilotes (Production): 595.84
Dernière version des pilotes (Fonctionnalité): 610.43.03
Derniers pilotes Legacy (470.xx série): 470.256.02 -
Nouveau drivers :
# nvidia-smi Thu Jul 16 17:27:40 2026 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 610.43.03 KMD Version: 610.43.03 CUDA UMD Version: 13.3 | +-----------------------------------------+------------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 NVIDIA GeForce RTX 3060 ... Off | 00000000:01:00.0 Off | N/A | | N/A 46C P8 11W / 115W | 1MiB / 6144MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ +-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | No running processes found | +-----------------------------------------------------------------------------------------+ # llama-bench -m /models/gemma-3-1b-it-q4_k_m.gguf ggml_cuda_init: found 1 CUDA devices (Total VRAM: 5803 MiB): Device 0: NVIDIA GeForce RTX 3060 Laptop GPU, compute capability 8.6, VMM: yes, VRAM: 5803 MiB | model | size | params | backend | ngl | test | t/s | | ------------------------------ | ---------: | ---------: | ---------- | --: | --------------: | -------------------: | | gemma3 1B Q4_K - Medium | 762.49 MiB | 999.89 M | CUDA | -1 | pp512 | 11250.23 ± 436.25 | | gemma3 1B Q4_K - Medium | 762.49 MiB | 999.89 M | CUDA | -1 | tg128 | 224.89 ± 0.24 | build: c3d47e696 (10030) -
Branchement de l’egpu :
# nvidia-smi Thu Jul 16 17:29:22 2026 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 610.43.03 KMD Version: 610.43.03 CUDA UMD Version: 13.3 | +-----------------------------------------+------------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 NVIDIA GeForce RTX 3060 ... Off | 00000000:01:00.0 Off | N/A | | N/A 46C P8 11W / 115W | 1MiB / 6144MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ | 1 NVIDIA GeForce RTX 5060 Ti Off | 00000000:05:00.0 Off | N/A | | 0% 34C P0 15W / 180W | 2MiB / 16311MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ +-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | No running processes found | +-----------------------------------------------------------------------------------------+ -
# sudo nvidia-persistenced --user XXXX # nvidia-smi Thu Jul 16 17:42:28 2026 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 610.43.03 KMD Version: 610.43.03 CUDA UMD Version: 13.3 | +-----------------------------------------+------------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 NVIDIA GeForce RTX 3060 ... On | 00000000:01:00.0 Off | N/A | | N/A 42C P8 11W / 115W | 1MiB / 6144MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ | 1 NVIDIA GeForce RTX 5060 Ti On | 00000000:05:00.0 Off | N/A | | 0% 41C P8 5W / 180W | 2MiB / 16311MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ +-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | No running processes found | +-----------------------------------------------------------------------------------------+ -
A noter :
The new nvidia drivers 610.43.02 do not need patching anymore.
Also, nvida_modeset and nvidia_drm can be safely loaded. However, they still need to be loaded after the bridge link speed was set. I believe the bridge link issue needs to be fixed in the kernel.Je change donc :
/etc/modprobe.d/nvidia.conf # Blacklist display modules for compute-only #blacklist nvidia_drm #blacklist nvidia_modeset -
A noter , pas possible de modifier la clock sur l’egpu :
# nvidia-smi --query-supported-clocks=memory --format=csv memory [MHz] 7001 MHz 6001 MHz 810 MHz 405 MHz 14001 MHz 13801 MHz 7001 MHz 810 MHz 405 MHz # nvidia-smi --lock-memory-clocks=810 Unable to determine the device handle for GPU1: 0000:05:00.0: Unknown Error Memory clocks set to "(memClkMin 810, memClkMax 810)" for GPU 00000000:01:00.0 Warning: persistence mode is disabled on device 00000000:01:00.0. See the Known Issues section of the nvidia-smi(1) man page for more information. Run with [--help | -h] switch to get more information on how to enable persistence mode. All done. -
Nouveau test :
# cat /etc/modprobe.d/nvidia-power-management.conf options nvidia_drm modeset=1 options nvidia NVreg_PreserveVideoMemoryAllocations=1 NVreg_TemporaryFilePath=/var/tmp # nvidia-smi --query-supported-clocks=memory --format=csv memory [MHz] 7001 MHz 6001 MHz 810 MHz 405 MHz 14001 MHz 13801 MHz 7001 MHz 810 MHz 405 MHz # nvidia-smi --lock-memory-clocks=810 Memory clocks set to "(memClkMin 810, memClkMax 810)" for GPU 00000000:01:00.0 Warning: persistence mode is disabled on device 00000000:01:00.0. See the Known Issues section of the nvidia-smi(1) man page for more information. Run with [--help | -h] switch to get more information on how to enable persistence mode. Memory clocks set to "(memClkMin 810, memClkMax 810)" for GPU 00000000:05:00.0 Warning: persistence mode is disabled on device 00000000:05:00.0. See the Known Issues section of the nvidia-smi(1) man page for more information. Run with [--help | -h] switch to get more information on how to enable persistence mode. All done. -
Resultat :
# llama-bench -m /models/gemma-3-1b-it-q4_k_m.gguf ggml_cuda_init: found 2 CUDA devices (Total VRAM: 21692 MiB): Device 0: NVIDIA GeForce RTX 5060 Ti, compute capability 12.0, VMM: yes, VRAM: 15888 MiB Device 1: NVIDIA GeForce RTX 3060 Laptop GPU, compute capability 8.6, VMM: yes, VRAM: 5803 MiB | model | size | params | backend | ngl | test | t/s | | ------------------------------ | ---------: | ---------: | ---------- | --: | --------------: | -------------------: | | gemma3 1B Q4_K - Medium | 762.49 MiB | 999.89 M | CUDA | -1 | pp512 | 2821.03 ± 23.15 | | gemma3 1B Q4_K - Medium | 762.49 MiB | 999.89 M | CUDA | -1 | tg128 | 26.17 ± 0.06 | build: c3d47e696 (10030)
Bonjour ! Vous semblez intéressé par cette conversation, mais vous n’avez pas encore de compte.
Marre de refaire défiler les mêmes messages ? Créez un compte pour retrouver votre position, recevoir des notifications des nouvelles réponses, sauvegarder vos favoris et voter pour les messages que vous appréciez.
Grâce à votre participation, ce message peut devenir encore meilleur 💗
S'inscrire Se connecter