tamiwiki:projects:egpu
Differences
This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revisionNext revision | Previous revision | ||
| tamiwiki:projects:egpu [2023/06/15 14:39] – [P40] yair | tamiwiki:projects:egpu [2023/11/04 11:07] (current) – [1080Ti] yair | ||
|---|---|---|---|
| Line 1: | Line 1: | ||
| ====== EGPU ====== | ====== EGPU ====== | ||
| - | https:// | + | {{ : |
| + | |||
| + | we are using the [[https:// | ||
| + | |||
| + | Linux Kernal notes > https:// | ||
| + | [[https:// | ||
| + | |||
| + | |||
| + | === ThunderBolt check and setup === | ||
| TLDR | TLDR | ||
| - upgrade kernel (??) | - upgrade kernel (??) | ||
| Line 46: | Line 55: | ||
| <code bash> | <code bash> | ||
| (base) user@eight: | (base) user@eight: | ||
| - | 1 | + | </code> |
| - | (base) user@eight: | + | |
| - | == /sys/ | + | |
| - | modalias : pci: | + | |
| - | vendor | + | |
| - | model : GF106GL [Quadro 2000] | + | |
| - | manual_install: | + | |
| - | driver | + | |
| - | driver | + | |
| - | </ | ||
| - | just an old card... | ||
| - | but EEK | ||
| + | ==== 1080Ti ==== | ||
| + | {{ : | ||
| + | looks legit | ||
| <code bash> | <code bash> | ||
| - | (base) user@eight: | + | $sudo dmesg -w |
| - | 08:04.0 PCI bridge: Intel Corporation JHL7540 Thunderbolt 3 Bridge | + | [96236.873213] nvidia-nvlink: Nvlink Core is being initialized, |
| - | 09:00.0 VGA compatible controller: NVIDIA | + | |
| - | 09: | + | [96236.874544] nvidia 0000: |
| + | [96236.874646] nvidia 0000:09:00.0: vgaarb: changed | ||
| + | [96236.991272] NVRM: loading | ||
| + | [96237.009537] nvidia-modeset: | ||
| + | [96237.013346] [drm] [nvidia-drm] [GPU ID 0x00000900] Loading driver | ||
| + | [96238.239429] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:09:00.0 on minor 1 | ||
| + | [96238.269008] nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint. | ||
| + | [96238.330257] nvidia-uvm: Loaded the UVM driver, major device | ||
| + | [96238.399348] NVRM: API mismatch: the client has the version 390.157, but | ||
| + | NVRM: this kernel module has the version 535.43.02. | ||
| + | NVRM: make sure that this kernel module and all NVIDIA | ||
| - | $sudo dmesg | ||
| - | [ 1041.053826] nvidia: module license ' | ||
| - | [ 1041.053831] Disabling lock debugging due to kernel taint | ||
| - | [ 1041.484017] nvidia-nvlink: | ||
| - | [ 1041.484032] NVRM: The NVIDIA Quadro 2000 GPU installed in this system is | ||
| - | | ||
| - | | ||
| - | | ||
| - | | ||
| - | [ 1041.501047] NVRM: No NVIDIA GPU found. | ||
| - | [ 1041.521176] nvidia-nvlink: | ||
| - | [ 1042.332830] nvidia-nvlink: | ||
| - | [ 1042.332842] NVRM: The NVIDIA Quadro 2000 GPU installed in this system is | ||
| - | | ||
| - | | ||
| - | | ||
| - | | ||
| - | [ 1042.335282] NVRM: No NVIDIA GPU found. | ||
| - | [ 1042.335835] nvidia-nvlink: | ||
| </ | </ | ||
| - | <WRAP center round alert 33%>WE ARE TAINTED</ | + | update the driver |
| - | + | ||
| - | ==== driver | + | |
| - | + | ||
| - | we went with ubuntu selection | + | |
| - | + | ||
| - | but cute https:// | + | |
| <code bash> | <code bash> | ||
| - | $ sudo apt installl nvidia-headless-535 | + | $ ubuntu-drivers devices |
| - | + | == / | |
| - | + | modalias : pci: | |
| - | #downgrade nvidia to quadro supported version | + | vendor |
| - | sudo apt install nvidia-headless-390 | + | model : GP102 [GeForce GTX 1080 Ti] |
| - | + | manual_install: | |
| - | # EEK | + | driver |
| - | RROR (dkms apport): kernel package linux-headers-6.3.7-060307-generic is not supported | + | driver |
| - | Error! Bad return status for module build on kernel: 6.3.7-060307-generic (x86_64) | + | driver |
| - | Consult / | + | driver |
| - | dpkg: error processing package | + | driver |
| - | installed | + | driver |
| - | dpkg: dependency problems prevent configuration of nvidia-headless-390: | + | driver |
| - | nvidia-headless-390 depends on nvidia-dkms-390; however: | + | driver |
| - | Package | + | driver |
| - | + | driver | |
| - | dpkg: error processing package | + | driver |
| - | | + | driver |
| - | Processing triggers for libc-bin (2.36-0ubuntu4) ... | + | |
| - | No apport report written because the error message indicates its a followup error from a previous failure. | + | |
| - | / | + | |
| - | ndi.so.4 is not a symbolic link | + | |
| - | + | ||
| - | Processing triggers for man-db (2.10.2-2) ... | + | |
| - | Processing triggers for initramfs-tools (0.140ubuntu17) ... | + | |
| - | update-initramfs: Generating / | + | |
| - | Errors were encountered while processing: | + | |
| - | | + | |
| - | | + | |
| + | $ sudo ubuntu-drivers autoinstall | ||
| </ | </ | ||
| - | downgrading but to headless,\\ | ||
| - | without touching the x config? | ||
| + | ==== P40 ==== | ||
| + | <WRAP center round important 60%> | ||
| + | this doesnt work on our test machine | ||
| + | </ | ||
| - | going with [[https:// | ||
| - | [1] A 3D video game environment and benchmark designed from scratch for reinforcement learning research | + | {{ : |
| - | <code bash> | + | the P40 needs modern motherboard that allow for '' |
| - | conda create | + | |
| - | conda activate avalon | + | |
| - | sudo apt install | + | NVIDIA Tesla P40 24GB DDR5 GPU Accelerator Card Dual PCI-E 3.0 x16\\ |
| + | need to retrofit with a FAN,it doesnt come with one | ||
| - | #this will also install torch... | + | got one on ebay for 200$(+shipping) ([[https:// |
| - | pip install avalon-rl[train] | + | |
| - | python -m avalon.install_godot_binary | + | some dude got it working, https:// |
| - | python | + | === SPECIFICATIONS: |
| - | </ | + | |
| - | why even bother, the quaDRO IS JUST A TEST. \\ | + | * GPU Architecture: |
| - | NEED TO CLEAN REMOVE THE 390 driver AND MOVE BACK TO | + | * Single-Precision Performance 12 TeraFLOPS* |
| + | * Integer Operations (INT8) 47 TOPS* (TeraOperations per Second) | ||
| + | * GPU Memory 24 GB | ||
| + | * Memory Bandwidth 346 GB/s | ||
| + | * System Interface PCI Express 3.0 x16 | ||
| + | * Form Factor 4.4” H x 10.5” L, Dual Slot, Full Height | ||
| + | * Max Power 250 W | ||
| + | * Enhanced Programmability with Page Migration Engine Yes | ||
| + | * ECC Protection Yes | ||
| + | * Server-Optimized for Data Center Deployment Yes | ||
| + | * Hardware-Accelerated Video Engine 1x Decode Engine, 2x Encode Engine /> | ||
| + | * NVPN: 699-2G610-0200-100 | ||
| + | * NVIDIA® CUDA® cores: 3840 | ||
| - | NVIDIA-CURRENT | ||
| - | |||
| - | |||
| - | ==== P40 ==== | ||
| - | https:// | ||
| installing | installing | ||
| <code bash> | <code bash> | ||
| - | sudo apt instakll | + | sudo apt install |
| </ | </ | ||
| there is some issue, unlike other cards the blue led doesnt turn green on thunderbolt connection.\\ | there is some issue, unlike other cards the blue led doesnt turn green on thunderbolt connection.\\ | ||
| no power passing to the gPU.\\ | no power passing to the gPU.\\ | ||
| - | unlike with other cards we tried (quadro 2000 and 660Ti) | ||
| :( | :( | ||
| + | |||
| ==== misc ==== | ==== misc ==== | ||
tamiwiki/projects/egpu.1686829194.txt.gz · Last modified: 2023/06/15 14:39 by yair
Page Tools
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Noncommercial-Share Alike 4.0 International
