公司開發需要測試新的 AMD顯卡，由於測試服務器上的顯卡是英偉達的顯卡所以換完后要安裝相應的驅動。由於之前裝機的同事裝的ubuntu是18.04.5 恰巧18.04.5在amd官網上沒有相匹配的驅動（這里由於開發之前在測試機上部署好了環境 就沒有及時將系統降級 耽誤了好久）
查看版本： lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 18.04.5 LTS
Release: 18.04
Codename: bionic
可能是版本不一致導致的中間的一些包裝不了 換了個系統后再找相應的驅動包：https://www.amd.com/en/support/graphics/amd-radeon-5700-series/amd-radeon-rx-5700-series/amd-radeon-rx-5700-xt這是我的對應驅動包
安裝：
我用的是 os x就用他的命令行scp過去后
解壓
$ cd ~/Downloads $ tar -Jxvf amdgpu-pro-YY.XX-NNNNNN.tar.xz $ cd ~/Downloads/amdgpu-pro-YY.XX-NNNNNN
他里面會有amdgpu-install 的自動安裝程序
執行./amdgpu-install 后觀察安裝過程有無報錯
Done后查看
出現紅框內的參數就說明安裝成功了
中間可能涉及到的輔助命令：
root@ubuntu18:~# dmesg | egrep "radeon|drm"
[ 2.091828] [drm] amdgpu kernel modesetting enabled.
[ 2.092435] [drm] amdgpu version: 5.6.0.20.20
[ 2.093056] [drm] OS DRM version: 4.15.0
[ 2.096776] fb: switching to amdgpudrmfb from EFI VGA
[ 2.097623] [drm] initializing kernel modesetting (NAVI10 0x1002:0x731F 0x1002:0x0B36 0xC1).
[ 2.097637] [drm] register mmio base: 0x9F200000
[ 2.097639] [drm] register mmio size: 524288
[ 2.116084] [drm] add ip block number 0 <nv_common>
[ 2.116086] [drm] add ip block number 1 <gmc_v10_0>
[ 2.116087] [drm] add ip block number 2 <navi10_ih>
[ 2.116088] [drm] add ip block number 3 <psp>
[ 2.116089] [drm] add ip block number 4 <smu>
[ 2.116091] [drm] add ip block number 5 <dm>
[ 2.116092] [drm] add ip block number 6 <gfx_v10_0>
[ 2.116093] [drm] add ip block number 7 <sdma_v5_0>
[ 2.116094] [drm] add ip block number 8 <vcn_v2_0>
[ 2.116095] [drm] add ip block number 9 <jpeg_v2_0>
[ 2.116174] [drm] VCN decode is enabled in VM mode
[ 2.116175] [drm] VCN encode is enabled in VM mode
[ 2.116177] [drm] JPEG decode is enabled in VM mode
[ 2.116195] [drm] vm size is 262144 GB, 4 levels, block size is 9-bit, fragment size is 9-bit
[ 2.116211] [drm] Detected VRAM RAM=8176M, BAR=256M
[ 2.116212] [drm] RAM width 256bits GDDR6
[ 2.116332] [drm] amdgpu: 8176M of VRAM memory ready
[ 2.116335] [drm] amdgpu: 257874M of GTT memory ready.
[ 2.116338] [drm] GART: num cpu pages 131072, num gpu pages 131072
[ 2.116461] [drm] PCIE GART of 512M enabled (table at 0x0000008000300000).
[ 2.118042] [drm] Found VCN firmware Version ENC: 1.9 DEC: 5 VEP: 0 Revision: 0
[ 2.118048] [drm] PSP loading VCN firmware
[ 2.776381] [drm] reserve 0x900000 from 0x81fe400000 for PSP TMR
[ 3.039879] [drm] Display Core initialized with v3.2.81!
[ 3.103858] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 3.103861] [drm] Driver supports precise vblank timestamp query.
[ 3.105902] [drm] kiq ring mec 2 pipe 1 q 0
[ 3.117203] [drm] VCN decode and encode initialized successfully(under DPG Mode).
[ 3.118253] [drm] JPEG decode initialized successfully.
[ 3.121604] [drm] fb mappable at 0x804CA000
[ 3.121606] [drm] vram apper at 0x80000000
[ 3.121607] [drm] size 8294400
[ 3.121608] [drm] fb depth is 24
[ 3.121610] [drm] pitch is 7680
[ 3.121670] fbcon: amdgpudrmfb (fb0) is primary device
[ 3.124830] amdgpu 0000:44:00.0: fb0: amdgpudrmfb frame buffer device
[ 3.157040] [drm] Initialized amdgpu 3.37.0 20150101 for 0000:44:00.0 on minor 0
root@ubuntu18:~# lspci -nn |grep VGA
44:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Device [1002:73
1f] (rev c1)
apt install radeontop
root@ubuntu18:~# radeontop
amdgpu DRM driver is used, but amdgpu VRAM size reporting is not enabled
amdgpu DRM driver is used, but amdgpu VRAM usage reporting is not enabled
Failed to get VRAM usage, kernel likely too old
Unknown Radeon card. <= R500 won't work, new cards might.
Collecting data, please wait....
由於我這邊的不支持以下命令，各位可以在自己的機器上試下
glxinfo | grep rendering（）
如果結果是“yes”，證明顯卡驅動已經成功安裝。
如果提示有問題，可能是系統里面沒有安裝mesa-utils，安裝過程：在終端輸入命令：sudo apt-get install mesa-utils
然后再運行命令：glxinfo | grep rendering
這就差不多了結束了 ，安裝不成功首先要檢查系統版本跟驅動版本的問題 要保持一致。
總結：
- 下載對應讀驅動包
- 解壓安裝
- reboot（重點）
- lshw -C display
按照amd官網給的安裝過程較好