美文网首页
看GPU NUMA

看GPU NUMA

作者: 陈先生_9e91 | 来源:发表于2019-01-15 10:21 被阅读0次

查看GPU NUMA

[TOC]

参考:https://blog.csdn.net/dcrmg/article/details/78146797

[yangxue.chen@yz-gpu002 ~]$ lspci | grep -i nvidia
02:00.0 VGA compatible controller: NVIDIA Corporation GM200 [GeForce GTX TITAN X] (rev a1)
02:00.1 Audio device: NVIDIA Corporation GM200 High Definition Audio (rev a1)
03:00.0 VGA compatible controller: NVIDIA Corporation GM200 [GeForce GTX TITAN X] (rev a1)
03:00.1 Audio device: NVIDIA Corporation GM200 High Definition Audio (rev a1)
82:00.0 VGA compatible controller: NVIDIA Corporation GM200 [GeForce GTX TITAN X] (rev a1)
82:00.1 Audio device: NVIDIA Corporation GM200 High Definition Audio (rev a1)
83:00.0 VGA compatible controller: NVIDIA Corporation GM200 [GeForce GTX TITAN X] (rev a1)
83:00.1 Audio device: NVIDIA Corporation GM200 High Definition Audio (rev a1)

前边的序号 "02:00.0"是显卡的代号

[yangxue.chen@yz-gpu002 ~]$ lspci -v -s 02:00.0
02:00.0 VGA compatible controller: NVIDIA Corporation GM200 [GeForce GTX TITAN X] (rev a1) (prog-if 00 [VGA controller])
        Subsystem: NVIDIA Corporation Device 1132
        Physical Slot: 2
        Flags: bus master, fast devsel, latency 0, IRQ 112, NUMA node 0
        Memory at cf000000 (32-bit, non-prefetchable) [size=16M]
        Memory at 23fe0000000 (64-bit, prefetchable) [size=256M]
        Memory at 23ff0000000 (64-bit, prefetchable) [size=32M]
        I/O ports at 6000 [size=128]
        [virtual] Expansion ROM at d0000000 [disabled] [size=512K]
        Capabilities: <access denied>
        Kernel driver in use: nvidia
        Kernel modules: nouveau, nvidia_drm, nvidia

NUMA 0

[yangxue.chen@yz-gpu002 ~]$ lspci -v -s 82:00.0
82:00.0 VGA compatible controller: NVIDIA Corporation GM200 [GeForce GTX TITAN X] (rev a1) (prog-if 00 [VGA controller])
        Subsystem: NVIDIA Corporation Device 1132
        Physical Slot: 8
        Flags: bus master, fast devsel, latency 0, IRQ 114, NUMA node 1
        Memory at fa000000 (32-bit, non-prefetchable) [size=16M]
        Memory at 27fe0000000 (64-bit, prefetchable) [size=256M]
        Memory at 27ff0000000 (64-bit, prefetchable) [size=32M]
        I/O ports at e000 [size=128]
        [virtual] Expansion ROM at fb000000 [disabled] [size=512K]
        Capabilities: <access denied>
        Kernel driver in use: nvidia
        Kernel modules: nouveau, nvidia_drm, nvidia

NUMA 1

只有一个 NUMA下的GPU才可以P2P

[yangxue.chen@yz-gpu002 ~]$ nvidia-smi
Mon Jan 14 09:57:51 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 396.45                 Driver Version: 396.45                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TIT...  Off  | 00000000:02:00.0 Off |                  N/A |
| 42%   83C    P2   163W / 250W |    793MiB / 12212MiB |     74%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX TIT...  Off  | 00000000:03:00.0 Off |                  N/A |
| 22%   33C    P8    16W / 250W |      0MiB / 12212MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX TIT...  Off  | 00000000:82:00.0 Off |                  N/A |
| 22%   32C    P8    15W / 250W |      0MiB / 12212MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   3  GeForce GTX TIT...  Off  | 00000000:83:00.0 Off |                  N/A |
| 34%   76C    P2   154W / 250W |    793MiB / 12212MiB |     67%      Default |
+-------------------------------+----------------------+----------------------+

通过前面的代号,可以看到对应GPU信息

相关文章

  • 看GPU NUMA

    查看GPU NUMA [TOC] 参考:https://blog.csdn.net/dcrmg/article/d...

  • CPU的制造和概念

    相关概念:多核、超线程、NUMA、睿频、功耗、GPU、大小核、分支预测、cache_line失效、加锁代价、IPC...

  • Linux内核load balance(一)

    [TOC] ## NUMA ### 为什么要有NUMA 在NUMA架构出现前,CPU欢快的朝着频率越来越高的方向发...

  • OpenStack 实现 GPU 与 CPU、内存的 NUMA

    “负责安卓云底层开发的小彭请了假,这个需求有点紧,你来做吧 ” 前言 这是我入职新公司的第二周。 我所接到的需求,...

  • ceph14.1.x的一下新功能(二)

    10、osd的numa亲和 ceph14开始支持osd的自动numa亲和,这样就不用麻烦再写脚本做numa绑定了。...

  • 【numa】关于numa

    # yum -y install numactl # yum info numactl # rpm -ql num...

  • 浅解NUMA机制

    导读 本文适合知道NUMA这个词但想进一步了解的新手。 以下的文章内容包括:NUMA的产生背景,NUMA的架构细节...

  • CPU虚拟化技术

    CPU虚拟化技术 NUMA技术 NUMA是一种解决多CPU共同工作的技术方案NUMA模式是每个处理器有自己的存储器...

  • 监测numa cpu mem 负载

    两个工具 node_exporter 和 numa top 至少要拿到numa cpu mem usage 统计信...

  • NUMA

网友评论

      本文标题:看GPU NUMA

      本文链接:https://www.haomeiwen.com/subject/lwuedqtx.html