MetaX-Tech Developer Forum 论坛首页
  • 沐曦开发者
search
Sign in

Heyerek

  • Members
  • Joined 2026年6月30日
  • message 帖子
  • forum 主题
  • favorite 关注者
  • favorite_border Follows
  • person_outline 详细信息

Heyerek has posted 10 messages.

  • See post chevron_right
    Heyerek
    Members
    C500的能否单卡跑一个Docker 2个vllm的4b小模型 已解决 2026年9月18日 10:49

    最新版本支持是吗

  • See post chevron_right
    Heyerek
    Members
    C500的能否单卡跑一个Docker 2个vllm的4b小模型 已解决 2026年9月18日 10:47

    !/bin/bash

    unset CUDA_VISIBLE_DEVICES

    启动Embedding

    vllm serve /mnt/model/ \
    --host 0.0.0.0 \
    --port 8080 \
    --trust_remote_code \
    --served-model-name Qwen3-Embedding-4B \
    --tensor-parallel-size 1 \
    --max-model-len 4096 \
    --gpu-memory-utilization 0.35 &
    EMB_PID=$!

    echo "等待embedding端口8080就绪..."
    until curl -s http://127.0.0.1:8080/v1/models >/dev/null 2>&1
    do
    sleep 3
    echo "8080还没就绪,继续等待..."
    done
    echo "✅ embedding 8080已经就绪"

    启动reranker,压低参数

    vllm serve /mnt/models/ \
    --host 0.0.0.0 \
    --port 8081 \
    --trust_remote_code \
    --served-model-name Qwen3-Reranker-4B \
    --tensor-parallel-size 1 \
    --max-model-len 512 \
    --gpu-memory-utilization 0.20 &
    RERANK_PID=$!

    echo "reranker已启动 PID $RERANK_PID"
    wait

    启动命令如下

  • See post chevron_right
    Heyerek
    Members
    C500的能否单卡跑一个Docker 2个vllm的4b小模型 已解决 2026年9月18日 10:45

    昨晚试了一下第二个一直起不来也没报错

  • See post chevron_right
    Heyerek
    Members
    C500的能否单卡跑一个Docker 2个vllm的4b小模型 已解决 2026年9月18日 10:45

    这个驱动也可以吗,驱动版本是2.多的

  • See post chevron_right
    Heyerek
    Members
    C500的能否单卡跑一个Docker 2个vllm的4b小模型 已解决 2026年9月18日 09:59

    版本驱动如下

    mx-smi

    mx-smi version: 2.2.3

    =================== MetaX System Management Interface Log ===================
    Timestamp : Fri Sep 18 09:57:10 2026

    Attached GPUs : 8
    +---------------------------------------------------------------------------------+
    | MX-SMI 2.2.3 Kernel Mode Driver Version: 2.14.6 |
    | MACA Version: 2.32.0.6 BIOS Version: 1.24.3.0 |
    |------------------------------------+---------------------+----------------------+
    | GPU NAME | Bus-id | GPU-Util |
    | Temp Pwr:Usage/Cap | Memory-Usage | |
    |====================================+=====================+======================|
    | 0 MetaX C500 | 0000:43:00.0 | 0% |
    | 41C 52W / 350W | 59366/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 1 MetaX C500 | 0000:45:00.0 | 0% |
    | 43C 54W / 350W | 60006/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 2 MetaX C500 | 0000:47:00.0 | 0% |
    | 46C 56W / 350W | 60006/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 3 MetaX C500 | 0000:4a:00.0 | 0% |
    | 47C 58W / 350W | 59366/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 4 MetaX C500 | 0000:c4:00.0 | 0% |
    | 44C 53W / 350W | 59715/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 5 MetaX C500 | 0000:c6:00.0 | 0% |
    | 45C 54W / 350W | 59605/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 6 MetaX C500 | 0000:c8:00.0 | 0% |
    | 47C 57W / 350W | 59635/65536 MiB | |
    +------------------------------------+---------------------+----------------------+
    | 7 MetaX C500 | 0000:ca:00.0 | 0% |
    | 42C 42W / 350W | 858/65536 MiB | |
    +------------------------------------+---------------------+----------------------+

    +---------------------------------------------------------------------------------+
    | Process: |
    | GPU PID Process Name GPU Memory |
    | Usage(MiB) |
    |=================================================================================|
    | 0 4104065 VLLM::Worker_TP 58500 |
    | 1 4104066 VLLM::Worker_TP 59140 |
    | 2 4104069 VLLM::Worker_TP 59140 |
    | 3 4104070 VLLM::Worker_TP 58500 |
    | 4 2504707 VLLM::EngineCor 58856 |
    | 5 2267768 VLLM::EngineCor 58746 |
    | 6 3061825 VLLM::EngineCor 58776 |
    +---------------------------------------------------------------------------------+

  • See post chevron_right
    Heyerek
    Members
    C500的卡能一次跑2个小模型吗 已解决 2026年9月18日 09:50

    C500的卡能一个容器通过vllm一次跑2个小模型吗,驱动是3.8.0

  • See post chevron_right
    Heyerek
    Members
    乌班图安装22.04 vbiso 1.33-c500-vf.bin 最新驱动 MACA Version: unknown报错 已解决 2026年6月30日 18:22

    已解决感谢

  • See post chevron_right
    Heyerek
    Members
    乌班图安装22.04 vbiso 1.33-c500-vf.bin 最新驱动 MACA Version: unknown报错 已解决 2026年6月30日 18:07

    怎么安装呢有没有教程

  • See post chevron_right
    Heyerek
    Members
    乌班图安装22.04 vbiso 1.33-c500-vf.bin 最新驱动 MACA Version: unknown报错 已解决 2026年6月30日 18:02

    物理机中系统是乌班图22.04 metax-driver-3.7.2.30-deb-x86_64.run 离线安装的,直接./metax-driver-3.7.2.30-deb-x86_64.run -- -f 然后mx-smi提示vbios需要更新,就更新了mx-smi -u /lib/firmware/metax/mxc500/mxvbios-1.33.5.0-1174-C500-VF.bin 更新后MACA出不来

  • See post chevron_right
    Heyerek
    Members
    乌班图安装22.04 vbiso 1.33-c500-vf.bin 最新驱动 MACA Version: unknown报错 已解决 2026年6月30日 17:50

    mx-smi
    mx-smi version: 2.3.1

    =================== MetaX System Management Interface Log ===================
    Timestamp : Tue Jun 30 17:44:54 2026

    Attached GPUs : 8
    +---------------------------------------------------------------------------------+
    | MX-SMI 2.3.1 Kernel Mode Driver Version: 3.8.30 |
    | MACA Version: unknown BIOS Version: 1.33.5.0 |
    |------------------+-----------------+---------------------+----------------------|
    | Board Name | GPU Persist-M | Bus-id | GPU-Util sGPU-M |
    | Pwr:Usage/Cap | Temp Perf | Memory-Usage | GPU-State |
    |==================+=================+=====================+======================|
    | 0 MetaX C500 | 0 Off | 0000:43:00.0 | 0% Disabled |
    | 40W / 350W | 39C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 1 MetaX C500 | 1 Off | 0000:45:00.0 | 0% Disabled |
    | 41W / 350W | 41C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 2 MetaX C500 | 2 Off | 0000:47:00.0 | 0% Disabled |
    | 43W / 350W | 43C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 3 MetaX C500 | 3 Off | 0000:4a:00.0 | 0% Disabled |
    | 44W / 350W | 44C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 4 MetaX C500 | 4 Off | 0000:c4:00.0 | 0% Disabled |
    | 40W / 350W | 41C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 5 MetaX C500 | 5 Off | 0000:c6:00.0 | 0% Disabled |
    | 41W / 350W | 42C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 6 MetaX C500 | 6 Off | 0000:c8:00.0 | 0% Disabled |
    | 43W / 350W | 43C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+
    | 7 MetaX C500 | 7 Off | 0000:ca:00.0 | 0% Disabled |
    | 42W / 350W | 43C P0 | 858/65536 MiB | Available |
    +------------------+-----------------+---------------------+----------------------+

    +---------------------------------------------------------------------------------+
    | Process: |
    | GPU PID Process Name GPU Memory |
    | Usage(MiB) |
    |=================================================================================|
    | no process found |
    +---------------------------------------------------------------------------------+

    dpkg -l | egrep "(metax|mxsmt|mxfw)"
    ii metax-linux 3.8.30 amd64 metax deb package
    ii mxfw 3.7.2 all mxfw binary deb package
    ii mxsmt 3.7.2.30 amd64 mxsmt deb package

  • 沐曦开发者论坛
powered by misago