Files
rocm-systems/tools/topo_expl/models/topo_8p_rome_4n_2.xml
T
Wenkai Du 5c8380ff5b Implement NIC identification and remapping (#420)
* Add 1H16P GPU model

* Implement NIC identification and remapping

* Revert "Sort IB devices based on device name (#413)"

This reverts commit 2d0ed8dff6.

* Fix permute and check order

* Correction on IB speed reporting

* Revert "Allow user to link layer with RCCL_IB_HCA_SKIP_LINK_LAYER (#361)"

This reverts commit caf5c9992a.
2021-08-24 09:42:04 -07:00

127 خطوط
9.5 KiB
XML

<system version="2">
<cpu numaid="1" affinity="00000000,00000000,00000000,ffff0000,00000000,00000000,00000000,ffff0000" arch="x86_64" vendor="AuthenticAMD" familyid="143" modelid="49">
<pci busid="0000:43:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:46:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:48:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:4a:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="0" sm="90" gcn="908" arch="38911" rank="0" gdr="1">
<xgmi target="0000:50:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:0a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:0f:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:4c:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:4e:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:50:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="1" sm="90" gcn="908" arch="38911" rank="1" gdr="1">
<xgmi target="0000:4a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:0a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:0f:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:45:00.0" class="0x020000" vendor="0x15b3" device="0x101b" subsystem_vendor="0x1590" subsystem_device="0x0345" link_speed="16 GT/s" link_width="16">
<nic>
<net name="mlx5_1" dev="3" speed="100000" port="1" guid="0x181515ffffa4e988" maxconn="262144" gdr="1"/>
</nic>
</pci>
</pci>
</cpu>
<cpu numaid="3" affinity="00000000,00000000,ffff0000,00000000,00000000,00000000,ffff0000,00000000" arch="x86_64" vendor="AuthenticAMD" familyid="143" modelid="49">
<pci busid="0000:03:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:05:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:08:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:0a:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="2" sm="90" gcn="908" arch="38911" rank="2" gdr="1">
<xgmi target="0000:4a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:50:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:0f:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:0b:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:0d:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:0f:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="3" sm="90" gcn="908" arch="38911" rank="3" gdr="1">
<xgmi target="0000:4a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:50:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:0a:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:13:00.0" class="0x020000" vendor="0x15b3" device="0x101b" subsystem_vendor="0x1590" subsystem_device="0x0345" link_speed="16 GT/s" link_width="16">
<nic>
<net name="mlx5_3" dev="2" speed="100000" port="1" guid="0x301515ffffa4e988" maxconn="262144" gdr="1"/>
</nic>
</pci>
</pci>
</cpu>
<cpu numaid="5" affinity="00000000,ffff0000,00000000,00000000,00000000,ffff0000,00000000,00000000" arch="x86_64" vendor="AuthenticAMD" familyid="143" modelid="49">
<pci busid="0000:c3:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:c6:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:c8:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:ca:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="4" sm="90" gcn="908" arch="38911" rank="4" gdr="1">
<xgmi target="0000:d0:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:8a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:90:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:cc:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:ce:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:d0:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="5" sm="90" gcn="908" arch="38911" rank="5" gdr="1">
<xgmi target="0000:ca:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:8a:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:90:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:c5:00.0" class="0x020000" vendor="0x15b3" device="0x101b" subsystem_vendor="0x1590" subsystem_device="0x0345" link_speed="16 GT/s" link_width="16">
<nic>
<net name="mlx5_5" dev="1" speed="100000" port="1" guid="0x601515ffffa4e988" maxconn="262144" gdr="1"/>
</nic>
</pci>
</pci>
</cpu>
<cpu numaid="7" affinity="ffff0000,00000000,00000000,00000000,ffff0000,00000000,00000000,00000000" arch="x86_64" vendor="AuthenticAMD" familyid="143" modelid="49">
<pci busid="0000:83:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:86:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:88:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:8a:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="6" sm="90" gcn="908" arch="38911" rank="6" gdr="1">
<xgmi target="0000:ca:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:d0:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:90:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:8c:00.0" class="0x060400" vendor="0x11f8" device="0x4000" subsystem_vendor="0x11f8" subsystem_device="0xbeef" link_speed="16 GT/s" link_width="16">
<pci busid="0000:8e:00.0" class="0x060400" vendor="0x1002" device="0x14a0" subsystem_vendor="0x0000" subsystem_device="0x0000" link_speed="16 GT/s" link_width="16">
<pci busid="0000:90:00.0" class="0x038000" vendor="0x1002" device="0x738c" subsystem_vendor="0x1002" subsystem_device="0x0c34" link_speed="16 GT/s" link_width="16">
<gpu dev="7" sm="90" gcn="908" arch="38911" rank="7" gdr="1">
<xgmi target="0000:ca:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:d0:00.0" count="1" tclass="0x038000"/>
<xgmi target="0000:8a:00.0" count="1" tclass="0x038000"/>
</gpu>
</pci>
</pci>
</pci>
<pci busid="0000:85:00.0" class="0x020000" vendor="0x15b3" device="0x101b" subsystem_vendor="0x1590" subsystem_device="0x0345" link_speed="16 GT/s" link_width="16">
<nic>
<net name="mlx5_7" dev="0" speed="100000" port="1" guid="0x64780affffa4e988" maxconn="262144" gdr="1"/>
</nic>
</pci>
</pci>
</cpu>
</system>