Accelerator scheduling based on cooling characteristics
An information handling system may include at least one central processing unit (CPU); and a plurality of special-purpose processing units. The information handling system may be configured to: receive information regarding cooling characteristics of the plurality of special-purpose processing units; and assign identification (ID) numbers to each of the plurality of special-purpose processing units in an order that is determined based at least in part on the cooling characteristics.
1. An information handling system comprising:
at least one central processing unit (CPU); and
a plurality of special-purpose processing units;
wherein the information handling system is configured to:
receive information regarding cooling characteristics of the plurality of special-purpose processing units;
assign identification (ID) numbers to each of the plurality of special-purpose processing units in an order that is determined based at least in part on the cooling characteristics, wherein the ID numbers are Peripheral Component Interconnect Express (PCIe) bus ID numbers;
execute a scheduling tool that is configured to assign jobs to the plurality of special-purpose processing units based solely on their respective ID numbers; and
perform the jobs at the plurality of special-purpose processing units according to the assignments of the scheduling tool.
2. The information handling system of claim 1 , wherein the special-purpose processing units are graphics processing units (GPUs).
3. The information handling system of claim 1 , further configured to, during a boot process of the information handling system:
execute a thermal stress test; and
based on results of the thermal stress test, reassign the ID numbers to each of the plurality of special-purpose processing units in a different order.
4. The information handling system of claim 1 , further configured to:
receive input from a user of the information handling system; and
based on the input, reassign the ID numbers to each of the plurality of special-purpose processing units in a different order.
5. The information handling system of claim 1 , wherein the ID numbers are assigned by a management controller of the information handling system.
6. The information handling system of claim 1 , wherein the order is further determined based at least in part on peer-to-peer connectivity of the plurality of special-purpose processing units.
7. A method comprising:
an information handling system that includes a plurality of special-purpose processing units receiving information regarding cooling characteristics of the plurality of special-purpose processing units;
the information handling system assigning identification (ID) numbers to each of the plurality of special-purpose processing units in an order that is determined based at least in part on the cooling characteristics, wherein the ID numbers are Peripheral Component Interconnect Express (PCIe) bus ID numbers;
the information handling system executing a scheduling tool that is configured to assign jobs to the plurality of special-purpose processing units based solely on their respective ID numbers; and
the information handling system performing the jobs at the plurality of special-purpose processing units according to the assignments of the scheduling tool.
8. The method of claim 7 , wherein the special-purpose processing units are graphics processing units (GPUs).
9. The method of claim 7 , further comprising, during a boot process of the information handling system:
the information handling system executing a thermal stress test; and
based on results of the thermal stress test, the information handling system reassigning the ID numbers to each of the plurality of special-purpose processing units in a different order.
10. The method of claim 7 , further comprising:
the information handling system receiving input from a user; and
based on the input, the information handling system reassigning the ID numbers to each of the plurality of special-purpose processing units in a different order.
11. The method of claim 7 , wherein the ID numbers are assigned by a management controller of the information handling system.
12. The method of claim 7 , wherein the order is further determined based at least in part on peer-to-peer connectivity of the plurality of special-purpose processing units.
13. An article of manufacture comprising a non-transitory, computer-readable medium having computer-executable code thereon that is executable by a processor of an information handling system that includes a plurality of special-purpose processing units, the code being executable for:
receiving information regarding cooling characteristics of the plurality of special-purpose processing units;
assigning identification (ID) numbers to each of the plurality of special-purpose processing units in an order that is determined based at least in part on the cooling characteristics, wherein the ID numbers are Peripheral Component Interconnect Express (PCIe) bus ID numbers;
executing a scheduling tool that is configured to assign jobs to the plurality of special-purpose processing units based solely on their respective ID numbers; and
performing the jobs at the plurality of special-purpose processing units according to the assignments of the scheduling tool.
14. The article of claim 13 , wherein the special-purpose processing units are graphics processing units (GPUs).
15. The article of claim 13 , wherein the code is further executable for, during a boot process of the information handling system:
executing a thermal stress test; and
based on results of the thermal stress test, reassigning the ID numbers to each of the plurality of special-purpose processing units in a different order.
16. The article of claim 13 , wherein the code is further executable for:
receiving input from a user; and
based on the input, reassigning the ID numbers to each of the plurality of special-purpose processing units in a different order.
17. The article of claim 13 , wherein the ID numbers are assigned by a management controller of the information handling system.
18. The article of claim 13 , wherein the order is further determined based at least in part on peer-to-peer connectivity of the plurality of special-purpose processing units.