Increased NAND performance under high thermal conditions
View Patent ↗Devices and techniques for increased NAND performance under high thermal conditions are disclosed herein. An indicator of a high-temperature thermal condition for a NAND device may be obtained. A workload of the NAND device may be measured in response to the high-temperature thermal condition. Operation of the NAND device may then be modified based on the workload and the high-temperature thermal condition.
1. A non-transitory machine readable medium including instructions to increase burst write performance in a NAND device, the instructions, when executed by processing circuitry, cause the processing circuitry to perform operations comprising:
operating a NAND device in accordance with high-temperature policy in response to a high-temperature thermal condition of the NAND device;
measuring a command queue of the NAND device to determine that a current workload exhibits a bursty workload profile, the bursty workload profile being a temporary rather than sustained increased change in activity for the NAND device;
suspending the high-temperature policy with respect to the current workload;
processing the current workload while the high-temperature policy is suspended; and
re-implementing the high-temperature policy for the NAND device in response to completing the current workload.
2. The machine readable medium of claim 1 , wherein the high-temperature thermal condition is one of an ordered set of thermal conditions for the NAND device, members of the set of thermal conditions having respective operating policies for the NAND device.
3. The machine readable medium of claim 2 , wherein the respective operating policies include throttling a resource of the NAND device, wherein a higher order in the set of thermal conditions corresponds to a greater degree of throttling.
4. The machine readable medium of claim 3 , wherein throttling the resource includes powering-off a component of the NAND device, a greater degree of throttling corresponding to powering-off a greater number of components of the NAND device.
5. The machine readable medium of claim 1 , wherein a bursty workload profile has a time-averaged command queue depth that is small in comparison to a depth of the command queue at a time that it includes commands.
6. The machine readable medium of claim 5 , wherein measuring the command queue of the NAND device to determine that a current workload exhibits a bursty workload profile includes comparing a current rate of commands into the command queue to an averaged historical rate of commands into the command queue over time.
7. A NAND device with increased burst write performance, the NAND device comprising:
a storage medium to hold commands in a command queue; and
processing circuitry to:
operate the NAND device in accordance with high-temperature policy in response to a high-temperature thermal condition of the NAND device;
measure the command queue of the NAND device to determine that a current workload exhibits a bursty workload profile, the bursty workload profile being a temporary rather than sustained increased change in activity for the NAND device;
suspend the high-temperature policy with respect to the current workload;
process the current workload while the high-temperature policy is suspended; and
re-implement the high-temperature policy for the NAND device in response to completion of the current workload.
8. The NAND device of claim 7 , wherein the high-temperature thermal condition is one of an ordered set of thermal conditions for the NAND device, members of the set of thermal conditions having respective operating policies for the NAND device.
9. The NAND device of claim 8 , wherein the respective operating policies direct the processing circuitry to throttle a resource of the NAND device, wherein a higher order in the set of thermal conditions corresponds to a greater degree of throttling.
10. The NAND device of claim 9 , wherein, to throttle the resource, the processing circuitry powers-off a component of the NAND device, a greater degree of throttling corresponding to powering-off a greater number of components of the NAND device.
11. The NAND device of claim 7 , wherein a bursty workload profile has a time-averaged command queue depth that is small in comparison to a depth of the command queue at a time that it includes commands.
12. The NAND device of claim 11 , wherein, to measure the command queue of the NAND device to determine that a current workload exhibits a bursty workload profile, the processing circuitry compares a current rate of commands into the command queue to an averaged historical rate of commands into the command queue over time.
13. A method to increase burst write performance, the method comprising:
operating a NAND device in accordance with high-temperature policy in response to a high-temperature thermal condition of the NAND device;
measuring a command queue of the NAND device to determine that a current workload exhibits a bursty workload profile, the bursty workload profile being a temporary rather than sustained increased change in activity for the NAND device;
suspending the high-temperature policy with respect to the current workload;
processing the current workload while the high-temperature policy is suspended; and
re-implementing the high-temperature policy for the NAND device in response to completing the current workload.
14. The method of claim 13 , wherein the high-temperature thermal condition is one of an ordered set of thermal conditions for the NAND device, members of the set of thermal conditions having respective operating policies for the NAND device.
15. The method of claim 14 , wherein the respective operating policies include throttling a resource of the NAND device, wherein a higher order in the set of thermal conditions corresponds to a greater degree of throttling.
16. The method of claim 15 , wherein throttling the resource includes powering-off a component of the NAND device, a greater degree of throttling corresponding to powering-off a greater number of components of the NAND device.
17. The method of claim 13 , wherein a bursty workload profile has a time-averaged command queue depth that is small in comparison to a depth of the command queue at a time that it includes commands.
18. The method of claim 17 , wherein measuring the command queue of the NAND device to determine that a current workload exhibits a bursty workload profile includes comparing a current rate of commands into the command queue to an averaged historical rate of commands into the command queue over time.