Data storage device managing low endurance semiconductor memory write cache
AssigneeWestern Digital Technologies, Inc.
InventorDavid R. Hall
Life of this patent
Estimating the term…
Abstract
A data storage device is disclosed comprising a head actuated over a disk, a first semiconductor memory (SM) having a first endurance, and a second SM having a second endurance lower than the first endurance. A write command is received from a host including write data. When a size of the write command is less than a threshold, the write data is stored in a first SM write cache in the first SM, and when the size of the write command is greater than the threshold, the write data is stored in a second SM write cache in the second SM.
Claims19
A data storage device comprising:
- a disk;
- a head actuated over the disk;
- a first semiconductor memory (SM) having a first endurance, wherein the first SM is distinct from the disk;
- a second SM having a second endurance lower than the first endurance, wherein the second SM is distinct from the disk and the first SM, and wherein the second endurance is based on a number of times the second SM is erased, and the second SM has a block size for garbage collection and erasure;
- and
- control circuitry configured to:
- receive a write command including write data from a host;
- determine a size of write data of the write command;
- when a determined size of the write command is less than a threshold that is based on the block size for garbage collection and erasure of the second SM, store, the write data in a first SM write cache in the first SM;
- and
- when the determined size of the write command is greater than the threshold, store the write data in a second SM write cache in the second SM.
A data storage device comprising:
- a disk;
- a head actuated over the disk;
- a semiconductor memory (SM), wherein the SM is distinct from the disk;
- a non-volatile semiconductor memory (NVSM), wherein the NVSM is distinct from the disk and the SM;
- and
- control circuitry configured to:
- receive a plurality of write commands from a host, wherein each write command includes write data;
- determine a size of write data of the write commands;
- store, based upon the determining, the write commands in a command queue having a command queue depth;
- store flushing, based upon the determining, at least some of the write data in a NVSM write cache;
- and
- throttle use of the NVSM write cache based on at least one of the command queue depth, an average size of the write data, or a write transfer rate of the disk.
A data storage device comprising:
- a disk;
- a head actuated over the disk;
- a semiconductor memory (SM), wherein the SM is distinct from the disk;
- a non-volatile semiconductor (NVSM) comprising a plurality of blocks, wherein the NVSM is distinct from the disk and the SM;
- and
- control circuitry configured to:
- receive a plurality of write commands from a host, wherein each write command includes write data;
- determine a size of write data of the write commands;
- store, based upon the determining, at least some of the write data in a NVSM write cache in the NVSM;
- and
- generate a command execution order for the write commands based on whether flushing the write data of each write command will enable erasing one or more blocks of the NVSM.
A data storage device comprising:
- a disk comprising a plurality of data tracks;
- a head actuated over the disk;
- a first semiconductor memory (SM) having a first endurance, wherein the first SM is distinct from the disk;
- a second semiconductor memory (SM) a second endurance lower than the first endurance, wherein the second SM is distinct from the disk and the first SM;
- and
- control circuitry configured to:
- receive a plurality of write commands from a host, wherein each write command includes write data;
- determine a size of write data of the write commands;
- store, based upon the determining, at least some of the write data in a write cache;
- and
- generate a command execution order for flushing cached write data to the disk, based on a refresh metric of victim data tracks adjacent to data tracks of the disk to which the cached write data are to be written.
Dependencies are inferred from the claim text.