Data channel optimization with smart black box algorithms
A system includes a host configured to communicate with a device under test. The host is configured to write test data to the device under test. An optimization engine is configured to optimize a plurality of parameters associated with a magnetic recording channel associated with the device under test. The optimization engine is configured to select a first set of parameters for the plurality of parameters and the host is configured to set the magnetic recording channel based on the first set of parameters. The host then measures the performance of the magnetic recording channel based on the first set of parameters. Based on the measured performance, the optimization engine then selects new parameter values for the plurality of parameters. Until the measured performance is within an acceptable threshold, the optimization engine will iteratively update the plurality of parameters based on the measured performance.
1. A method comprising:
selecting parameters associated with a second algorithm using a first algorithm;
selecting an initial set of parameters for a plurality of parameters associated with a magnetic recording channel using the second algorithm, wherein the second algorithm is configured to simultaneously optimize the plurality of parameters associated with the magnetic recording channel, the simultaneous optimization being part of an iterative process comprising:
(a) performing a write operation to write test data, the write operation being performed with the magnetic recording channel having the plurality of parameters;
(b) performing a read operation that comprises reading the test data, the read operation being performed with the magnetic recording channel having the plurality of parameters;
(c) calculating a performance of the magnetic recording channel for the read operation, an updating the plurality parameters if the calculated performance does not meet a certain condition,
wherein (a), (b) and (c) are repeated until the certain condition is met;
responsive to the certain condition being met, optimizing the parameters associated with the second algorithm using the first algorithm;
responsive to the optimizing the parameters associated with the second algorithm using the first algorithm, repeating (a), (b) and (c) until the certain condition is met; and
responsive to the certain condition being met, further optimizing the parameters associated with the second algorithm using the first algorithm.
2. The method of claim 1 , wherein the selecting the parameters associated with the second algorithm and wherein the selecting the initial set of parameters associated with the magnetic recording channel is random.
3. The method of claim 1 , wherein the first algorithm is an upper confidence bound (UCB) algorithm.
4. The method of claim 1 , wherein the second algorithm is selected from a group consisting of the Truncated Newton (TN) method, Sequential Least Squares Programming (SLSP) and Limited Memory version of Broyden-Fletcher-Goldfarb-Shanno (L-BFGS-B) algorithm.
5. The method of claim 1 , wherein the performance is measured by bit error rate (BER).
6. The method of claim 1 , wherein the certain condition is a bit error rate (BER) being less than a predetermined threshold.
7. The method of claim 1 , wherein the magnetic recording channel comprises an analog front end and a digital front end.
8. The method of claim 7 , wherein the analog front end comprises A/C coupling filter, a time gain recovery unit, an asymmetric correction, and a low pass filter.
9. The method of claim 7 , wherein the digital front end comprises a FIR filter equalizer, a synchronization mark detector, and a time-gain error generator.
10. A system comprising:
a host configured to communicate with a device under test, wherein the host is further configured to write test data to the device under test; and
an optimization engine configured to simultaneously optimize a plurality of parameters associated with a magnetic recording channel associated with the device under test,
wherein the optimization engine is configured select a first set of parameters for the plurality of parameters, and wherein the host is configured to set the magnetic recording channel based on the first set of parameters for the plurality of parameters,
wherein the host is configured to measure performance of the magnetic recording channel based on the first set of parameters, and
wherein the optimization engine is further configured to determine whether the measured performance with the magnetic recording channel set based on the first set of parameters is within a predetermined acceptable threshold,
and wherein, for a condition where the optimization engine determines that the measured performance with the magnetic recording channel set based on the first set of parameters is not within the predetermined acceptable threshold;
(a) the optimization engine is further configured to update the plurality of parameters,
(b) the host is configured to set the magnetic recording channel based on the updated plurality of parameters and configured to the measure performance of the magnetic recording channel based on the updated plurality of parameters, and
(c) the optimization engine is further configured to determine whether the measured performance with the magnetic recording channel set based on the updated plurality of parameter is within the predetermined acceptable threshold, and
wherein the optimization engine and the host are configured to perform (a), (b) and (c) iteratively until the measured performance is within the predetermined acceptable threshold.
11. The system of claim 10 , wherein the optimization engine comprises a first algorithm and a second algorithm, wherein parameters of the second algorithm are selected using the first algorithm and wherein the parameters of the second algorithm are optimized iteratively, and wherein the second algorithm is configured to simultaneously optimize the plurality of parameters associated with the magnetic recording channel.
12. The system of claim 11 , wherein the first algorithm is an upper confidence bound (UCB) algorithm.
13. The system of claim 11 , wherein the second algorithm is selected from a group consisting of the Truncated Newton (TN) method, Sequential Least Squares Programming (SLSP) and Limited Memory version of Broyden-Fletcher-Goldfarb-Shanno (L-BFGS-B) algorithm.
14. The system of claim 10 , wherein the performance is measured by bit error rate (BER).
15. The system of claim 10 , wherein the magnetic recording channel comprises an analog front end and a digital front end.
16. The system of claim 15 , wherein the analog front end comprises A/C coupling filter, a time gain recovery unit, an asymmetric correction, and a low pass filter.
17. The system of claim 15 , wherein the digital front end comprises a FIR filter equalizer, a synchronization mark detector, and a time-gain error generator.
18. A system comprising:
a host configured to communicate with a device under test, wherein the host is further configured to write test data to the device under test; and
an optimization engine comprising a first algorithm and a second algorithm, wherein the first algorithm is configured to perform first optimization operations in which parameters associated with the second algorithm are optimized, and wherein the second algorithm is configured to perform second optimization operations in which a plurality of parameters associated with a magnetic recording channel associated with the device under test are simultaneously optimized, and
wherein the first optimization operations and the second optimization operations are configured to be performed iteratively and responsive to performance measurements.
19. The system of claim 18 , wherein the first algorithm and the second algorithm are machine learning based algorithms.