Memristor crossbar arrays to activate processors
In one example, a device to process analog sensor data is described. For example, a device may include at least one analog sensor to generate a first set of analog voltage signals and a crossbar array including a plurality of memristors. In one example, the crossbar array is to receive an input vector of the first set of analog voltage signals, generate an output vector comprising a second set of analog voltage signals that is based upon a dot product of the input vector and a matrix comprising resistance values of the plurality of memristors, detect a pattern of the output vector, and activate a processor upon a detection of the pattern.
1. A device, comprising:
at least one analog sensor to generate a first set of analog voltage signals;
a crossbar array comprising a plurality of memristors, the crossbar array configured to:
receive an input vector of the first set of analog voltage signals;
generate an output vector comprising a second set of analog voltage signals that is based upon a dot product of the input vector and a matrix comprising resistance values of the plurality of memristors, and
detect a pattern of the output vector; and
a processor configured to:
activate when at least a portion of the second set of analog voltage signals match a pattern,
map matrix values into programming signals to generate a third set of analog voltage signals based on the programming signals, and
provide the third set of analog voltage signals to reprogram the plurality of memristors of the crossbar array.
2. The device of claim 1 ,
wherein the processor comprises a digital microprocessor.
3. The device of claim 2 , wherein the processor is to remain off or in a low power state until the processor is activated by the crossbar array.
4. The device of claim 2 , further comprising:
a set of digital-to-analog converters coupled to the processor and to the crossbar array to output a third set of analog voltage signals to program the resistance values of the plurality of memristors.
5. The device of claim 1 , wherein the crossbar array comprises a set of sense circuits to convert output currents of a plurality of rows or a plurality of columns of the crossbar array to the second set of analog voltage signals.
6. The device of claim 1 , further comprising:
a set of analog buffers to sample and hold the first set of analog voltage signals and to feed the first set of analog voltage signals to the crossbar array.
7. The device of claim 1 , further comprising:
a set of analog-to-digital converters to convert the second set of analog voltage signals to a set of digital output signals.
8. The device of claim 1 , wherein the resistance values of the plurality of memristors are to cause the crossbar array to function as a neural network.
9. The device of claim 1 , wherein at least one of the second set of analog voltage signals of the output vector is to function as an activation signal to activate the processor.
10. A system, comprising:
at least one analog sensor to generate a first set of analog voltage signals;
a crossbar array comprising a plurality of memristors, the crossbar array configured to:
generate a second set of analog voltage signals that is based upon the first set of analog voltage signals and resistance values of the plurality of memristors;
a processor configured to:
activate when at least a portion of the second set of analog voltage signals match a pattern, and
map matrix values into programming signals to generate a third set of analog voltage signals based on the programming signals; and
a programming unit to determine the resistance values of the plurality of memristors.
11. The system of claim 10 , further comprising:
a wireless communication interface for communications between the programming unit and the processor.
12. The system of claim 10 , wherein the programming unit comprises a remote network-based component.
13. The system of claim 10 , wherein the third set of analog voltage signals are used to by the programming unit to program the resistance values of the plurality of memristors.
14. A device, comprising:
a processor; and
a non-transitory computer-readable medium storing instructions which, when executed by the processor, cause the processor to:
determine target resistance values for a plurality of memristors of a crossbar array to represent a neural network for processing analog sensor data to detect a pattern;
provide a first set of voltage signals to program the plurality of memristors of the crossbar array, the first set of voltage signals based upon the target resistance values;
determine an accuracy of the crossbar array in detecting the pattern in the analog sensor data;
provide a second set of voltage signals to reprogram the plurality of memristors to improve the accuracy of the crossbar array in detecting the pattern in the analog sensor data;
map matrix values into programming signals to generate a third set of voltage signals based on the programming signals; and
provide the third set of voltage signals to reprogram the plurality of memristors of the crossbar array.
15. The device of claim 14 , wherein the instructions, when executed by the processor, further cause the processor to:
inject a test input vector to the crossbar array;
read an output vector from the crossbar array; and
determine a deviation of the output vector from an expected output vector.
16. The device of claim 14 , wherein the instructions, when executed by the processor, further cause the processor to:
apply the first set of voltage signals to the plurality of memristors in the crossbar array to program resistances of the plurality of memristors.