Decision tree node instruction unification for parallel processing
An information processing device performs a decision tree based on a decision tree which has condition determination nodes and leaf nodes. In the information processing device, an instruction unification means generates a unified instruction by unifying an instruction, which each of the condition determination nodes included in the decision tree executes, to be suitable for a parallel processing. An acquisition means acquires a plurality of pieces of input data. A condition determination means performs, by the parallel processing, a condition determination with respect to the plurality of pieces of input data for each of the condition determination nodes.
1. An information processing device comprising:
a memory storing instructions; and
one or more processors configured to execute the instructions to:
convert each of instructions executed by all condition determination nodes included in a decision tree to a unified instruction;
acquire a plurality of pieces of input data; and
perform, by parallel processing, condition determinations with respect to the plurality of pieces of input data using the unified instruction for each of the condition determination nodes,
wherein the one or more processors convert each of the instructions to the unified instruction by:
converting each of the instructions executed by all the condition determination nodes to a comparison instruction which determines whether or not a determination target value in each of the condition determination nodes is equal to or greater than a first threshold value and is equal to or less than a second threshold value; and
determining the first threshold value and the second threshold value.
2. The information processing device according to claim 1 , wherein the one or more processors convert an instruction that determines whether or not the determination target value in one of the condition determination nodes is less than a predetermined value, into an instruction that determines whether or not the determination target value in the one of the condition determination nodes is equal to or greater than minus infinity and is equal to or less than an adjacent value.
3. The information processing device according to claim 2 , wherein the one or more processors convert an instruction that determines whether or not the determination target value in a second one of the condition determination nodes is greater than a predetermined value, into an instruction that determines whether or not the determination target value in the second one of the condition determination nodes is equal to or greater than an adjacent value being greater than the predetermined value and is equal to or less than plus infinity.
4. The information processing device according to claim 1 , wherein the one or more processors convert an instruction that determines whether or not the determination target value in one of the condition determination nodes matches a predetermined value, into an instruction that determines whether or not the determination target value in the one of the condition determination nodes is equal to or less than the predetermined value and is equal to or greater than the predetermined value.
5. The information processing device according to claim 1 , wherein the one or more processors convert an instruction that determines whether or not the determination target value in one of the condition determination nodes is equal to or less than a predetermined value, into an instruction that determines whether or not the determination target value in the one of the condition determination nodes is equal to or greater than minus infinity and is equal to or less than the predetermined value.
6. The information processing device according to claim 5 , wherein the one or more processors convert an instruction that determines whether or not the determination target value in a second one of the condition determination nodes is equal to or greater than a predetermined value, into an instruction that determines whether or not the determination target value in the second one of the condition determination nodes is equal to or greater than the predetermined value and is equal to or less than plus infinity.
7. The information processing device according to claim 1 , wherein the one or more processors perform parallel SIMD processing.
8. An information processing method performed by one or more processors and comprising:
converting each of instructions executed by all condition determination nodes included in a decision tree to a unified instruction;
acquiring a plurality of pieces of input data; and
performing, by parallel processing, condition determinations with respect to the plurality of pieces of input data using the unified instruction for each of the condition determination nodes,
wherein converting each of the instructions to the unified instruction comprises:
converting each of the instructions executed by all condition determination nodes to a comparison instruction which determines whether or not a determination target value in each of the condition determination nodes is equal to or greater than a first threshold value and is equal to or less than a second threshold value; and
determining the first threshold value and the second threshold value.
9. A non-transitory computer-readable recording medium storing a program causing a computer to perform a process comprising:
converting each of instructions executed by all condition determination nodes included in a decision tree to a unified instruction;
acquiring a plurality of pieces of input data; and
performing, by a parallel processing, condition determinations with respect to the plurality of pieces of input data using the unified instruction for each of the condition determination nodes,
wherein converting each of the instructions to the unified instruction comprises:
converting each of the instructions executed by all condition determination nodes to a comparison instruction which determines whether or not a determination target value in each of the condition determination nodes is equal to or greater than a first threshold value and is equal to or less than a second threshold value; and
determining the first threshold value and the second threshold value.