Accelerating decision tree inferences

A computer program product for accelerating machine learning inferences, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by processing means to cause the processing means to:

• set, in a memory of a computerized system, a memory buffer that includes contiguous memory blocks;
• obtain a decision tree that comprises nodes including split nodes and leaf nodes, wherein the split nodes include at least two child nodes ordered according to a likelihood of accessing one of the at least two child nodes after an included split node;
• and
• map the nodes of the decision tree onto respective blocks of the memory blocks, the memory blocks storing attributes of the mapped nodes, wherein the mapping is performed according to a depth-first search order, wherein the split nodes and included child nodes are mapped onto successive blocks of the memory blocks, wherein the ordered child nodes an included split node are mapped onto successive blocks of the memory blocks.

A computer system comprising:

• a processor set;
• and
• a computer readable storage medium;
• wherein:
• the processor set is structured, located, connected and programmed to run program instructions stored on the computer readable storage medium;
• and
• the stored program instructions include:
• program instructions programmed to set a memory buffer that includes contiguous memory blocks;
• program instructions programmed to obtain a decision tree that comprises nodes including split nodes and leaf nodes, wherein the split nodes include at least two child nodes ordered according to a likelihood of accessing one of the at least two child nodes after an included split node;
• and
• program instructions programmed to map the nodes of the decision tree onto respective blocks of the memory blocks, the memory blocks storing attributes of the mapped nodes, wherein the mapping is performed according to a depth-first search order, wherein the split nodes and included child nodes are mapped onto successive blocks of the memory blocks, wherein the ordered child nodes an included split node are mapped onto successive blocks of the memory blocks;
• program instructions programmed to execute the nodes of the decision tree by processing the attributes of the nodes as accessed from the memory according to an order of the memory blocks in the memory buffer;
• and
• program instructions programmed to return an inference result based on an outcome of executing the nodes.