TCAM allocation method based on graph division

Abstract

The invention provides a TCAM distribution method based on graph division. The TCAM distribution method comprises the steps of converting a control flow chart of a program execution process into a condition variable chart; greedy selecting a next-level sub-graph root node from the alternative nodes of the current conditional variable graph according to the utility value until a final set of sub-graph root nodes is obtained; dividing the conditional variable graph into a plurality of final sub-graphs based on the set of root nodes of the final sub-graphs; and allocating TCAM to the program execution process according to the final sub-graph.

Description

technical field [0001] The invention relates to the field of computers, in particular to a method for allocating TCAM based on graph division. Background technique [0002] In recent years, with the emergence of many new technologies, the new generation memory market is heating up, and the chip architecture is developing along with the new market. Unlike traditional Static Random-Access Memory (SRAM), Ternary Content-Addressable Memory (TCAM) provides a method of searching large data sets using masks indicating ranges. method. It allows searching stored information by its content rather than retrieving memory cells by their physical address. In addition to the two states of "0" and "1", each byte in TCAM also has a third state called "other" (that is, "don't care"). It is this third state feature of TCAM that enables it to perform both exact match search and fuzzy match search, so its performance is far superior to that of SCAM, which can only perform exact match search. ...

AI Technical Summary

Problems solved by technology

[0004] However, the graph partition method based on the principle of multiplication and addition has the following problems: First, the method adopts a recursive partition strategy, which only considers how to reduce the scale of the graph, and the purpose of partitioning is only to reduce the graph. view, thus lacking a global scheduling strategy

This may reduce the size of the local subgraph, but the overall effect may be poor partition results

Secondly, due to the variety of hardware resource specifications, different subgraphs may require different termination conditions to limit, and it is difficult to use this method to guide partition optimization through indicators, which will further affect the partition results

In addition, the preconditions for cross-cutting are relatively strict, and a large number of vertical cuts are required to create conditions for cross-cutting, which may make the division results cross-redundant

Low-quality graph partition results will affect TCAM allocation results and compiler efficiency

Images