A global data address allocation method and related device

Abstract

An embodiment of the invention discloses a method and a relevant device for allocating addresses to global data. The method for allocating the addresses to the global data includes grouping the global data in source programs; allocating the corresponding addresses to the global data in each group, allocating the global data in each group in the same cache line and aligning initial positions of the global data in each group to an initial position of the corresponding cache line; allocating the corresponding addresses to ungrouped global data if the ungrouped global data are available. All the global data in the same group meet conditions that all the global data in the same group are used in the same basic block, the sum of the data sizes of all the global data in the same group is not larger than the size of the corresponding cache line of a computer system, and the total execution number of the same basic block is larger than a preset total execution number. According to the technical scheme, the method and the relevant device have the advantage that the probability of cache missing due to access to global data can be effectively reduced.

Description

technical field [0001] The invention relates to the field of computer science and technology, in particular to a global data address allocation method and a related device. Background technique [0002] In the field of computer science and technology, how to correctly and effectively convert the source program written by the programmer into a file that can be executed on the computer has always been one of the hot spots of research. At present, the generation of executable files from the source program is uniformly divided into three stages, that is, the source program is processed by a compiler, an assembler, and a linker successively. The processing process consisting of the above three stages needs to complete two tasks: one is to convert the high-level language code in the source program into a computer-recognizable instruction sequence; the other is to allocate a memory address for the data used in the source program to store the data. Among the data involved in the s...

AI Technical Summary

Problems solved by technology

When the executable file executes an instruction that uses global data a, the central processing unit (CPU, Central Processing Unit) issues an instruction to put global data a into a register. Since this is the first access to global data a, therefore The global data a is not in the cache, causing the first cache miss

After the instruction using global data a is executed, global data a and several adjacent data will be placed in a cache line (ie cacheline) of the cache. If the instruction using global data b is executed at this time, it will be the same as The above process is similar, since the global data b is not in the cache, the instruction using the global data b will cause the second cache miss

Images