Bounds-checking elimination

In computer science, bounds-checking elimination is a compiler optimization useful in programming languages or runtimes that enforce bounds checking, the practice of checking every index into an array to verify that the index is within the defined valid range of indexes. Its goal is to detect which of these indexing operations do not need to be validated at runtime, and eliminating those checks.

One common example is accessing an array element, modifying it, and storing the modified value in the same array at the same location. Normally, this example would result in a bounds check when the element is read from the array and a second bounds check when the modified element is stored using the same array index. Bounds-checking elimination could eliminate the second check if the compiler or runtime can determine that neither the array size nor the index could change between the two array operations.

Another example occurs when a programmer loops over the elements of the array, and the loop condition guarantees that the index is within the bounds of the array. It may be difficult to detect that the programmer's manual check renders the automatic check redundant. However, it may still be possible for the compiler or runtime to perform proper bounds-checking elimination in this case.

Implementations

In natively compiled languages

One technique for bounds-checking elimination is to use a typed static single assignment form representation and for each array to create a new type representing a safe index for that particular array. The first use of a value as an array index results in a runtime type cast (and appropriate check), but subsequently the safe index value can be used without a type cast, without sacrificing correctness or safety.

In JIT-compiled languages

Just-in-time compiled languages such as Java often check indexes at runtime before accessing arrays. Some just-in-time compilers such as HotSpot are able to eliminate some of these checks if they discover that the index is always within the correct range, or if an earlier check would have already thrown an exception.^[1]^[2]

References

↑ Kawaguchi, Kohsuke (2008-03-30). "Deep dive into assembly code from Java". Retrieved 2008-04-02.
↑ "Fast, Effective Code Generation in a Just-In-Time Java Compiler" (PDF). Intel Corporation. Retrieved 2007-06-22.

External links

W. Amme, J. von Ronne, M. Franz. Using the SafeTSA Representation to Boost the Performance of an Existing Java Virtual Machine (2002).

Compiler optimizations

Basic block	Peephole optimization

Loop optimization	Induction variable Strength reduction Loop fusion Loop inversion Loop interchange Loop-invariant code motion Loop nest optimization Loop unrolling Loop splitting Loop unswitching Software pipelining Automatic parallelization

Data-flow analysis	Common subexpression elimination Constant folding Induction variable recognition and elimination Dead store elimination Use-define chain Live variable analysis Available expression

SSA-based	Global value numbering Sparse conditional constant propagation

Code generation	Register allocation Instruction selection Instruction scheduling Rematerialization

Functional	Tail call elimination Deforestation

Global	Interprocedural optimization

Other	Bounds-checking elimination Dead code elimination Inline expansion Jump threading

Static analysis	Alias analysis Pointer analysis Shape analysis Escape analysis Array access analysis Dependence analysis Control flow analysis Data flow analysis

This article is issued from Wikipedia - version of the 9/15/2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.