Computation reuse for loops with irregular accesses
A compiler selects a nested loop within software code that includes an outer loop and an inner loop. The outer loop includes an outer induction variable and the inner loop includes an inner induction variable. The compiler identifies a computation included in the nested loop that generates an irregular array access, which includes an expression of both the outer induction variable and the inner induction variable. Next, the compiler identifies a redundant calculation for the computation based upon the outer induction variable and the inner induction variable, and generates a temporary variable to correspond with the redundant calculation. The compiler replaces the computation with the temporary variable in the nested loop and, in turn, compiles the nested loop with the included temporary variable.
1. A computer-implemented method comprising:
selecting a nested loop within software code that includes an outer loop and an inner loop, the outer loop including an outer induction variable and the inner loop including an inner induction variable;
identifying, by one or more processors, a computation included in the nested loop that generates an irregular array access, wherein the irregular array access includes an expression of both the outer induction variable and the inner induction variable;
identifying, by one or more of the processors, a redundant calculation for the computation based upon the outer induction variable and the inner induction variable;
generating, by one or more of the processors, a temporary variable that corresponds to the identified redundant calculation;
replacing, by one or more of the processors, the computation with the temporary variable in the nested loop; and
compiling the nested loop that includes the temporary variable.
2. The method of claim 1 further comprising:
identifying a first operand and a second operand included in the computation; and
determining that the computation is reusable based upon:
detecting that the first operand generates the irregular array access; and
detecting that the second operand is loop invariant with respect to the outer loop and the inner loop.
3. The method of claim 1 further comprising:
identifying a first operand and a second operand included in the computation; and
determining that the computation is reusable based upon:
detecting that the first operand generates the irregular array access; and
detecting that the second operand generates a subsequent irregular array access having similar array access coefficients as the first operand.
4. The method of claim 1 further comprising:
identifying a first operand and a second operand included in the computation; and
determining that the computation is legal based upon:
determining that the first operand and the second operand fail to be a dependence target; and
determining that an outer variable range corresponding to the outer induction variable and an inner variable range corresponding to the inner induction variable are finite.
5. The method of claim 1 further comprising:
computing a reuse factor for the computation, wherein the reuse factor includes a non-optimized calculations quantity compared with an optimized calculations quantity; and
determining that reusing the redundant calculation is profitable based upon comparing the reuse factor with a reuse threshold.
6. The method of claim 1 further comprising:
including a temporary variable computation in the software code, the temporary variable computation residing external to the nested loop and generating a computation result for the temporary variable; and
compiling the software code with the included temporary variable computation.
7. The method of claim 1 wherein the nested loop resides within an n-nested loop.
8. A system comprising:
one or more processors;
a memory accessible by at least one of the processors;
a set of instructions stored in the memory and executed by at least one of the processors in order to perform actions of:
selecting a nested loop within software code that includes an outer loop and an inner loop, the outer loop including an outer induction variable and the inner loop including an inner induction variable;
identifying a computation included in the nested loop that generates an irregular array access, wherein the irregular array access includes an expression of both the outer induction variable and the inner induction variable;
identifying a redundant calculation for the computation based upon the outer induction variable and the inner induction variable;
generating a temporary variable that corresponds to the identified redundant calculation;
replacing the computation with the temporary variable in the nested loop; and
compiling the nested loop that includes the temporary variable.
9. The system of claim 8 wherein the set of instructions when executed by one or more of the processors further performs actions of:
identifying a first operand and a second operand included in the computation; and
determining that the computation is reusable based upon:
detecting that the first operand generates the irregular array access; and
detecting that the second operand is loop invariant with respect to the outer loop and the inner loop.
10. The system of claim 8 wherein the set of instructions when executed by one or more of the processors further performs actions of:
identifying a first operand and a second operand included in the computation; and
determining that the computation is reusable based upon:
detecting that the first operand generates the irregular array access; and
detecting that the second operand generates a subsequent irregular array access having similar array access coefficients as the first operand.
11. The system of claim 8 wherein the set of instructions when executed by one or more of the processors further performs actions of:
identifying a first operand and a second operand included in the computation; and
determining that the computation is legal based upon:
determining that the first operand and the second operand fail to be a dependence target; and
determining that an outer variable range corresponding to the outer induction variable and an inner variable range corresponding to the inner induction variable are finite.
12. The system of claim 8 wherein the set of instructions when executed by one or more of the processors further performs actions of:
computing a reuse factor for the computation, wherein the reuse factor includes a non-optimized calculations quantity compared with an optimized calculations quantity; and
determining that reusing the redundant calculation is profitable based upon comparing the reuse factor with a reuse threshold.
13. The system of claim 8 wherein the set of instructions when executed by one or more of the processors further performs actions of:
including a temporary variable computation in the software code, the temporary variable computation residing external to the nested loop and generating a computation result for the temporary variable; and
compiling the software code with the included temporary variable computation.
14. The system of claim 8 wherein the nested loop resides within an n-nested loop.
15. A computer program product stored in a computer readable storage medium, comprising functional descriptive material that, when executed by an information handling system, causes the information handling system to perform actions that include:
selecting a nested loop within software code that includes an outer loop and an inner loop, the outer loop including an outer induction variable and the inner loop including an inner induction variable;
identifying a computation included in the nested loop that generates an irregular array access, wherein the irregular array access includes an expression of both the outer induction variable and the inner induction variable;
identifying a redundant calculation for the computation based upon the outer induction variable and the inner induction variable;
generating a temporary variable that corresponds to the identified redundant calculation;
replacing the computation with the temporary variable in the nested loop; and
compiling the nested loop that includes the temporary variable.
16. The computer program product of claim 15 comprising functional descriptive material that, when executed by the information handling system, causes the information handling system to further performs actions of:
identifying a first operand and a second operand included in the computation; and
determining that the computation is reusable based upon:
detecting that the first operand generates the irregular array access; and
detecting that the second operand is loop invariant with respect to the outer loop and the inner loop.
17. The computer program product of claim 15 comprising functional descriptive material that, when executed by the information handling system, causes the information handling system to further performs actions of:
identifying a first operand and a second operand included in the computation; and
determining that the computation is reusable based upon:
detecting that the first operand generates the irregular array access; and
detecting that the second operand generates a subsequent irregular array access having similar array access coefficients as the first operand.
18. The computer program product of claim 15 comprising functional descriptive material that, when executed by the information handling system, causes the information handling system to further performs actions of:
identifying a first operand and a second operand included in the computation; and
determining that the computation is legal based upon:
determining that the first operand and the second operand fail to be a dependence target; and
determining that an outer variable range corresponding to the outer induction variable and an inner variable range corresponding to the inner induction variable are finite.
19. The computer program product of claim 15 comprising functional descriptive material that, when executed by the information handling system, causes the information handling system to further performs actions of:
computing a reuse factor for the computation, wherein the reuse factor includes a non-optimized calculations quantity compared with an optimized calculations quantity; and
determining that reusing the redundant calculation is profitable based upon comparing the reuse factor with a reuse threshold.
20. The computer program product of claim 15 comprising functional descriptive material that, when executed by the information handling system, causes the information handling system to further performs actions of:
including a temporary variable computation in the software code, the temporary variable computation residing external to the nested loop and generating a computation result for the temporary variable; and
compiling the software code with the included temporary variable computation.