F. ERG UN, Testing multivariate linear functions: Overcoming the generator bottleneck,in Proceedings of the Twenty-Seventh Annual ACM Symposium on the Theory of Computing, Las Vegas, Nevada, jun 1995, pp. 407--416.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....similar results. I could not determine the exact source of this difference. Testing FFTW. FFTW uses different plans on each platform, and some codelets are not used at all on the machines available to me. How do we ensure that FFTW is correct FFTW uses the self testing algorithm by Funda Ergun [49], a randomized test that guarantees that a given program computes the DFT for an overwhelmingly large fraction of all possible inputs. The self tester does not require any other DFT program to be available. In the past, we checked FFTW against the program by Singleton [132] assuming that any bug ....

F. ERG UN, Testing multivariate linear functions: Overcoming the generator bottleneck,in Proceedings of the Twenty-Seventh Annual ACM Symposium on the Theory of Computing, Las Vegas, Nevada, jun 1995, pp. 407--416.

....similar results. I could not determine the exact source of this difference. Testing FFTW. FFTW uses different plans on each platform, and some codelets are not used at all on the machines available to me. How do we ensure that FFTW is correct FFTW uses the self testing algorithm by Funda Ergun [49], a randomized test that guarantees that a given program computes the DFT for an overwhelmingly large fraction of all possible inputs. The self tester does not require any other DFT program to be available. In the past, we checked FFTW against the program by Singleton [132] assuming that any bug ....

F. ERG UN, Testing multivariate linear functions: Overcoming the generator bottleneck, in Proceedings of the Twenty-Seventh Annual ACM Symposium on the Theory of Computing, Las Vegas, Nevada, jun 1995, pp. 407--416.

....real numbers on legal domain [ Gamma1; 1] Assume we have verified that P is correct on these random input vectors ( Sigma an allowable error delta) a version of our simple checker might be employed to facilitate this verification. Through such a testing stage (or by use of a self tester [Ergun 1995]) we can satisfy ourselves that (with high probability) the fraction of inputs on which P returns incorrect output is very small: say, 1 10;000;000 . Once we have this assurance, we can employ the following self corrector for P, whose time bound is two calls to P plus O(n) Algorithm 4. On ....

Erg un, F. 1995. Testing multivariate linear functions: overcoming the generator bottleneck. In Proc. 27th ACM Symp. Theory of Computing (1995), pp. 407--416. Extends self-testing to functions (e.g., the Fourier Transform) for which traditional self-testers prove inefficient. See also [Ravikumar and Sivakumar 1995].

.... programs that take P as an oracle, and in addition, take one or more of the following parameters as input: an accuracy parameter ffl that specifies the conditions that P is expected to This paper unifies the preliminary versions which appeared in the 27th Annual Symposium on Theory of Computing [10] and in the 15th Annual Foundations of Software Technology and Theoretical Computer Science [16] y Department of Computer Science, Cornell University, Ithaca, NY 14853 7501 (ergun cs.cornell.edu) This work is partially supported by ONR Young Investigator Award N00014 93 1 0590, the Alfred P. ....

F. Erg un, Testing multivariate linear functions: Overcoming the generator bottleneck, in Proc. 27th Annual ACM Symposium on the Theory of Computing, 1995, pp. 407--416.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC