P.D. Coddington, Tests of random number generators using Ising model simulations, in Proc. of the  Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Random Number Generators for Parallel Applications - Srinivasan, Ceperley, al. (1998) (Correct)
....quite differently in each algorithm. Thus this application is very popular in testing random number generators and has often detected defects in generators. We can test parallel generators on the Ising model application by assigning distinct random number sequences to subsets of lattice sites [37]. How good are the popular generators on the Ising model Ferrenberg, et al. [3] found that certain generators such as the particular shift register generator they used (R250) failed with the Wolff algorithm, while the simple and much maligned 32 bit LCG called CONG did well. Similar defects in ....
....now give some test results for the LCG with (parameterized) prime addend and a modified version of the LFG. Both of these generators performed acceptably in the sequential tests with 10 11 random numbers. Preliminary results from other tests of PPRNG can also be found in the paper by Coddington [37]. We first tested the generators by interleaving about 1000 pairs of random sequences each containing about 10 8 PRNs. Both the generators passed this tests. Next we simulated parallelism on the 16 Theta 16 Ising model by using a different random sequence for each lattice site in the Metropolis ....
P. Coddington. Tests of random number generators using Ising model simulations. Int. J. of Mod. Phys. "C", 7(3):295--303, 1996.
Testing Parallel Random Number Generators - Ashok Srinivasan (Correct)
....algorithms, such as the Metropolis and the Wolff algorithms, that can be used to simulate the Ising model, and the random numbers enter quite differently in each algorithm. Thus this application is very popular in testing random number generators and has often detected defects in generators [8,15]. We can test parallel generators on the Ising model application by assigning distinct random number sequences to different subsets of lattice sites [15] In our parallel tests we use a different stream for each site. 3 Test Results We next describe the results of our tests. We first show that ....
....in each algorithm. Thus this application is very popular in testing random number generators and has often detected defects in generators [8,15] We can test parallel generators on the Ising model application by assigning distinct random number sequences to different subsets of lattice sites [15]. In our parallel tests we use a different stream for each site. 3 Test Results We next describe the results of our tests. We first show that the popular 32 bit Unix LCG, rand fails even sequential RNG tests. For example, the collisions test fails for the following parameters: d = 4, k = 5, n = 2 ....
P. Coddington. Tests of random number generators using Ising model simulations. Int. J. of Mod. Phys. "C", 7(3):295--303, 1996.
Random Number Generators for Parallel Computers - Coddington (1997) (10 citations) Self-citation (Coddington) (Correct)
....parallel random number generators is to use them with parallel implementations of the Monte Carlo algorithms used for simulating the two dimensional Ising model, which have proven to be very effective at testing sequential generators. Such tests have been done for a number of parallel generators [23, 24]. 3 Random Number Generators In this section we introduce the most common algorithms used in random number generators the linear congruential, lagged Fibonacci, and shift register generators, and techniques which combine two or more of these generators. More detailed information can be found ....
....numbers (the lower order bits were not even uniformly distributed) due to a poor initialization of the seed tables on each processor, which left them highly correlated. This was greatly improved in a later release, although the newer version still failed a Monte Carlo Ising model test [23, 24], presumably because the new method for initializing the generator is still introducing some correlations between the seed tables on each processor. As with sequence splitting, just because the sequences on each processor do not overlap, does not necessarily mean they are uncorrelated. However in ....
[Article contains additional citation context not shown here]
P.D. Coddington, Tests of random number generators using Ising model simulations, in Proc. of the
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