OpenMx Developer Forums

At the Developer's Meeting on 1/6/12 we discussed the following:

• Upcoming grant proposals to fund research related to OpenMx.
• Mike Neale is experiencing an issue with the parallelization in the current version of OpenMx 1.2. There appears to be an issue with PBS that allows his 24-core sever to experience an average load of 144 when running certain models. The group believes this can be fixed by reading and setting several PBS environment variables.

At the OpenMx Developers meeting on 12/30/11 we discussed the following:

• The grant proposal to fund research related to OpenMx.
• The group discussed releasing a new beta release of OpenMx 1.2 containing the the parallel ordinal FIML performance regression bug fix.
• The group discussed beginning to diagnos the issue Mike Cheung came across in OpenMx 1.2 fitting a model with one binary variable and one latent variable.

At the OpenMx Developers meeting on 12/16/11 we discussed the following:

• The content for a grant(s) to fund research related to OpenMx.

At the OpenMx meeting on 12/9/11 we discussed the following:

• The status of the grant proposal. Our strategy is to attempt to address any possible concerns or criticisms a reviewer may have. In particular liability concerns were discussed. The group believes that the grant can be phrased with in such a way that what we our proposing is seen in the same light (liability wise) as existing medical equipment such as MRI machines, etc. Framing goals for the grant in terms of a decision tree to address possible issues that become intractable was also discussed.

At developers meeting on 12/2 we discussed the following:

• The group discussed several funding opportunities. In particular opportunities in industry and the Dept. of Defense were discussed. It was pointed out that mentioning on a grant application that the work was being considered for corporate funding could be an advantage for more traditional funding sources like NIH and NSF.

At developers meeting on 11/4 we discussed the following:

• The group discussed Dan Hackett's latest progress with PPML. Dan has achieved some dramatic speedups (up to 30x speedup) for models with fixed structures that allow an analytical solution to be computed for the error matrices and in some cases the covariance matrices.
• The group discussed the changes that would need to be made to the current implementation to adapt PPML to handle ordinal data. Parallelizing several aspects of PPML was also discussed.

Starting in OpenMx 1.2, we will have support for multicore execution of the FIML objective function, and for the numerical calculation of the Hessian that is used to compute standard error estimates. A mechanism is needed to specify the number of threads to be assigned to the computations.

We currently have an option available under ?mxOption() that is named "Number of Cores". I'd like to rename the option to "Number of Threads" if there are no objections. And I'd like to propose the following behavior for the value (N) that is stored in this option.

At developers meeting on 11/4 we discussed the following:

• The source of the standard error issue identified by Timo has been discovered. The source of the error is a precision issue. OpenMx's numerical solution to calculating the hessian was (0.02) while the analytical solution was (-.04). Because the numerical solution crossed the zero boundary the numerical solution was significantly off. The group discussed implementing an analytical hessian calculation to address this.

At developers meeting on 10/27 we discussed the following:

• The group discussed the state of the effort to run OpenMx on the CRAN platform. Mike Wilde and Tim Armstrong are currently addressing some straight forward issues to get OpenMx to run.
• Michael Spiegel and Tim Armstrong have both observed a bottle neck of serializing and unserializing data. Spiegel has submitted a patch to the R Developers' group to address this bottleneck. Spiegel has seen a 2.4x benchmark on a micro-benchmark that just serializes and unserializes data. He has seen a 10% on a real benchmark.

I tried to run the script given in:

http://openmx.psyc.virginia.edu/docs/OpenMx/latest/FIML_RowObjective.html

It seems to fail with the following error:

Running FIML BivCor
Error: Unknown reference 'filteredDataRow' detected in the entity 'secondHalfCalc' in model 'FIML BivCor'

am i doing something wrong? i merely copy pasted it.

Michel