R module Prediction generates an error after Parglm

Question

My runs are taking several hours, and I would like to take advantage of the multiprocessor capabilities of the R module PARGLM in conjunction with the PREDICTION module, but the PREDICTION module generates the following error after a PARGLM call. It runs fine after the GLM call.

Error in qr.R(qr.lm(object)) : argument is not a QR decomposition

Is there a way around this?

Thanks.

Bob

Toy example below:

library(parglm)

library(prediction)

library(sandwich)

a <- structure(list(Family = c(1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 3, 
                          4, 4, 4, 4, 4, 2), 
               Sex = c("F", "F", "F", "M", "M", "F", "F", 
                        "M", "M", "M", "F", "M", "F", "F", "M", "M", "M", "F"),
               Height = c(67, 66, 64, 71, 72, 63, 67, 69, 68, 70, 63, 64, 67, 66, 67, 69, 67, 63)),
               row.names = c(NA, 18L), class = "data.frame") 

**Call to GLM**

fit <-       glm(Height~Sex, 
             family=Gamma(link="log"),
             data=a
)

v<-vcovCL(fit,cluster=a$Family)
summary(prediction(fit,data=a,vcov=v),type = "link")

Prediction SE z p lower upper 66.83 0.6002 111.4 0 65.66 68.01

**Call to PARGLM** 

fit <-       parglm(Height~Sex, 
                 data=a, family=Gamma(link="log"), cluster="Family",
                 control = parglm.control(nthreads = 1L)
                
                 
)

v<-vcovCL(fit,cluster=a$Family)
summary(prediction(fit,data=a,vcov=v),type = "link")

Error in qr.R(qr.lm(object)) : argument is not a QR decomposition