fastR slower than GNU R

[abhinaynagpal]

library('forecast')
data("AirPassengers")
time_func <- function() {
startTime <- Sys.time()
mymodel <- auto.arima(AirPassengers)
myforecast <- forecast(mymodel, level=c(95), h=10*12)
endTime <- Sys.time()
print(endTime - startTime)
}

time_func()

GNU R takes 1 second VS fastR : 3.9 seconds, any pointers how this can be optimized?

[abhinaynagpal]

I had also tried
/opt/java/graalvm/languages/R/bin/R --R.BackEnd=llvm

library('forecast')
Warning message:
Loading of '/root/R/x86_64-pc-linux-gnu-library/fastr-21.3.0-4.0/lmtest/libs/lmtest.so' in LLVM mode failed. You may load this package via the native mode by adding it to /opt/java/graalvm/languages/R/etc/native-packages or by running FastR with option --R.BackEndNative=packageName.
Error in polyglot evaluation : External LLVMFunction pan_ cannot be found.

[abh1nay]

On a related note will https://github.com/flang-compiler/flang integration with this help?