Dr. Matias, Thanks for the reply. The problems are quite simple:

(i) Users cannot handle the size of virtual memory in any way, even it's not effective to move the location of swap files to another storage drive.

(ii) Users cannot handle when OSX kernel kills a "not-safe" process by itself though we have no idea which one is safe or not.

I found only few contents from the web, which are looking it seems hopeless to get a gigantic virtual memory to handle a large data set in OSX:

http://r.789695.n4.nabble.com/R-process-killed-when-allocating-too-large-matrix-Mac-OS-X-td4720400.html

Best,

I got 64-bit OS (10.12.5) installed in mac pro with 32GB RAM and 4TB HDD.

I thought VRAM should work until the physical HDD is full, I observed the activity monitor, and the R process was killed when about a part of data was being loaded.

I only found the following two lines in the log:

default 06:14:27.024571 -0500 kernel low swap: killing largest compressed process with pid 673 (R) and size 188 MB

default 06:14:27.024593 -0500 kernel memorystatus_thread: idle exiting pid 69 [iconservicesagent]

Then, I have no idea on how to resolve this.

Thank you in advance for any suggestion.

Briefly, swap partition setting didn't figure this out. I'm using my institute's clustering server for peace in mind.

Thanks!