This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Robert, It sounds like the process of -O2 uses a lot of resources when compiling which ran you out of resources for that routine. Its possible that you could compile the rest of GEMPAK with -O2, but I have seen problems with OSF/1 where input pointer and output pointer was the same variable, and high levels of optimization collapse those argument lists. Glad you have now moved on though, Steve >From: Robert Mullenax <address@hidden> >Organization: UCAR/Unidata >Keywords: 200306101536.h5AFa5Ld002607 >-O2 was the problem. -O works fine. > >> <br /> >> Robert,<br /> >> <br /> >> Seems like your compiler is out of space (are you redefining <br /> >> larger arrays?). You didn't mention what version of gcc/g77 you <br /> >> were using, or if you have changed the array sizes from<br /> >> what I tested here. The routine you mention is rather simple,<br /> >> and the arrays aren't all that big in the grphgd.cmn, so maybe<br /> >> its the optimization level you are using. Also, you probably need<br /> >> to check for any patches for RH 8.0.<br /> >> <br /> >> I have built with RH 7.3 and RH 8.0 without problems<br /> >> <br /> >> gcc version 2.96 20000731 (Red Hat Linux 7.3 2.96-113)<br /> >> gcc version 3.2 20020903 (Red Hat Linux 8.0 3.2-7)<br /> >> <br /> >> <br /> >> Steve Chiswell<br /> >> <br /> >> <br /> >> >From: Robert Mullenax <address@hidden><br /> >> >Organization: UCAR/Unidata<br /> >> >Keywords: 200306092054.h59Ks0Ld008396<br /> >> <br /> >> >Steve, We got some new Dells here at NMSU/NSBF. I had intended to install > So<br /> >> > laris on them,<br /> >> >but it will not install, even though it installs on another new Dell with s > imi<br /> >> > lar<br /> >> >specs..but not quite the same.<br /> >> ><br /> >> >Anyway I stuck with RH 8.0 which came with the box (after a fresh install). > I<br /> >> > can't compile<br /> >> >all of GEMPAK 5.6J. I get the error below, which later leads to a failed b > uil<br /> >> > d of other things<br /> >> >including nmap2.<br /> >> ><br /> >> >Any ideas?<br /> >> ><br /> >> >Thanks,<br /> >> >Robert Mullenax<br /> >> ><br /> >> >gweigs.f: In subroutine `gweigs':<br /> >> >gweigs.f:144: unable to find a register to spill in class `FLOAT_REGS'<br / > > >> >gweigs.f:144: this is the insn:<br /> >> >(insn 677 675 687 (set (subreg:SF (reg/v:DI 29 rmm0 [99]) 0)<br /> >> > (plus:SF (subreg:SF (reg/v:DI 29 rmm0 [99]) 0)<br /> >> > (reg:SF 8 st(0) [228]))) 531 {*fop_sf_comm_nosse} (insn_list 67 > 4 (<br /> >> > nil))<br /> >> > (expr_list:REG_DEAD (reg:SF 8 st(0) [228])<br /> >> > (nil)))<br /> >> >gweigs.f:144: confused by earlier errors, bailing out<br /> >> >make[5]: *** [/home/gempak/GEMPAK/lib/linux/grphgd.a(gweigs.o)] Error 1<br > /> >> >make[5]: Leaving directory `/home/gempak/GEMPAK/gempak/source/programs/gd/g > rph<br /> >> > gd'<br /> >> ><br /> >> <br /> >> **************************************************************************** > <<br /> >> Unidata User Support UCAR Unidata Program > <<br /> >> (303)497-8643 P.O. Box 3000 > <<br /> >> address@hidden Boulder, CO 80307 > <<br /> >> ---------------------------------------------------------------------------- > <<br /> >> Unidata WWW Service http://my.unidata.ucar.edu/content/support > <<br /> >> **************************************************************************** > <<br /> >