Version française
Home     About     Download     Resources     Contact us    

This site is updated infrequently. For up-to-date information, please visit the new OCaml website at

Browse thread
[Caml-list] problem building ocaml 3.08.1 on MIPS/IRIX
[ Home ] [ Index: by date | by threads ]
[ Search: ]

[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
Date: -- (:)
From: TiM <timreid@g...>
Subject: [Caml-list] problem building ocaml 3.08.1 on MIPS/IRIX

i am attemping to build ocaml 3.08.1 on a MIPS/IRIX system, but i am
having a few troublems.

IRIX v6.5
gcc v2.95.3
cc v7.41

./configure -prefix /home/treid


** Configuration summary **

Directories where Objective Caml will be installed:
        binaries.................. /home/treid/bin
        standard library.......... /home/treid/lib/ocaml
        manual pages.............. /home/treid/man (with extension .1)
Configuration for the bytecode compiler:
        C compiler used........... gcc
        options for compiling..... -fno-defer-pop -Wall
        options for linking.......   -lm  -lcurses -lpthread
        shared libraries are supported
        options for compiling..... -fPIC -fno-defer-pop -Wall
        command for building...... ld -shared -rdata_shared -o
-rpath /a/path objs
Configuration for the native-code compiler:
        hardware architecture..... mips
        OS variant................ irix
        C compiler used........... cc
        options for compiling..... -n32 -D_FILE_OFFSET_BITS=64 -D_REENTRANT
        options for linking....... -n32 -Wl,-woff,84  -lm
        assembler ................ $(AS) -n32 -O2
        preprocessed assembler ... $(AS) -n32 -O2
        profiling with gprof ..... not supported
        ocamlopt -pack ........... not supported (no binutils)
Source-level replay debugger: supported
Additional libraries supported:
        unix str num dynlink bigarray systhreads threads graph dbm
Configuration for the "num" library:
        target architecture ...... mips (asm level 1)
Configuration for the "graph" library:
        options for compiling .... 
        options for linking ...... -lX11
The "labltk" library: not found

so gcc is being used to build the bytecode compiler, and cc is being
used for the native compiler. i believe that gcc is operating in 64b
mode, but the "-n32" option to cc definately means that cc is
operating in 32b mode.

my goal is maximum 64b floating point performance, so i am pretty sure
that the "-n32" option needs to be replaced with the "-64" option, and
i would like to explore "-O3" and various other optimizations,
although i understand that fancy optimizations may break the ocaml

anyhow, without me messing with the Makefile at all, "make world" and
"make bootstrap" complete just fine. however, "make opt" dies with:

make[1]: Entering directory `/u/treid/ocaml/ocaml-3.08.1/stdlib'
../boot/ocamlrun ../ocamlopt -warn-error A -nostdlib `./Compflags
pervasives.cmx` -c
as: Error: /tmp/camlasm86466e.s, line 445: Overflow: 4372995238176751616
      .word 4372995238176751616
as: Error: /tmp/camlasm86466e.s, line 451: Overflow: 4503599627370496
      .word 4503599627370496
as: Error: /tmp/camlasm86466e.s, line 457: Overflow: 9218868437227405311
      .word 9218868437227405311
as: Error: /tmp/camlasm86466e.s, line 463: Overflow: 9218868437227405313
      .word 9218868437227405313
as: Error: /tmp/camlasm86466e.s, line 469: Overflow: 4503599627370496
      .word -4503599627370496
as: Error: /tmp/camlasm86466e.s, line 475: Overflow: 9218868437227405312
      .word 9218868437227405312
as: Error: /tmp/camlasm86466e.s, line 3163: li expression must be
32-bit immediate
      li $5, -9223372036854775807
as: Error: /tmp/camlasm86466e.s, line 3164: li expression must be
32-bit immediate
      li $4, -9223372036854775807
as: Error: /tmp/camlasm86466e.s, line 3166: li expression must be
32-bit immediate
      li $3, 9223372036854775807
Assembler error, input left in file /tmp/camlasm86466e.s
make[1]: *** [pervasives.cmx] Error 2
make[1]: Leaving directory `/u/treid/ocaml/ocaml-3.08.1/stdlib'
make: *** [libraryopt] Error 2

so, nievely, i start over from a freshly unpacked tarball, configure,
and change all occurances of "-n32" in config/Makefile to "-64", but
"make opt" dies in exactly the same way.

any help appreciated ...


To unsubscribe, mail Archives:
Bug reports: FAQ:
Beginner's list: