This is from a BLAS routine. gfortran -O3 -g -march=athlon-xp -mfpmath=sse -mno-80387 -c ddot.f ddot.f: In function 'ddot': ddot.f:46: error: insn does not satisfy its constraints: (insn 365 364 366 12 ddot.f:24 (set (reg:DF 21 xmm0 [104]) (mem:DF (plus:SI (reg/f:SI 6 bp) (const_int -120 [0xffffff88])) [0 S8 A8])) 64 {*movdf_nointeger} (nil) (nil)) ddot.f:46: internal compiler error: in reload_cse_simplify_operands, at postreload.c:391 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions. $ gfortran -v ; gfortran -dumpmachine Using built-in specs. Configured with: ../gcc/configure --prefix=/home/ig25 --enable-languages=c,f95 Thread model: posix gcc version 4.0.0 20050129 (experimental) i686-pc-linux-gnu $ cat BLAS/SRC/ddot.f double precision function ddot(n,dx,incx,dy,incy) c c forms the dot product of two vectors. c uses unrolled loops for increments equal to one. c jack dongarra, linpack, 3/11/78. c modified 12/3/93, array(1) declarations changed to array(*) c double precision dx(*),dy(*),dtemp integer i,incx,incy,ix,iy,m,mp1,n c ddot = 0.0d0 dtemp = 0.0d0 if(n.le.0)return if(incx.eq.1.and.incy.eq.1)go to 20 c c code for unequal increments or equal increments c not equal to 1 c ix = 1 iy = 1 if(incx.lt.0)ix = (-n+1)*incx + 1 if(incy.lt.0)iy = (-n+1)*incy + 1 do 10 i = 1,n dtemp = dtemp + dx(ix)*dy(iy) ix = ix + incx iy = iy + incy 10 continue ddot = dtemp return c c code for both increments equal to 1 c c c clean-up loop c 20 m = mod(n,5) if( m .eq. 0 ) go to 40 do 30 i = 1,m dtemp = dtemp + dx(i)*dy(i) 30 continue if( n .lt. 5 ) go to 60 40 mp1 = m + 1 do 50 i = mp1,n,5 dtemp = dtemp + dx(i)*dy(i) + dx(i + 1)*dy(i + 1) + * dx(i + 2)*dy(i + 2) + dx(i + 3)*dy(i + 3) + dx(i + 4)*dy(i + 4) 50 continue 60 ddot = dtemp return end Line 46 is the '50 continue' line. Thomas
I should note (I think others have to before) that -mno-80387 *******changes********* the ABI.
Created attachment 8103 [details] Failing preprocessed C source code To see wether this problem can also be exposed with the C frontend, I ran the subroutine through f2c and preprocessed it with "gcc -E". Still the same error: $ gcc -O3 -g -march=athlon-xp -mfpmath=sse -mno-80387 ddot.i ddot.c: In function 'ddot_': ddot.c:91: error: insn does not satisfy its constraints: (insn 346 345 347 8 ddot.c:53 (set (reg:DF 21 xmm0 [103]) (mem:DF (plus:SI (reg/f:SI 6 bp) (const_int -128 [0xffffff80])) [0 S8 A8])) 64 {*movdf_nointeger} (nil) (nil)) ddot.c:91: internal compiler error: in reload_cse_simplify_operands, at postreload.c:391 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions. so this is indeed independent of the front end. Thomas
Subject: Bug 19690 CVSROOT: /cvs/gcc Module name: gcc Changes by: rth@gcc.gnu.org 2005-01-30 03:14:52 Modified files: gcc : ChangeLog gcc/config/i386: i386.md Log message: PR target/19690 * config/i386/i386.md (movdf_nointeger, movdf_integer): Fix ordering of # and * in constraints. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.7329&r2=2.7330 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/i386/i386.md.diff?cvsroot=gcc&r1=1.614&r2=1.615
Fixed.