Fixes are available
April 2012 PTF for XL Fortran for AIX, V13.1
July 2012 PTF for XL Fortran for AIX, V13.1
October 2012 PTF for XL Fortran for AIX, V13.1
February 2013 PTF for XL Fortran for AIX, V13.1
May 2013 PTF for XL Fortran for AIX, V13.1
XL Fortran for AIX Fix Pack 15 (August 2013 PTF) for 13.1
XL Fortran for AIX Fix Pack 16 (November 2013 PTF) for 13.1
XL Fortran for AIX Fix Pack 17 (February 2014 PTF) for 13.1
XL Fortran for AIX Fix Pack 21 (April 2015 PTF) for 13.1
XL Fortran for AIX Fix Pack 22 (September 2015 PTF) for 13.1
XL Fortran for AIX Fix Pack 23 (March 2016 PTF) for 13.1
XL Fortran for AIX Fix Pack 24 (September 2016 PTF) for 13.1
APAR status
Closed as program error.
Error description
IPA64 abends when compiling the following test case at -O3. ===== Compile Command: xlf90_r -q64 -qfixed=72 -O3 -qmaxmem=-1 -qdbg -qfullpath -qwarn64 -qstrict -qfloat=rrm:norsqrt:nofold -qreport -qflttrap=ov:zero:inv:imp:enable -qsigtrap -qsuppress=1501-245 -qalias=intptr -qreport=smplist -qextname -qintsize=8 -qrealsize=8 -qarch=pwr7 -qtune=pwr7 -qsimd=auto -qnocheck -qspillsize=32648 -qnoescape -qxflag=ngenstub -bmaxstack:0xF0000000 -qnoundef -qsmp=omp -qnosave -qtbtable=full -qxref -qattr -bnoquiet -bdatapsize:64K -bstackpsize:64K -btextpsize:64K -qsource -c mieint.F ===== Test case: $ cat mieint.F Subroutine MieInt(Dx, SCm, Inp, Dqv, Dqxt, Dqsc, Dbsc, Dg, Xs1, 1 Xs2, DPh, Error) Integer * 2 Imaxx Parameter (Imaxx = 12000) Real * 4 RIMax ! largest real part of refractive inde Parameter (RIMax = 2.5) Real * 4 IRIMax ! largest imaginary part of refractive Parameter (IRIMax = -2) Integer * 2 Itermax Parameter (Itermax = 12000 * 2.5) ! must be large enough to cope with the ! largest possible nmx = x * abs(scm) + ! or nmx = Dx + 4.05*Dx**(1./3.) + 2.0 Integer * 2 Imaxnp Parameter (Imaxnp = 10000) ! Change this as required ! INPUT Real * 8 Dx Complex * 16 SCm Integer * 4 Inp Real * 8 Dqv(Inp) ! OUTPUT Complex * 16 Xs1(InP) Complex * 16 Xs2(InP) Real * 8 Dqxt Real * 8 Dqsc Real * 8 Dg Real * 8 Dbsc Real * 8 DPh(InP) Integer * 4 Error ! LOCAL Integer * 2 I Integer * 2 NStop Integer * 2 NmX Integer * 4 N ! N*N > 32767 ie N > 181 Integer * 4 Inp2 Real * 8 Chi,Chi0,Chi1 Real * 8 APsi,APsi0,APsi1 Real * 8 Pi0(Imaxnp) Real * 8 Pi1(Imaxnp) Real * 8 Taun(Imaxnp) Real * 8 Psi,Psi0,Psi1 Complex * 8 Ir Complex * 16 Cm Complex * 16 A,ANM1,APB Complex * 16 B,BNM1,AMB Complex * 16 D(Itermax) Complex * 16 Sp(Imaxnp) Complex * 16 Sm(Imaxnp) Complex * 16 Xi,Xi0,Xi1 Complex * 16 Y ! ACCELERATOR VARIABLES Integer * 2 Tnp1 Integer * 2 Tnm1 Real * 8 Dn Real * 8 Rnx Real * 8 S(Imaxnp) Real * 8 T(Imaxnp) Real * 8 Turbo Real * 8 A2 Complex * 16 A1 If ((Dx.Gt.Imaxx) .Or. (InP.Gt.ImaxNP)) Then Error = 1 Return EndIf Cm = SCm Ir = 1 / Cm Y = Dx * Cm If (Dx.Lt.0.02) Then NStop = 2 Else If (Dx.Le.8.0) Then NStop = Dx + 4.00*Dx**(1./3.) + 2.0 Else If (Dx.Lt. 4200.0) Then NStop = Dx + 4.05*Dx**(1./3.) + 2.0 Else NStop = Dx + 4.00*Dx**(1./3.) + 2.0 End If End If End If NmX = Max(Real(NStop),Real(Abs(Y))) + 15. If (Nmx .gt. Itermax) then Error = 1 Return End If Inp2 = Inp+1 D(NmX) = Dcmplx(0,0) Do N = Nmx-1,1,-1 A1 = (N+1) / Y D(N) = A1 - 1/(A1+D(N+1)) End Do Do I =1,Inp2 Sm(I) = Dcmplx(0,0) Sp(I) = Dcmplx(0,0) Pi0(I) = 0 Pi1(I) = 1 End Do Psi0 = Cos(Dx) Psi1 = Sin(Dx) Chi0 =-Sin(Dx) Chi1 = Cos(Dx) APsi0 = Psi0 APsi1 = Psi1 Xi0 = Dcmplx(APsi0,Chi0) Xi1 = Dcmplx(APsi1,Chi1) Dg = 0 Dqsc = 0 Dqxt = 0 Tnp1 = 1 Do N = 1,Nstop DN = N Tnp1 = Tnp1 + 2 Tnm1 = Tnp1 - 2 A2 = Tnp1 / (DN*(DN+1D0)) Turbo = (DN+1D0) / DN Rnx = DN/Dx Psi = Dble(Tnm1)*Psi1/Dx - Psi0 APsi = Psi Chi = Tnm1*Chi1/Dx - Chi0 Xi = Dcmplx(APsi,Chi) A = ((D(N)*Ir+Rnx)*APsi-APsi1) / ((D(N)*Ir+Rnx)* Xi- Xi1) B = ((D(N)*Cm+Rnx)*APsi-APsi1) / ((D(N)*Cm+Rnx)* Xi- Xi1) Dqxt = Tnp1 * Dble(A + B) + Dqxt Dqsc = Tnp1 * (A*Conjg(A) + B*Conjg(B)) + Dqsc If (N.Gt.1) then Dg = Dg + (dN*dN - 1) * Dble(ANM1*Conjg(A) + BNM1 * 1 Conjg(B)) / dN + TNM1 * Dble(ANM1*Conjg(BNM1)) 2 / (dN*dN - dN) End If Anm1 = A Bnm1 = B APB = A2 * (A + B) AMB = A2 * (A - B) Do I = 1,Inp2 If (I.GT.Inp) Then S(I) = -Pi1(I) Else S(I) = Dqv(I) * Pi1(I) End If T(I) = S(I) - Pi0(I) Taun(I) = N*T(I) - Pi0(I) Sp(I) = APB * (Pi1(I) + Taun(I)) + Sp(I) Sm(I) = AMB * (Pi1(I) - Taun(I)) + Sm(I) Pi0(I) = Pi1(I) Pi1(I) = S(I) + T(I)*Turbo End Do Psi0 = Psi1 Psi1 = Psi Apsi1 = Psi1 Chi0 = Chi1 Chi1 = Chi Xi1 = Dcmplx(APsi1,Chi1) End Do If (Dg .GT.0) Dg = 2 * Dg / Dqsc Dqsc = 2 * Dqsc / Dx**2 Dqxt = 2 * Dqxt / Dx**2 Do I = 1,Inp Xs1(I) = (Sp(I)+Sm(I)) / 2 Xs2(I) = (Sp(I)-Sm(I)) / 2 Dph(I) = 2 * Dble(Xs1(I)*Conjg(Xs1(I)) + Xs2(I)*Conjg(Xs2(I))) 1 / (Dx**2 * Dqsc) End Do Dbsc = 4 * Abs(( (Sp(Inp2)+Sm(Inp2))/2 )**2) / Dx**2 Error = 0 Return End subroutine MieInt $ ===== ACTUAL OUTPUT: $xlf90_r -q64 -qfixed=72 -O3 -qmaxmem=-1 -qdbg -qfullpath -qwarn64 -qstrict -qfloat=rrm:norsqrt:nofold -qreport -qflttrap=ov:zero:inv:imp:enable -qsigtrap -qsuppress=1501-245 -qalias=intptr -qreport=smplist -qextname -qintsize=8 -qrealsize=8 -qarch=pwr7 -qtune=pwr7 -qsimd=auto -qnocheck -qspillsize=32648 -qnoescape -qxflag=ngenstub -bmaxstack:0xF0000000 -qnoundef -qsmp=omp -qnosave -qtbtable=full -qxref -qattr -bnoquiet -bdatapsize:64K -bstackpsize:64K -btextpsize:64K -qsource -c mieint.F ** mieint === End of Compilation 1 === The assert subroutine failed: lit_len <= 8 || (lit_len == 16 && dt == DT_COMPLEX), file literal.cpp, line 2738 Calling signal handler... 1586-494 (U) INTERNAL COMPILER ERROR: Signal 6. /gsa/tlbgsa/projects/x/xlcmpbld/run/xlf/131/aix/solution/111020/ usr/bin/.orig/xlf90_r: 1501-230 (S) Internal compiler error; please contact your Service Representative. For more information visit: http://www.ibm.com/support/docview.wss?uid=swg21110810 1501-511 Compilation failed for file mieint.F. $ ===== EXPECTED OUTPUT: Should compile clean
Local fix
Use -O2 instead of -O3.
Problem summary
USERS AFFECTED: XL Fortran users using O3 and up may potentially be affected by this issue. PROBLEM DESCRIPTION: A specific internal representation of literals was not designed to be supported in certain scenarios.
Problem conclusion
The specific internal representation of literals is now supported via this fix.
Temporary fix
Comments
APAR Information
APAR number
IV13177
Reported component name
XL FORTRAN FOR
Reported component ID
5724X1500
Reported release
D10
Status
CLOSED PER
PE
NoPE
HIPER
NoHIPER
Special Attention
NoSpecatt
Submitted date
2012-01-13
Closed date
2012-03-28
Last modified date
2012-03-28
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Fix information
Fixed component name
XL FORTRAN FOR
Fixed component ID
5724X1500
Applicable component levels
RD10 PSY U850834
UP C
[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSGH4D","label":"XL Fortran for AIX"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"13.1","Edition":"","Line of Business":{"code":"LOB57","label":"Power"}}]
Document Information
Modified date:
28 March 2012