Menu ▾ ▴

#552 Linux PowerPC and ibook G4 7447A

both_stable_and_developer
closed-fixed
R. Clint Whaley
Problems during config (126)
5
2008-07-23
2008-06-27
Bin Zhang
No

Hi,

atlas-3.8.2 does not detect my ibook G4 (linux powerpc).

-------------------
Assembly configured as GAS_PPC (4)
atlconf_misc.o: In function `CmndResults':
/home/zhang/src/ATLAS/my_build_dir/..//CONFIG/src/atlconf_misc.c:306: warning: the use of `tmpnam' is dangerous, better use `mkstemp'

Bad VECFLAG value=0, ierr=0, ln2='VECFLAG=0
'

Vector ISA Extension configured as (0,0)
atlconf_misc.o: In function `CmndResults':
/home/zhang/src/ATLAS/my_build_dir/..//CONFIG/src/atlconf_misc.c:306: warning: the use of `tmpnam' is dangerous, better use `mkstemp'
atlconf_misc_BE.o: In function `CmndResults':
/home/zhang/src/ATLAS/my_build_dir/..//CONFIG/src/atlconf_misc.c:306: warning: the use of `tmpnam' is dangerous, better use `mkstemp'

Bad MACHTYPE value=0, ierr=0, ln2='MACHTYPE=0
'

Architecture configured as UNKNOWN (0)

Bad CPU MHZ value=0, ierr=0, ln2='CPU MHZ=0
'

Clock rate configured as 0Mhz
-------------------------------------

----------------
$ cat /proc/cpuinfo
processor : 0
cpu : 7447A, altivec supported
clock : 1199.999000MHz
revision : 0.1 (pvr 8003 0101)
bogomips : 73.47
timebase : 18432000
platform : PowerMac
machine : PowerBook6,5
motherboard : PowerBook6,5 MacRISC3 Power Macintosh
detected as : 287 (iBook G4)
pmac flags : 0000001b
L2 cache : 512K unified
pmac-generation : NewWorld
---------------------------------

If I add
else if (strstr(res, "7447")) mach = PPCG4;
after
else if (strstr(res, "7410")) mach = PPCG4;
in CONFIG/src/backend/archinfo_linux.c, configure can detect Architecture PPCG4:

-----------------
OS configured as Linux (1)

Assembly configured as GAS_PPC (4)

Bad VECFLAG value=0, ierr=0, ln2='VECFLAG=0
'

Vector ISA Extension configured as (0,0)
atlconf_misc_BE.o: In function `CmndResults':
/home/zhang/src/ATLAS/my_build_dir/..//CONFIG/src/atlconf_misc.c:306: warning: the use of `tmpnam' is dangerous, better use `mkstemp'

Architecture configured as PPCG4 (4)

Bad CPU MHZ value=0, ierr=0, ln2='CPU MHZ=0
'

Clock rate configured as 0Mhz
-------------------------------------

But I still have two errors "Bad ....", are they important?

And how do I know if altivec is used?

Thanks for your help,
Bin

Discussion

  • R. Clint Whaley

    R. Clint Whaley - 2008-07-18

    Logged In: YES
    user_id=182470
    Originator: NO

    Bin,

    Unfortunately the "ISA extension configured as (0,0)" means that altivec was not detected!

    It's hard to say what is going wrong from the above output. Post the complete output of your configure (you can use script, man script if you don't know), but pass the "-v 2" flag to configure (telling it to be verbose).

    Thanks,
    Clint

     

  • R. Clint Whaley

    R. Clint Whaley - 2008-07-18
    • labels: --> Problems during config
    • milestone: --> both_stable_and_developer
    • assigned_to: nobody --> rwhaley
     

  • Bin Zhang

    Bin Zhang - 2008-07-19

    Logged In: YES
    user_id=2129713
    Originator: YES

    Hi Clint,

    Attached configure.log of the complete output of my "configure -v 2" with atlas 3.9.0.

    Thanks,
    Bin
    File Added: configure.log

     

  • Bin Zhang

    Bin Zhang - 2008-07-19

    output "configure -v 2"

     
    configure.log

  • R. Clint Whaley

    R. Clint Whaley - 2008-07-21

    Logged In: YES
    user_id=182470
    Originator: NO

    Bin,

    This install failed because CPU throttling was turned on (maybe you had your laptop running on the battery?):

    CPU Throttling apparently enabled!^M
    It appears you have cpu throttling enabled, which makes timings^M
    unreliable and an ATLAS install nonsensical. Aborting.^M
    See ATLAS/INSTALL.txt for further information^M
    xconfig exited with 1^M

    You can find more info on CPU throttling in ATLAS/INSTALL.txt and :
    http://math-atlas.sourceforge.net/atlas_install/atlas_install.html#SECTION00032000000000000000

    Also, since you are using linux, you need to issue:
    http://math-atlas.sourceforge.net/errata.html#G4gcc

    So, apply your above fix (so configure recognizes it's a G4), and scope the above, and then post the full configure script command again.

    Cheers,
    Clint

     

  • Bin Zhang

    Bin Zhang - 2008-07-22
     
    configure.ppcG4_7447A.log

  • Bin Zhang

    Bin Zhang - 2008-07-22

    Logged In: YES
    user_id=2129713
    Originator: YES

    Thanks. ATLAS 3.9.0 on my slow ibook G4 1.2Ghz (linux):

    - Add G4 7447:
    --- ATLAS/CONFIG/src/backend/archinfo_linux.c.orig 2008-07-21 21:26:04.000000000 +0200
    +++ ATLAS/CONFIG/src/backend/archinfo_linux.c 2008-07-21 21:36:07.000000000 +0200
    @@ -72,6 +72,7 @@
    if (strstr(res, "G4")) mach = PPCG4;
    else if (strstr(res, "7400")) mach = PPCG4;
    else if (strstr(res, "7410")) mach = PPCG4;
    + else if (strstr(res, "7447")) mach = PPCG4;
    else if (strstr(res, "7455")) mach = PPCG4;
    else if (strstr(res, "PPC970FX")) mach = PPCG5;
    else if (strstr(res, "POWER5")) mach = IbmPwr5;

    - script -c "../configure --cflags='-mregnames' -v 2 -D c -DATL_AVgcc" configure.ppcG4_7447A.log
    (Bad CPU MHZ value=0, ierr=0, ln2='CPU MHZ=0, see attached output configure.ppcG4_7447A.log)

    - "time make" gives me
    real 53m56.651s
    user 47m10.284s
    sys 4m8.267s

    - "make check"
    DONE BUILDING TESTERS, RUNNING:
    SCOPING FOR FAILURES IN BIN TESTS:
    fgrep -e fault -e FAULT -e error -e ERROR -e fail -e FAIL \
    bin/sanity.out
    8 cases: 8 passed, 0 skipped, 0 failed
    4 cases: 4 passed, 0 skipped, 0 failed
    8 cases: 8 passed, 0 skipped, 0 failed
    4 cases: 4 passed, 0 skipped, 0 failed
    8 cases: 8 passed, 0 skipped, 0 failed
    4 cases: 4 passed, 0 skipped, 0 failed
    8 cases: 8 passed, 0 skipped, 0 failed
    4 cases: 4 passed, 0 skipped, 0 failed
    DONE
    SCOPING FOR FAILURES IN CBLAS TESTS:
    fgrep -e fault -e FAULT -e error -e ERROR -e fail -e FAIL \
    interfaces/blas/C/testing/sanity.out | \
    fgrep -v PASSED
    make[1]: [sanity_test] Error 1 (ignored)
    DONE
    SCOPING FOR FAILURES IN F77BLAS TESTS:
    fgrep -e fault -e FAULT -e error -e ERROR -e fail -e FAIL \
    interfaces/blas/F77/testing/sanity.out | \
    fgrep -v PASSED
    make[1]: [sanity_test] Error 1 (ignored)
    DONE
    make[1]: Leaving directory `/home/zhang/src/ATLAS/my_build_dir'

    File Added: configure.ppcG4_7447A.log

     

  • R. Clint Whaley

    R. Clint Whaley - 2008-07-23

    Logged In: YES
    user_id=182470
    Originator: NO

    I think you wanted "make time", not "time make" :)

    Looks like your install is working now, so I'm closing this support request. Reopen if you have further questions.

    If you install again, get 3.9.1 since it fixes several small errors in 3.9.0.

    Cheers,
    Clint

     

  • R. Clint Whaley

    R. Clint Whaley - 2008-07-23
    • status: open --> closed-fixed
     

  • Bin Zhang

    Bin Zhang - 2008-07-25

    Logged In: YES
    user_id=2129713
    Originator: YES

    No. I wanted to know the time of compilation. So I did a "time make"
    On my slow ibook G4 1.2Ghz, the compilation time is approximately one hour, with or without AltiVec detected. Is this normal?

     

  • Bin Zhang

    Bin Zhang - 2008-07-25

    Logged In: YES
    user_id=2129713
    Originator: YES

    No. I wanted to know the time of compilation. So I did a "time make"
    On my slow ibook G4 1.2Ghz, the compilation time is approximately one hour, with or without AltiVec detected. Is this normal?

     

  • Bin Zhang

    Bin Zhang - 2008-08-08

    Logged In: YES
    user_id=2129713
    Originator: YES

    All my apologies. I forgot to mention that the patch "+ else if (strstr (res," 7447 ")) = mach PPCG4" is found in sagemath. It is written by Michael Abshoff <michael.abshoff@googlemail.com> to whom I present my apologies.

     

Log in to post a comment.