Something is amiss between -fext-core and optimisations.

Validating (--testsuite-only --slow) brought my box to its knees badly, the 
culprit turned out to be T7239 with the optasm, hpc, and optllvm ways:

=====> T7239(hpc) 1434 of 3536 [0, 4, 0]
cd ./ext-core && '/home/dafis/GHC/bghc/bindisttest/install   dir/bin/ghc' -
fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -
rtsopts -fno-ghci-history -c T7239.hs -O -fhpc -fext-core  >T7239.comp.stderr 
2>&1
Compile failed (status 256) errors were:
stack overflow: use +RTS -K<size> to increase it

*** unexpected failure for T7239(hpc)

Running it manually,

$ /home/dafis/GHC/bghc/inplace/bin/ghc-stage2 +RTS -s -M1200M -RTS -fforce-
recomp -rtsopts -c T7239.hs -O -fllvm -fext-core > T7239.comp.stderr 2>&1
dafis@schwartz:~/GHC/bghc/testsuite/tests/ext-core> cat T7239.comp.stderr
ghc-stage2: Heap exhausted;
Current maximum heap size is 1258291200 bytes (1200 MB);
use `+RTS -M<size>' to increase it.
   1,938,087,952 bytes allocated in the heap
   1,835,432,552 bytes copied during GC
   1,241,472,688 bytes maximum residency (19 sample(s))
     174,344,440 bytes maximum slop
            1673 MB total memory in use (8 MB lost due to fragmentation)

                                    Tot time (elapsed)  Avg pause  Max pause
  Gen  0       954 colls,     0 par    1.79s    1.79s     0.0019s    0.0054s
  Gen  1        19 colls,     0 par   10.10s   10.20s     0.5369s    1.7456s

  TASKS: 3 (1 bound, 2 peak workers (2 total), using -N1)

  SPARKS: 0 (0 converted, 0 overflowed, 0 dud, 0 GC'd, 0 fizzled)

  INIT    time    0.00s  (  0.03s elapsed)
  MUT     time    0.74s  (  0.82s elapsed)
  GC      time   11.89s  ( 11.99s elapsed)
  EXIT    time    0.00s  (  0.00s elapsed)
  Total   time   12.63s  ( 12.85s elapsed)

  Alloc rate    2,629,160,866 bytes per MUT second

  Productivity   5.9% of total user, 5.8% of total elapsed

gc_alloc_block_sync: 0
whitehole_spin: 0
gen[0].sync: 0
gen[1].sync: 0

it gobbles up a lot of memory, rendering the box unresponsive for several 
minutes if the heap isn't restricted to be small enough.

Removing either -O or -fext-core makes it well-behaved.

_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to