[Bug middle-end/37448] [4.3 Regression] gcc 4.3.1 cannot compile big function

[hubicka at gcc dot gnu dot org]

------- Comment #21 from hubicka at gcc dot gnu dot org  2008-09-26 22:15 -------
I've just tested Kenny's df scan ref union patch.  We still have over 2.2GB
footprint that is not that much different from unpatched compiler.
allocpool statistics now give some more useful data:
df_scan_ref base pool          868 -1871497632  553770408          0
copies                         434     214464     134512          0
df_scan_ref artificial pool    868   77289168   11231352          0
df_scan_ref regular pool       868  737848848   72493272          0

So largest one is df_scan_ref base, not artifical defs as was suspected.  About
0.5GB comming to that storage.

Other large pool (now at -O2) is:
allocno live ranges      434  167266736  167062320          0

And bitmaps:
df-problems.c:311 (df_rd_alloc)           110920 1071636320 1071611720
1071611720          0
df-problems.c:312 (df_rd_alloc)           110920  543723720  534523840 
534523840          0
df-problems.c:539 (df_rd_transfer_functio  52865  541674400  537236800 
537236760     101861
tree-ssa-pre.c:584 (bitmap_set_new)       547137 1368222720 1366138880
1366138880     784222
tree-ssa-pre.c:585 (bitmap_set_new)       547137  951660760  949608960 
949608960    1467881

and GGC garbage:
emit-rtl.c:3343 (make_insn_raw)                    33796400: 4.4%          0:
0.0%         88: 0.0%    3072408: 3.1%     384051
tree-inline.c:3600 (copy_tree_r)                   36731592: 4.8%          0:
0.0%          0: 0.0%    2702392: 2.7%     478660
lists.c:143 (alloc_EXPR_LIST)                      37188960: 4.8%          0:
0.0%          0: 0.0%    7437792: 7.5%     929724
tree-ssanames.c:141 (make_ssa_name_fn)             52532760: 6.8%          0:
0.0%       3480: 0.0%    3502416: 3.5%     437802


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37448