I wouldn't worry if it produces correct output.

sha1.c is likely to be an implementation of the SHA1 cryptographic
hash algorithm.  Cryptographic algorithms typically give compilers a
lot of room to run wild in optimization.  PIC makes things more
complex because memory-references get more complex.  When compiled -O
-fpic, it likely is causing the `n' in the several O(n**2) (or slower)
optimization algorithms in gcc to get somewhat larger.

To see where the time's going, you can run the `cc1' pass manually,
and it will print out the time spent in each pass:

% gcc -v -E -O md4c.c >md4c.pp 
% /usr/libexec/cc1 md4c.pp -O6 -fpic -o md4c.o
 MD4Init MD4Update MD4Final MD4Transform Encode Decode
time in parse: 0.702689
time in integration: 0.012710
time in jump: 0.143015
time in cse: 0.535831
time in loop: 0.060145
time in cse2: 0.408662
time in flow: 0.166294
time in combine: 1.598138
time in sched: 0.000000
time in local-alloc: 0.676442
time in global-alloc: 1.676413
time in sched2: 0.000000
time in dbranch: 0.000000
time in shorten-branch: 0.000776
time in stack-reg: 0.019758
time in final: 0.355848
time in varconst: 0.013127
time in symout: 0.000000
time in dump: 0.000000


					- Bill