[erlang-questions] native

[Tony Rogvall]

Hi list.

I started playing with a simple implementation of MD5 (basically straight from the wiki page) 
When the implementation produce the correct output, I did some timing for fun and also
tried to native compile it. 
To my astonishment I discovered that the code was slower when native compiled.
I am using R15B01 on mac.  Can this be true? And in this case why? 

Note that I do not care to optimize the implementation it self! I am more interested why
the native compiler produce slower code on this example.

Thanks

Tony


%%%%%%%
-module(md5).

-compile(export_all).

k(I) when I >= 0, I < 64 -> %% may construct table!
    trunc(abs(math:sin(I+1)) * (1 bsl 32)).

r(I) when I >= 0, I < 64 ->
    element(I+1,
	    {7, 12, 17, 22, 7, 12, 17, 22, 7, 12, 17, 22, 7, 12, 17, 22,
	     5,  9, 14, 20, 5,  9, 14, 20, 5,  9, 14, 20, 5,  9, 14, 20,
	     4, 11, 16, 23, 4, 11, 16, 23, 4, 11, 16, 23, 4, 11, 16, 23,
	     6, 10, 15, 21, 6, 10, 15, 21, 6, 10, 15, 21, 6, 10, 15, 21}).

hex(Bin) ->
    [ element(H+1,{$0,$1,$2,$3,$4,$5,$6,$7,$8,$9,$a,$b,$c,$d,$e,$f}) ||
	<<H:4>> <= Bin ].

md5_test(Input) ->
    {hex(md5(Input)), hex(erlang:md5(Input))}.

md5(List) when is_list(List) ->
    md5(list_to_binary(List));
md5(Bits) when is_bitstring(Bits) ->
    Length = bit_size(Bits),
    Pad    = (512 - (Length+1+64) rem 512) rem 512,
    Bits1  = <<Bits/bitstring,1:1,0:Pad,Length:64/little>>,
    md5(Bits1,16#67452301,16#efcdab89,16#98badcfe,16#10325476).

md5(<<>>, H0,H1,H2,H3) ->
    <<H0:32/little,H1:32/little,H2:32/little,H3:32/little>>;
md5(Bits,H0,H1,H2,H3) ->
    <<Chunk:64/binary, Rest/binary>> = Bits,
    W = list_to_tuple([ X || <<X:32/little>> <= Chunk ]),
    {A,B,C,D} = md5_(0, W, H0, H1, H2, H3),
    md5(Rest, H0+A,H1+B,H2+C,H3+D).

md5_(I, W, A, B, C, D) when I < 16 ->
    F = (B band C) bor ((bnot B) band D),
    G = I,
    R = rotate(A+F+k(I)+element(G+1,W), r(I)),
    md5_(I+1,W,D,B+R,B,C);
md5_(I, W, A, B, C, D) when I < 32 ->
    F = (D band B) bor ((bnot D) band C),
    G = (5*I + 1) rem 16,
    R = rotate(A+F+k(I)+element(G+1,W), r(I)),
    md5_(I+1,W,D,B+R,B,C);
md5_(I, W, A, B, C, D) when I < 48 ->
    F = (B bxor C) bxor D,
    G = (3*I+5) rem 16,
    R = rotate(A+F+k(I)+element(G+1,W), r(I)),
    md5_(I+1,W,D,B+R,B,C);
md5_(I, W, A, B, C, D) when I < 64 ->
    F = C bxor (B bor (bnot D)),
    G = (7*I) rem 16,
    R = rotate(A+F+k(I)+element(G+1,W), r(I)),
    md5_(I+1,W,D,B+R,B,C);
md5_(64, _W, A, B, C, D) ->
    {A,B,C,D}.

rotate(X0, C) ->
    X = X0 band 16#ffffffff,
    ((X bsl C) bor (X bsr (32-C))) band 16#ffffffff.


%%%%%%%
"Installing applications can lead to corruption over time. Applications gradually write over each other's libraries, partial upgrades occur, user and system errors happen, and minute changes may be unnoticeable and difficult to fix"



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20120909/3612a2d7/attachment.htm>