Yet Another Benchmark Interface ! :)

[oex]

AMD Sempron(tm) Processor 3100+
Results 8 pass average
timing utoa    615 ms
timing utoa_ex 189 ms
timing utoa2   338 ms
timing utoa3   587 ms
timing utoa_exLingo   176 ms

[hutch--]

 :U

Very good results on this Core2 Quad. It has upped the time of Paul's algo by about 20%.

Code Select Expand


Intel(R) Core(TM)2 Quad CPU Q9650 @ 3.00GHz
Results 8 pass average
timing utoa    285 ms
timing utoa_ex 78 ms
timing utoa2   156 ms
timing utoa3   234 ms
timing utoa_exLingo   62 ms


[hutch--]

Lingo,

Sad to say your modification of Paul Dixon's algo is broken. Attached is the test piece.

[dedndave]

for 900000000-999999999, it returns the string with a leading 0 (0900000000-0999999999)

[lingo]

Corrected.  Thank you for testing.. :wink
The technical error was:
Wrong: jc  Lo0
Corrected: jbe Lo0

[dedndave]

w/lingo's fixed proc
prescott w/htt

Code Select Expand

Intel(R) Pentium(R) 4 CPU 3.00GHz
Results 8 pass average
timing utoa            802 ms
timing utoa_ex         302 ms
timing utoa2           615 ms
timing utoa3           734 ms
timing utoa_exLingo    247 ms

[brethren]

Code Select Expand

AMD Turion(tm) 64 X2 Mobile Technology TL-52
Results 8 pass average
timing utoa    664 ms
timing utoa_ex 208 ms
timing utoa2   367 ms
timing utoa3   626 ms
timing utoa_exLingo   193 ms

[hutch--]

Compliments, this version works fine.

Decyphered and fit to be read by human beings. Will put it into the benchmark a bit later.

[lingo]

Thanks Hutch but you used my old variant in lingo_fixed.zip,
so pls reload my  bmtemplate1.zip. file  :wink

[hutch--]

All I did was tweak the jump you posted.

[clive]

Code Select Expand

Intel(R) Pentium(R) Dual CPU T2330 @ 1.60GHz
Results 8 pass average
timing utoa    563 ms
timing utoa_ex 173 ms
timing utoa2   329 ms
timing utoa3   466 ms
timing utoa_exLingo   163 ms

Code Select Expand

AMD Athlon(tm) II X2 215 Processor
Results 8 pass average
timing utoa    419 ms
timing utoa_ex 153 ms
timing utoa2   224 ms
timing utoa3   374 ms
timing utoa_exLingo   144 ms

Code Select Expand

Intel(R) Atom(TM) CPU N450 @ 1.66GHz
Results 8 pass average
timing utoa    1343 ms
timing utoa_ex 536 ms
timing utoa2   824 ms
timing utoa3   1339 ms
timing utoa_exLingo   526 ms

Code Select Expand

Intel(R) Celeron(R) CPU 900 @ 2.20GHz
Results 8 pass average
timing utoa    429 ms
timing utoa_ex 126 ms
timing utoa2   247 ms
timing utoa3   364 ms
timing utoa_exLingo   111 ms

[hutch--]

Here is your later version.

Code Select Expand


Intel(R) Core(TM)2 Quad CPU Q9650 @ 3.00GHz
Results 8 pass average
timing utoa    281 ms
timing utoa_ex 78 ms
timing utoa2   156 ms
timing utoa3   234 ms
timing utoa_exLingo   62 ms


Lingo,

One question, it would be a much more useful algo if it balanced the CALL / RET pairing. I don't like to fiddle with complex algos, do you think it could be done as no big deal ?

[lingo]

"...do you think it could be done as no big deal ?"

You can change the end

Code Select Expand


from:
        pop   esi
jmp   edx
to:
        pop   esi
push  edx
ret
but you will lose a tick... :wink
You can try to preserve ecx too (like jj) ... :lol

[hutch--]

No change in timing.

Code Select Expand


Intel(R) Core(TM)2 Quad CPU Q9650 @ 3.00GHz
Results 8 pass average
timing utoa    283 ms
timing utoa_ex 78 ms
timing utoa2   156 ms
timing utoa3   234 ms
timing utoa_exLingo   62 ms


[FORTRANS]

Hi,

   This is the bmlog.txt from hutch's code in Reply #6.

Code Select Expand





Results 8 pass average
timing utoa       1634 ms
timing utoa_ex    646 ms
timing udw2str    2443 ms
timing vcrt ustr$ 4435 ms


   The code in Reply #13 does not run under Windows 2000
on this machine.

Regards,

Steve N.