Author: Dann Corbit
Date: 23:46:27 12/04/03
Go up one level in this thread
On December 05, 2003 at 00:22:11, Russell Reagan wrote:
>I found an old piece of code I had written to test the speed of a few
>bitscanning routines (ex. FirstOne() in Crafty). I was curious how different
>they would run on a 64-bit machine, so I sent them to Dr. Hyatt and he was kind
>enough to run them for me on the Opteron. Thanks Bob ;-)
>
>The source code of my program is below. The output of the program is in the
>form: <name of routine> <checksum> <seconds>. Lower times are better.
>
>Here are the results on my 32-bit 2GHz Athlon:
>
> magic bitscan 2980130816 11.987
> magic bitscan 2980130816 12.207
> magic bitscan 2980130816 12.338
> magic bitscan 2980130816 12.448
> ------------------------------
> table 16 2980130816 7.250
> table 16 2980130816 7.251
> table 16 2980130816 7.140
> table 16 2980130816 7.040
> ------------------------------
> gerd 2980130816 9.243
> gerd 2980130816 9.264
> gerd 2980130816 9.253
> gerd 2980130816 9.273
> ------------------------------
> eugene 2980130816 7.201
> eugene 2980130816 7.170
> eugene 2980130816 7.190
> eugene 2980130816 7.191
> ------------------------------
> eugene2 2980130816 7.230
> eugene2 2980130816 7.261
> eugene2 2980130816 7.250
> eugene2 2980130816 6.780
>
>Here are the results from the 1.8ghz Opteron:
>
> magic bitscan 2980130816 4.040
> magic bitscan 2980130816 4.350
> magic bitscan 2980130816 4.350
> magic bitscan 2980130816 4.350
> ------------------------------
> table 16 2980130816 4.600
> table 16 2980130816 4.820
> table 16 2980130816 4.240
> table 16 2980130816 4.340
> ------------------------------
> gerd 2980130816 5.790
> gerd 2980130816 5.780
> gerd 2980130816 5.440
> gerd 2980130816 5.440
> ------------------------------
> eugene 2980130816 3.990
> eugene 2980130816 3.960
> eugene 2980130816 3.770
> eugene 2980130816 3.470
> ------------------------------
> eugene2 2980130816 3.690
> eugene2 2980130816 3.690
> eugene2 2980130816 3.680
> eugene2 2980130816 3.690
>
>Here is the output of Bob's FirstOne() written in assembly, run on the Opteron:
>
> ------------------------------
> hyatt 2980130816 5.100
> hyatt 2980130816 5.090
> hyatt 2980130816 5.100
> hyatt 2980130816 5.100
Output using MS VC++ .Net 2003 compiler, on 950 MHz Athlon:
magic bitscan 2980130816 18.616
magic bitscan 2980130816 18.407
magic bitscan 2980130816 18.437
magic bitscan 2980130816 18.457
------------------------------
table 16 1490065408 9.633
table 16 1490065408 9.604
table 16 1490065408 9.664
table 16 1490065408 9.624
------------------------------
gerd 2980130816 9.164
gerd 2980130816 9.183
gerd 2980130816 9.323
gerd 2980130816 9.404
------------------------------
eugene 2980130816 11.336
eugene 2980130816 11.267
eugene 2980130816 11.226
eugene 2980130816 11.266
------------------------------
eugene2 2980130816 10.165
eugene2 2980130816 10.115
eugene2 2980130816 10.154
eugene2 2980130816 10.135
3.0 GHz Pentium 4:
magic bitscan 2980130816 10.421
magic bitscan 2980130816 10.422
magic bitscan 2980130816 10.453
magic bitscan 2980130816 10.421
------------------------------
table 16 1490065408 2.422
table 16 1490065408 2.422
table 16 1490065408 2.422
table 16 1490065408 2.421
------------------------------
gerd 2980130816 2.453
gerd 2980130816 2.438
gerd 2980130816 2.437
gerd 2980130816 2.422
------------------------------
eugene 2980130816 3.250
eugene 2980130816 3.250
eugene 2980130816 3.250
eugene 2980130816 3.250
------------------------------
eugene2 2980130816 2.062
eugene2 2980130816 2.078
eugene2 2980130816 2.078
eugene2 2980130816 2.079
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.