Re:CCL:Re:CCL:Summary: AMD-K6/linux for G98

 > > For pinene using DFT/B3LYP with large basis sets, for
 > >example, the machine i bought for $11,000 was faster than a Cray
 T3E-600
 > >with 4 processors and only 15% slower than a Cray T3E-1200 with 4
 > >processors, both running parallel Gaussian 94. (The G94 calculations
 were
 > >done by Sosa and Frisch, so were probably pretty well optimized for the
 > >machine, see last July 15 number of J. Comput. Chem. Sosa et al. Vol
 19,
 > >p. 1053). I haven't priced those machines but they must be >$100k.
 If you
 > >want to benchmark against one of your own machines, Pulay would be
 happy
 > >to do some calculations for you.
 > The comparison here is *completely* invalid.  If you want to compare
 > the speed of machines, you must be running the same program.  In this
 > case, a highly-optimized DFT implementation (the one from Pulay) is being
 > compared to an implementation which can charitably be called "not so
 > great", this does not say a single thing about the relative speeds of
 the
 > computers running those programs.
 >
 > Having Pulay do a benchmark for you would be altering two variables at the
 > same time:
 >  (a) the hardware
 >  (b) the software
 > it is impossible to decouple the effects of the two changes.
 > One more time, for good measure:
 > If you want reliable benchmarks of *hardware* performance, you must
 > compare the same programs on the different hardware.
 >
 Moreover, just comparing the same program is not enough. The program
 must be compiled with the same compiler. Even providing compilers
 from one vendor were used, the compilers must produce assembler code
 with the same efficiency. Next, exactly the same operation systems
 should be used on the different platforms, isn't it?
    etc., etc., ...
 Ideal hardware benchmarking on a *real* ab-inito task seems to be a
 very difficult problem.
 I agree with Greg's opinion that 1) hardware benchmarking and 2) software
 benchmarking  require different strategies in order to get correct results.
 However, I would not call the PII/Cray comparison completely invalid.
 This may be the third type of benchmarking: the best performance/price
 ratio. Another words, how to achieve the the best results within a given
 budget. For an average computational chemist this type of benchmarking
 is really important and the original post does give us useful information.
 Finely, I would suggest to support all kinds of benchmarks. Of course a
 care should be taken comparing the results from the different benchmark
 types.
 best regards,
 Valentin.
 ====================================================================
                                              ,         ,      ,   ,
 Valentin P. Ananikov                         |\\\\ ////|     /////|
 NMR Group                                    | \\\|/// |    ///// |
 ND Zelinsky Institute of Organic Chemistry   |  |~~~|  |   |~~~|  |
 Leninsky Prospect 47                         |  |===|  |   |===|  |
 Moscow  117913                               |  |   |  |   |   |  |
 Russia                                       |  | A |  |   | Z |  |
                                               \ |   | /    |   | /
 e-mail: val { *at * } cacr.ioc.ac.ru                     \|===|/     |===|/
 http://nmr.ioc.ac.ru/Staff/AnanikovVP/          '---'
 '---'
   Fax +7 (095)1355328   Phone +7 (095)9383536
 ====================================================================