I still believe that P4 is better for scientific applications, what SPECfp tries to analyze. Nevertheless, something must be considered.
Both, P4 and Athlon, use profiling to improve branch prediction. Clearly, P4 is more sensitive to this technology because the large pipeline.
In real world, nobody does profiling (sad but true). Without profiling the differences are closer.
Currently PostgreSQL is better than MySQL, but this is changing. See the great MySQL improvements from
innodb.
Once this code is more stable it would give lots of image problems to postgres. Additionally, MySQL has a very clean code.
I still believe that P4 is better for scientific applications, what SPECfp tries to analyze. Nevertheless, something must be considered.
Both, P4 and Athlon, use profiling to improve branch prediction. Clearly, P4 is more sensitive to this technology because the large pipeline.
In real world, nobody does profiling (sad but true). Without profiling the differences are closer.
innodb.
Once this code is more stable it would give lots of image problems to postgres. Additionally, MySQL has a very clean code.
Why not to create an extension to the GPL/BSD/LPGL... licenses so that they are free for everybody but for Microsoft?