Fairly stated, a mathematical simulation model  has to accept the following parameters: locations (longitude    and latitude of impact); vertical angle of impact (0 to 180 degrees - how steep a descent the impactor made through the atmosphere); horizontal angle of impact (0 to 360 degrees - what direction the impactor approached from); speed of the impactor (roughly 10,000 to 70,000 miles per hour); size of the impactor (diameter in miles of an idealized sphere - probably one half mile to five miles, although some work remains to be done in limiting the range); density of the impactor (measured in tons per cubic meter - this varies widely depending on whether the impactor was a comet - so largely water ice - or what type of rock the meteor was made of) and what sort of target the impactor hit - all deep water; a mixed mode coastal area, a desert (mostly sand) or all rocky land.

The model must then predict crater size and the distribution of tektites. The former is possible. The latter has a set of challenges all its own. As some seventy (70) percent of the surface of Planet Earth is covered by water it is reasonable to expect that not all major impacts have been found yet. Even relatively dry land holds surprises: for example, Luizi crater, with a diameter of 17 kilometers, was just recently discovered in the Democratic Republic of the Congo. Apparently, not all major impacts produce tektite strewn fields. There are currently 11 fields known - from almost 160 craters. Tektites from the Australasian impact were found several years ago in Antarctica meaning the Australasian field covers close to 15% of the planet's surface. For the Chesapeake and Australasian strewn fields there are multiple distinct types of tektites. Then there are geological consequences - would there have been massive fires, volcanic eruptions, earthquakes and megatsunamis? If so, can evidence be found so many years later? A lot of computing. 

To complicate matters, preliminary indications highlight a performance difference between Windows 7 and Windows 8, especially in multi-core configurations with aggressive multi-threading. However, instead of decreasing, the variance of execution among seemingly identical threads increases.