Tamarin and SpiderMonkey fail

Apparently, Tamarin and SpiderMonkey can't handle large datasets. When running our port of JavaGrande, at size C, both Tamarin and SpiderMonkey crash my system. Rhino won't run on my machine since it uses more than 1.5 gigs of ram at size C. However, size B should be enough :). We can just run size C of rhino versus our generated bytecode on xen.