Benchmarking.NET.
One of the projects we are working on is the benchmarking of .Net, SSCLI and other CLI vm's from the perspective of high-performance computing. Chris Re has ported the Java Grande sections and a few other benchmarks to C# and we have developed a some micro-benchmarks to look at particular details. We are now validating whether the numbers we are getting make sense and are reasonably accurate.
<snip />
C# on the 1.1 .NET framework performs consistently better than Java on the best VM I know of (IBM 1.3). On the small memory models the C# version the composite result is about 20%-25% better than the Java one.
<snip />
[All Things Distributed]
Interesting results. It will be good to see more information as to what the details of the different areas of performance are.
I've not gotten around to running my highly scientific "Run Freespace 2 under the CLR" test to see how performance measures up to the unmanaged version so I'll do it tonight.
What other blogs are saying about this post.
11:14:51 AM 
|
