I'm actually seeing 2.5.3.3 running faster than 2.5.1.4 as well. It is other users who have reported that 2.5.3.3 is slower than 2.5.1.4, and I cannot seem to cause that to happen unless I use online symbols for 2.5.3.3 and then use the local cache for 2.5.1.4
Maybe others had issues with the symbols download or DRT download when using 2.5.3.3 and that resulted in laying blame on the apps?
Based on what was changed for 2.5.3.3, more resources are now available for the kernel debugger, so I would expect it to be faster, not slower. You and I are finding that is the case; not sure why others are seeing slower runtimes with 2.5.3.3...