This is currently a limitation of the UD Agent software, not Rosetta. There are 3rd party tools which get around this limitation, but we do not recommend their use. Use them at your own risk.

should have known this is a MicroSoft problem. :-)

Intel hyperthreading is very peculiar. The idea is to keep a long pipeline full. Sometimes a pipeline stalls because the next instruction in a thread requires resources in use. The answer is to add about 5% to the size of the core and have a second thread trying to execute. If the next instruction for that thread can execute, then great!! It executes instead. This increases total throughput

The problem is that 2 threads are executing on the same processor using the same cache. The hardware cannot tell them apart, since they are on the same processor, so there is no mediation circuitry balancing their cache requests. In the worst case, the cache thrashes and total throughput actually decreases. Some tasks, like video editing, are very easy to hyperthread and can show a 50% speedup with hyperthreading. In general, people expect speedups of 10%-30% (or another often quoted rule of thumb is 14%-26%). The trouble is that this puts the burden on the programmer to follow arcane rules to try and make the cache behave.

Now contrast this with a multi-core cpu. There are several cores accessing the cache through standard hardware mediation circuitry which does a fair job of load-balancing with little if any burden placed on the programmer. And the result should be speeds at least as good as the best that hyperthreading can deliver.

So the 50% usage that Task Manager shows simply means that only 1 thread is executing, but probably the power expended is about the same as 100% usage with 2 threads executing. Actual throughput is probably not greatly less than it would be with 2 threads. Hyperthreading is probably the slowest method of running simultaneous multiple threads, and one of the more demanding ways from the perspective of a programmer.

I can use realtime priority class, maybe because my front side bus is 800 am running ddr500 (only shows 400) on a deskboard d875pbz with 2 gigs of memory ... however either on low priority or real time priority the amount of memory meg's holds at about 29 megs for memory usage. so I'm saying it's possible to use realtime and also saying your right in suggesting someone system may degrade without the proper equipment ThankYou for your memo swl