i'm working on a project that makes use of SSE2 on a Dual Xeon 2Ghz workstation. During the debugging session using MS visual studio, only 8 XMM registers are visible. so as far as dual cpu is concerned, how many XMM registers are actually available? right now the performance turns out to be quite the same as running on a single P4 CPU with similar speed. thus it indicates during the execution of SSE2 code portion, only one of the CPUs is working, am i right?
if i am able to make the other CPU work on the SSE2 code at the same time, the program is ought to run at about twice as fast. so i'd like to ask how do i address the other set of 8 XMM registers and as well as other registers that are physically existing on the second processor, if provided they exist? is openMP an option?
SSE2 programming on Xeon Dual CPU