Subj : Re: AMD has cmpxchg16b now - sort of To : comp.programming.threads From : Joe Seigh Date : Tue Mar 08 2005 06:24 pm On 08 Mar 2005 17:52:02 -0500, Nathan J. Williams wrote: > "Joe Seigh" writes: >> The synchronization stuff for powerpc was pretty well in place from the >> very beginning. > > Synchronization, sure, but there are lots of other features one is > interested in testing and that have varied across PowerPC > implementations (FPU, SIMD/vector instructions (including, but not > limited to AltiVec), string instructions, handling unaligned data > automatically, and so on). > All of those things can be simulated, albeit with a possibly significant performance hit. Significant enough that they'd likely advertize or explicitly document the fact. Why else go through the trouble of putting the feature in. The synchronization stuff can't be simulated and AFAICT not documented on Intel's or AMD's web site. Though I suppose I could email some Intel or AMD marketing type and ask for quotes for quantities of 100000 of processors that have that feature. :) -- Joe Seigh .