Subj : Re: RCU+SMR To : comp.programming.threads From : Randy Howard Date : Fri Jul 08 2005 06:58 pm Joe Seigh wrote (in article ): > Randy Howard wrote: >> >> That's why I was wondering about standardization. With >> all the pushing for SMP, dual-core, multi-core, (and the >> abomination known as "hyperthreading", you would think some of >> the big names (like Intel and AMD) would be pushing for ways to >> get the most out of new processor features and parallelism >> without everyone having to roll their own. > > Intel and AMD are basically hardware companies and are run by > hardware not software types. True, but I think they (especially Intel) realize that the profit is in systems, not isolating hardware from software. IOW, recognizing that they go together. Intel has had a lot of effort in software over the years, including compilers, IDF, software development forums on their website, reams of documentation, white papers, example code, etc. They have been harping on threading and the need to develop parallel solutions for a long time. > They know HPC (high performance computing) and parallel programming. Exactly. If they have the energy to push something like OpenMP and their own compilers, then working on pushing lock-free adoption doesn't seem /that/ unlikely. > So while you could do some really neat things with hyperthreading, > no one will because it will be gone by the time the OSes could > provide an api to exploit it. I hope so. Hyperthreading is beneficial in a very few special cases, but in general it sucks compared to SMP or multi-core solutions. I suppose keeping it around inside of multi-core processors isn't horrible, as long as you can turn it off. In some cases, it makes things much worse enabled than disabled. > It's not just Intel and AMD. True, but they represent the bulk of the market. If they go in a direction, then it can happen. If they don't, then the idea will die on the vine, more often than not. > Sun Research has a whole group that does synchronization, mostly > lock-free. But I have yet to see any evidence that Sun exploits any of this. If sun doesn't use it themselves, you can't expect any of their customers or developer community to do so. They *could* push it, but they've got bigger fish to fry right now trying to keep Java alive in the onslaught of alternatives coming along now. > And they're coming out with Niagara which will have up to 8 cores, > 4 threads per core. You would think. I hope those folks in India and China are up on threaded programming. :-) >> Oh well, I guess it will be a long time before this sort of >> thing is widely used enough to become more generally applicable >> to portable software. >> > Well, a killer app will speed things up. True. I've yet to hear of any commercial or popular OSS application bragging about the massive performance benefits achieved through the use of lock-free techniques. Real-world examples convince people more than research papers. Perhaps it is being widely used but not talked about, or I just missed the claims? -- Randy Howard (2reply remove FOOBAR) .