Stewart Stremler wrote: > begin quoting Gregory K. Ruiz-Ade as of Fri, Mar 30, 2007 at 07:07:14PM > -0700: >> On Mar 30, 2007, at 2:43 PM, Stewart Stremler wrote: >> >>> I think we need some more languages with built-in threading support, >>> 'cuz the future is looking like multi-core shared-memory systems. >>> We're >>> going to be dealing with concurrent programming issues no matter what. >>> We should be exploring that problem space, not avoiding it. >> Not just multi-core, but Intel now is following Sun's lead (though >> I'm left to ask why Intel didn't just lead themselves) in adding >> Symmetric Multi-Threading (a.k.a. HyperThreading) back into their >> processor architectures. In a couple years, you won't be able to buy >> an Intel CPU that's not multi-core and multi-threaded-per-core. > > I wonder if this is a factor in introducing threads into the TCL core? > If they can show a major speedup by changing to a new CPU, it'll > do a lot for TCL among the performance-minded crowd.
Given the overhead TCL imposes, I find it hard to believe that this is really about major performance improvements (process overheads just aren't that major). I'd buy it more along the lines that it just made it easier to solve certain kinds of problems. >> Sun's Niagra CPU packages already have 8 cores with 2-4 threads per >> core (can't recall the exact numbers), and they're loved by people > > At SCALE, they said 4 threads per core. Damn, you were there and I didn't see you! ;-) >> running large-scale databases (which, I guess, would really mean >> Oracle). > > Heh. > > I wish they'd make a Blade 100-class machine (cheap-as-in-beer) > with a Niagra CPU. Niagra isn't exactly all that. I had several people warn me that their experiences with it had not exactly lived up to the hype. It appears to very much be a case of YMMV. --Chris -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
