Subj : Re: Hazard Pointers w/o memory barrier. To : comp.programming.threads From : Joe Seigh Date : Wed May 04 2005 04:17 pm On Wed, 04 May 2005 07:00:24 -0400, Joe Seigh wrote: >> >> Ah wait, I must have been looking at the write rate. It >> gets well over 500,000 reads/sec/thread at the default >> poll interval, 50 msec. That's more reasonable. Now >> to try it on Linux to see how it works there. Linux >> signaling is preemptive I believe. >> > > Mystery cleared up. I was getting read rates of > 9 reads/sec/thread on occasion. It seems that > even though I'm using a barrier to start the readers > and writers all at once, sometimes the writers can > get way ahead of the readers. Increasing the duration > averaged things out better. The read rates works out > to be around 1,500,000+ reads/sec/thread in that case. > > It's a shock going back to Linux with its poor throughput rate compared to OS X. And signals on Linux definitely slow things down. So we can forget about a common reference implementation. Platform specific implementations are the way to go. Nothing to see here. Move along. :) -- Joe Seigh When you get lemons, you make lemonade. When you get hardware, you make software. .