Messages in this thread | | | Subject | Re: hackbench [pthread mode] regression with 2.6.29-rc3 | From | Peter Zijlstra <> | Date | Mon, 02 Feb 2009 21:50:48 +0100 |
| |
On Mon, 2009-02-02 at 17:49 +0000, Bryon Roche wrote:
don't drop CCs
> On Mon, 02 Feb 2009 09:53:38 +0100, Peter Zijlstra wrote: > > > On Mon, 2009-02-02 at 09:12 +0800, Zhang, Yanmin wrote: > > > > I would rather go back to the old model where we iterate all threads, > > and find a way to not make programs with too many threads for their own > > good lock up the kernel, but instead get poor service. > > Now, there's an interesting question, what is the definition of too many > threads for a program's own good? When evaluating this, please assume > that you > do actually have enough RAM to keep thread stacks/other userspace > threading > resources in-core.
In my book a program has too many threads when it outnumbers the number of cpus available to it by an order of so.
| |