Messages in this thread | | | Date | Wed, 4 Oct 2006 20:50:13 -0700 | From | Andrew Morton <> | Subject | Re: [PATCH] mm: balance dirty pages |
| |
On Thu, 5 Oct 2006 04:03:35 +0400 "Ananiev, Leonid I" <leonid.i.ananiev@intel.com> wrote:
> The throughput of iozone benchmark is changed as > > serial random > write -55% -99% > read +60% +13% > > after "[PATCH] mm: balance dirty pages". > iozone is running with option -B (using mmap) for file size 120% of RAM. >
That is expected and intentional. Previously, `iozone -B' was able to swamp the whole machine with dirty memory. That gives good benchmark numbers, but causes everything else on the machine to be adversely affected.
The new behaviour will punish iozone for being bad, and will hopefully cause less damage to other applications. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |