On Tue, Jan 15, 2002 at 12:20:25PM -0500, Hubertus Franke wrote: > Conclusion > ---------- > Manfred Spraul's new zero copy patch showed a very good performance > improvement for pipeflex as well as grep on both 2-way and 1-way. > By adding our large pipe support, performance improvement up to > 42% for pipeflex and 96% for grep benchmarks on 2-way systems. > No significant different has been observed for lmbench on 1-way systems. > The right way to configure the system than would be for UP and 1-way to > set the buffer size by default to 4K while for the SMP larger pipe buffers > should be encouraged. Any conclusions are incomplete without mentioning the serious (30%+) degredation on UP systems under a good chunk of the benchmarks. Such aspects of the patch make it unsuitable as is for both the mainstream and vendor kernels. -ben - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo _at_ vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Follow-Ups:
- Re: [Lse-tech] zerocopy pipe, new version"Manfred Spraul" <manfred _at_ colorfullife.com>
- zerocopy pipe, new version"Manfred Spraul" <manfred _at_ colorfullife.com>
- Re: [Lse-tech] zerocopy pipe, new version"Hubertus Franke" <frankeh _at_ watson.ibm.com>
- Prev by Date: high resolution timer for packet timestamping
- Next by Date: Re: Penelope builds a kernel
- Previous by thread: Re: [Lse-tech] zerocopy pipe, new version
- Next by thread: Re: [Lse-tech] zerocopy pipe, new version
- Indexes:[Main][Thread]