On Fri, Apr 25, 2008 at 11:35 PM, Andrew Wilson <[EMAIL PROTECTED]> wrote: > > BTW: there is still some inconsistence through all these documentation: > > http://www.solarisinternals.com/wiki/index.php/FileBench > > http://www.opensolaris.org/os/community/performance/filebench/ > > and some docs in the source tree; > > 1. what is the binary f scripts interpreter's name: go_filebench or > filebench? > > > The f scripts are now interpreted by go_filebench. "filebench was the old > name for it" > > > 2. and the perf(or conf) files driven perl script: filebench or > > runbench or benchpoint? There are several different named filebench > > perl scripts, which is the correct one? Could you please the > > maintainer remove others? > > > Use filebench. it used to be called runbench, I think, but has been renamed > filebench. > Thus, use "filebench <xxx.prof>" to run suites of benchmarks, > "go_filebench" to interactively run a single .f workload file, as in: > %go_filebench > filebench> load randomread > filebench> run 60 > [lots stuff prints out here, including the results] > filebench> quit if go_filebench needs to run without absolute path name(/usr/.../go_filebench), the execl call in the source could be changed to execlp;
> > 3. Please give some more description about every part of the source > > tree, which provided on the sourceforge? > There is a fairly detailed description of how filebench works on > SolarisInternals.com. Also documentation on the "f" language is there. you mean the wiki documentation are the latest? http://www.solarisinternals.com/wiki/index.php/FileBench so many docs in the source tree and comments on code need to be fixed, (especially `go_filebench -h`, it's much misleading and have confused me). On Sat, Apr 26, 2008 at 1:14 AM, Spencer Shepler <[EMAIL PROTECTED]> wrote: > > > On Sat, Apr 19, 2008 at 7:17 AM, eric kustarz <[EMAIL PROTECTED]> > wrote: > Hi. Would you be willing to complete an OpenSolaris contributor agreement? > The process is described here: > http://opensolaris.org/os/about/sun_contributor_agreement/ > > Even though you are contributing for the Linux port, the source will > be common between all platforms and the Sun/OpenSolaris contributor > agreement is how we are managing contributions to source changes > at this point. The Instructions for submitting the agreement might be too complex, do you agree? I just work with the origin of filebench-1.2.4 tarball from sourceforge, after some effort run it on gentoo linux, now have some patches, I want to send the patches to upstream; so if sun people are willing, you can review/integrate the patch or then add me to the developer list on sourceforge. The CDDL License of this agreement is not the key, since I sent patches, I have agreed it. Or if you are willing, you can switch to GPL. On Sat, Apr 26, 2008 at 7:56 AM, Richard McDougall <[EMAIL PROTECTED]> wrote: > I had to add quite a bit of code to make the database (oltp) workload run > properly on Linux, around the way that I/O is submitted (using Oracle's > io_submit instead of aio_write), and the way shared memory is allocated and What do you mean by Oracle's io_submit? Linux currently(from 2.6.1x) have a series of native aio interface (io_setup/io_submit/...) implemented with system call, that's not Oracle's. So the Linux have two types of aio: the native aio implemented with system calls and POSIX aio implemented with pthreads, filebench currently only support POSIX aio, I think filebench could also add native-aio as a future work TODO entry. > used. I also added a feature to enable raw: syntax in the filenames, so that > we can use raw devices for the database workload. > > The makefiles/autoconf stuff isn't an issue (I didn't make any specific > changes there). The original configure/Makefile.am have some problems to conform the Linux FHS, so I regenerated them using autoconf/automake. > > We need to integrate these set of changes to enable Linux to be complete... > > Richard. -- Cheng _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
