| 1 |
OpenMosix is not a viable option for compiling with "make -j". |
| 2 |
The threads of a source file building do not exist long enough |
| 3 |
to be migrated. I tired it... Didn't do very well. |
| 4 |
|
| 5 |
> Paul de Vrieze wrote: |
| 6 |
> > On Sunday 09 June 2002 22:31, Marko Mikulicic wrote |
| 7 |
> > |
| 8 |
> > |
| 9 |
> > While I didn't test it yet, openmosix could be a solution for better |
| 10 |
> > utilization of your computing power. Parallel dependencies are sometimes |
| 11 |
> > difficult (cross dependencies), also it would require tweaks to emerge |
| 12 |
> > |
| 13 |
> |
| 14 |
> Well, I delibertely left out openmosix because it is automatically |
| 15 |
> supported with make -j n. However there is a big issue with it: |
| 16 |
> Openmosix process migration can distribute the load of a single system |
| 17 |
> through a cluster but if you install n packages they are build in |
| 18 |
> sequence so that only compiles from a single build get eventually |
| 19 |
> distributed. And it's even worst when some packages have parallel build |
| 20 |
> disabled. |
| 21 |
> Moreover openmosix is not applicable in a etherogeneous network, whith |
| 22 |
> freebsd, linux, win32, tru64, linux-alpha machines, while distcc can |
| 23 |
> execute cross-compilers on every architecture (windows too). |
| 24 |
> and openmosix requires all to run a special kernel, administrative |
| 25 |
> issues,.... I think that something lighter is needed, for it to be |
| 26 |
> really useful. I have access to a LAN with 16 PIV 1.5Gh but with any |
| 27 |
> kindof linux distros, versions, libraries, users. The only |
| 28 |
> administrative power I have is to run niced jobs and a good amount of |
| 29 |
> ram. |
| 30 |
> |
| 31 |
> However the problem is not openmosix vs. distcc, while both are |
| 32 |
> completely transparent with portage and both have the same problem: |
| 33 |
> sequential builds. |
| 34 |
> |
| 35 |
> As for cross dependencies, how does portage handle the now? There could |
| 36 |
> a queue for each pending build where dependent build are added. |
| 37 |
> RDEPEND does not conflict with paralell builds. |
| 38 |
> I agree this would require tweaks to emerge. I think it's not |
| 39 |
> urgent in portage1, but it could go in portage2, which with it's added |
| 40 |
> performance could resolve this kind of dependence in acceptable time. |
| 41 |
> What do you think ? |
| 42 |
> |
| 43 |
> Marko |
Archives