2015/05/24 13:08 "Warner Losh" <i...@bsdimp.com>: > > > > On May 23, 2015, at 4:46 AM, Joel Rees <joel.r...@gmail.com> wrote: > > > > > >> Warner suggested that emulated hardware could be faster than the native > > hardware being used. If emulated hardware were twice as fast, would that > > justify using it in addition to native builds? > >> > > > > If only such theoretical speed advantage were real, it might be worth > > talking about. Maybe. On the other hand, the openbsd project never used > > emulation on the 68k platforms, where the clockspeed difference was very > > real. > > The speed advantage is real. > > To build the same set of packages natively for amd64 it takes 12 or > so hours.
Irrelevant. > To build them using user-level emulation of the native binaries > takes about 30 hours, though some of the binaries are native producing > target binaries. If you are doing this with openbsd packages, where are the patches? > To build 100% native on something like a Wandboard > takes more like 20-30 days, since swapping hurts a lot... Irrelevant to your question. > The advantage for this is two fold: memory and core count. While you > can do something about the core count by running lots of boards in parallel, > that takes a more complex build system.You can’t do much about memory, > however. When you have a machine with 256GB of RAM, clang runs much > better than it could ever run on an arm board with 2GB of RAM. The native > binaries help, and the user-level emulation copes with those silly programs > that expect to run binaries in the target environment to produce their output. Okay, you have the experience doing this kind of thing for some other system? Great. Give it a try for openbsd. On your time, on your dime, on your hardware. Then tell the devs here what you get. And give them the patches to try on their stuff, to see if your setup actually helps flush out real bugs. > But hey, I’m not trying to tell anybody what to do. I’m just giving you data > from what a sister project has been able to do lately. They've seen false positives like this before, and they have mentioned that fact in this thread. > Warner Of course, I'm wondering what your host OS will be, at what layer you plan to start library emulation, and such things. And whether you are asking for packages compiled on a foreign host OS to be made available to be downloaded via pkg_add without re-compile on real hardware. (Hello, Ken Thompson?) I'd personally prefer to be spending my time adding physical memory to certain of the existing devices, over getting emulation on a foreign processor host running. But neither of us are providing patches to the project in the near future, are we? -- Joel Rees
