An interesting tidbit... I was talking to the JavaFX RE team which does builds on Windows XP and Windows 2003 X64. They have always used CYGWIN for builds, but there isn't many Makefiles used on Windows, it's mostly Ant scripts. So CYGWIN is used but probably not a major player. No MKS was ever used. Both windows builds had been the long pole, maybe 2 hours each, and past experiments showed that using physical hardware rather than VMware virtual machines could reduce the builds to about 1hour. So we knew that VMs would cost us, but assumed that as we got faster hosts this would improve.
After McAfee was installed, the Windows XP builds went down the toilet, but the Windows X64 builds were not impacted as much. Apparently the policies in place for Windows XP (considered a desktop environment) are considerably different from that of Windows 2003 (considered a server environment). Another... I now have 3 Windows 7 X64 systems I can now use to test jdk8 builds. I built the entire openjdk8 (hotspot included, product only) in slightly under an hour, just the Windows 64bit product. This is a VMware VM on a fully loaded X4170M2 system hosting just 6 VMs, each VM has 4 vCPUs with lots of RAM. This was using MKS 9.0p4 I can use it to do some additional experiments on Volker's changes, but it will take me a little time. -kto On Feb 15, 2012, at 4:10 AM, Fredrik Öhrström wrote: > ----- [email protected] skrev: > >> So I'm with you on the stat() theory, makes a great deal of sense. > > The stat theory is very interesting, but it is unclear to me if it explains > all of the problem. > > I setup a quadruple boot x86_64 machine with 4GB of ram and 4 cores: > Winxp 32bit > Win7 64bit > Solaris 64bit > Ubuntu 64bit > > And tested the build times on the different OS:es. > > Ubuntu Fastest by far. > > Solaris, slower, but this is only because of bad CC performance. > > Winxp, even slower but still ok. > > Win7, ridiculously slow. The configure script prints one line per second! > > Clearly, just running a bash script in cygwin/win7/64bit is problematic. > If we get 10% speedup from dash, then that is not going to help because > the slowdown is a factor 10. > > Could someone try out the difference between a 32bit win7 clean install and a > 64 bit win7 clean install when running the latest cygwin and just the > build-infra/jdk8/common/autoconf/configure script? > > (My patience for installing many OSes into the same box, just ran out. And > virtualization > testing can give a hint, but cannot be entirely trusted.) > > //Fredrik
