> I've complained about the same issues before. I > understand the need to move onto Xorg and drop Xsun. > At the same time, I understand that Sun really isn't > in the SPARC workstation business anymore and is > hoping that people will buy a Intel PC from them for > running Solaris 10 or OpenSolaris on today. The > reality is that only folks who want a complete > support contact on the desktop hardware and OS will > bother, because it runs on a wide variety of PC > hardware. By doing this Sun has killed it's own > workstation market which is drying up and obviously > pissed off all of it's ISV's, vendors, and > integrators that marketed, developed, and resold > SPARC workstations. This even forced me to bite the > bullet and buy a used Ultra20 to use OpenSolaris. > Notice, used.. so no money went back to Sun! Hell, > I'll probably just keep the case and upgrade > everything inside of it! This is the future Sun has > chosen outside of servers and thin clients, limited > desktop/workstation > sales. Not to mention that Sun has killed the > ecosystem for SPARC by making the entry-point > extremely expensive.. $12,795 for a T5120 > (http://www.sun.com/servers/coolthreads/overview/prod > cts.jsp) and $14,795 for a M3000 > (http://www.sun.com/servers/midrange/m3000/index.xml). > > So you have to wonder what is the thinking here? Are > x64 server sales out-pacing SPARC for Sun? Hate to > burst the bubble of every Sun employee working on > Solaris/OpenSolaris with their x64 laptop.. NO! The > lion's share of server sales are SPARC, specifically > the T-Series CMT servers! So how does this make > sense? The sales are on SPARC, but the majority of > the development effort with OpenSolaris has been > focused on x64. So Sun has to decide if they want to > keep soo much of the development effort with > OpenSolaris focused on x64, which brings in little > profit or if they want to focus on the needs of their > paying customers. Hey, I'm not saying the stuff that > has been done on OpenSolaris doesn't matter, it's > great work. But focus has been lost on turning that > into profit and you have to wonder at the end of the > day how Oracle will look at this. They have committed > themselves to investing more into SPARC, because they > know that's where the money is. > > Sun/Oracle needs to bring back SPARC workstations and > rebuilding the ecosystem. > > Here's a fun test that shouldn't cost millions for > Sun to do today. How about get Nvidia or ATI to build > a PCI and PCI-E 3D video card that will work with > OpenSolaris Xorg on SPARC. Make it work on UltraSPARC > III/IIIi workstations and sell it for under $300. I'm > willing to bet that this would sell like crazy and > make tons of people happy. It would also let Sun see > that there is still a large SPARC workstation market > out there that they have been ignoring. This isn't > rocket science, look on Ebay and check out the number > of SPARC workstations being sold every week! And why? > Because businesses and professionals need a SPARC > workstation.
While I sympathize, one thing you're overlooking is those systems for which the UPA slot(s) are the fastest; I'm in bad shape, because my only 66MHz slot (SB1K/SB2K) has a SAS controller, so all my other slots are either 33MHz PCI, or UPA. Another thing is that the XVR-100, although only 2D and not very fast, AFAIK should work with Xorg. Not available new anymore, but probably not too hard to find used. Last but not least, to use it as console, a SPARC frame buffer needs on-board fcode for OpenBoot; a PC frame buffer would typically have a video BIOS or some such support more suitable for a PC. Could Xorg be built to handle a particular list of loadable (separate) community-supported drivers, so that those could be added without replacing Xorg itself? That wouldn't be perfect, but it would be a lot better than nothing. -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
