[OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Hi, Upgraded to OI 151 things for OI are looking better with each release.:-) I have a question, it appears Oracle has removed 32 bit hardware support for the x86 CPU for solaris 11. so if oracle do release solaris code after solaris 11 is out, it will most likely only for 64bit only. so my question is if things do indeed go that way will OI continue 32bit support in future oi releases or do 64bit only to hold some compatibility with solaris 11 or will I have to toss out my Pentium m laptop along with other units. Functionality: * 32-bit Kernel click here to expand or collapse Support for the 32-bit only x86 hardware has been removed. Support for running 32-bit applications and libraries will continue http://www.oracle.com/technetwork/systems/end-of-notices/eonsolaris11-392732.html ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
One that strikes me as odd on that list is under "OpenWindows Libraries" "... However, if required, the applications that use OpenWindows Libraries can be run in Oracle Solaris 10 Zones" I thought that the renumbering of zones was just that and not a rebranding with libraries ... in fact I was pretty certain that that was advised against. does anyone know if we still have access to the old legacy drivers (not necessarily the old graphics drivers) ... in case we need to support legacy hardware in the future? Jon On 22 June 2011 02:06, Edward Martinez wrote: > Hi, > > Upgraded to OI 151 things for OI are looking better with each release.:-) > I have a question, it appears Oracle has removed 32 bit hardware support for > the x86 CPU for solaris 11. so if oracle > do release solaris code after solaris 11 is out, it will most likely only > for 64bit only. so my question is if > things do indeed go that way will OI continue 32bit support in future oi > releases or do 64bit only to hold some compatibility with solaris 11 or will > I have to toss out my Pentium m laptop along with other units. > > > > > Functionality: > > > * 32-bit Kernel click here to expand or collapse > > Support for the 32-bit only x86 hardware has been removed. Support > for running 32-bit applications and libraries will continue > > > http://www.oracle.com/technetwork/systems/end-of-notices/eonsolaris11-392732.html > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On Wed, Jun 22, 2011 at 4:05 AM, Jonathan Adams wrote: > One that strikes me as odd on that list is under "OpenWindows Libraries" > > "... However, if required, the applications that use OpenWindows > Libraries can be run in Oracle Solaris 10 Zones" > > I thought that the renumbering of zones was just that and not a > rebranding with libraries ... in fact I was pretty certain that that > was advised against. "Oracle Solaris 10 Zones" are branded zones that run on Oracle Solaris 11 Express 2010.11 and the notice you cite implies will continue to run on Solaris 11. This is analogous to the Solaris 8 Containers and Solaris 9 Containers products that used branded zones to run Solaris 8 and Solaris 9 userland bits on Solaris 10. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
32-bit still makes sense for an appliance. Unless you're seriously _poor_, I don't see that it matters if it won't run on some old dumpster diving relic. And a point against it: UFS root support is also said to be going away. That means ZFS root instead. And ZFS is really happier if you have lots of memory. Yes, it's running now on 32-bit, but I bet it's not being pushed really hard with lots of data. Keeping it would be a divergence large enough to make further divergence almost inevitable, too. Which leads to the point that I'd rather defer something like that until we see whether Oracle dumps some more code over the wall after Solaris 11 is no longer "Express" (beta). They had said that they'd basically sit on code until there was a commercial product, so that competitors weren't profiting from their work before they were. If they stick to that, then staying reasonable close will be possible, if more tedious than if they were more participatory in their approach. If they fail to stick to that, then aside from being liars (or not following through, if you want a less charged description), there wouldn't be much point to worrying about what they were doing anyway, aside from trying to keep the ABI and on-disk structures compatible at some common subset of features, and maybe trying to limit or manage other application and user-visible divergence. On Jun 21, 2011, at 9:06 PM, Edward Martinez wrote: > Hi, > > Upgraded to OI 151 things for OI are looking better with each release.:-) > I have a question, it appears Oracle has removed 32 bit hardware support for > the x86 CPU for solaris 11. so if oracle > do release solaris code after solaris 11 is out, it will most likely only for > 64bit only. so my question is if > things do indeed go that way will OI continue 32bit support in future oi > releases or do 64bit only to hold some compatibility with solaris 11 or will > I have to toss out my Pentium m laptop along with other units. > > > > > Functionality: > > > * 32-bit Kernel click here to expand or collapse > > Support for the 32-bit only x86 hardware has been removed. Support > for running 32-bit applications and libraries will continue > > > http://www.oracle.com/technetwork/systems/end-of-notices/eonsolaris11-392732.html > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > -- The waitress asked, "Do you want lemon or no lemon with that iced tea?" Naturally, I said "yes", and then burst out laughing, because there simply wasn't any other answer in Boolean logic. She didn't get it, but I got the lemon, which I wanted anyway. Later, I realized a quantum computer could have offered another answer: Schroedinger's Lemon! ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
In general, OpenWindows means two things to me: libraries: libxview (XView toolkit), libXol (OLIT toolkit), and associated libraries (libolgx and so on). The Open Look apps went away at I think Solaris 9 (although most of them would still run on Solaris 10 if one brought forward all the right bits), but the libraries had been kept around for any 3rd party or in-house apps that linked to the shared libraries. Xsun: biggie for those of us with older SPARC hardware, lacking Xorg graphics driver support I may be wrong, but I thing I've seen things implying that at least a modest effort may be made to see that Xsun runs in a branded zone with the proper setup, for those on such old hardware. Of course, that wouldn't be open source (mostly) or redistributable (at all). On Jun 22, 2011, at 5:05 AM, Jonathan Adams wrote: > One that strikes me as odd on that list is under "OpenWindows Libraries" > > "... However, if required, the applications that use OpenWindows > Libraries can be run in Oracle Solaris 10 Zones" > > I thought that the renumbering of zones was just that and not a > rebranding with libraries ... in fact I was pretty certain that that > was advised against. > > does anyone know if we still have access to the old legacy drivers > (not necessarily the old graphics drivers) ... in case we need to > support legacy hardware in the future? > > Jon > > On 22 June 2011 02:06, Edward Martinez wrote: >> Hi, >> >> Upgraded to OI 151 things for OI are looking better with each release.:-) >> I have a question, it appears Oracle has removed 32 bit hardware support for >> the x86 CPU for solaris 11. so if oracle >> do release solaris code after solaris 11 is out, it will most likely only >> for 64bit only. so my question is if >> things do indeed go that way will OI continue 32bit support in future oi >> releases or do 64bit only to hold some compatibility with solaris 11 or will >> I have to toss out my Pentium m laptop along with other units. >> >> >> >> >> Functionality: >> >> >> * 32-bit Kernel click here to expand or collapse >> >> Support for the 32-bit only x86 hardware has been removed. Support >> for running 32-bit applications and libraries will continue >> >> >> http://www.oracle.com/technetwork/systems/end-of-notices/eonsolaris11-392732.html >> ___ >> OpenIndiana-discuss mailing list >> OpenIndiana-discuss@openindiana.org >> http://openindiana.org/mailman/listinfo/openindiana-discuss >> > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > -- The waitress asked, "Do you want lemon or no lemon with that iced tea?" Naturally, I said "yes", and then burst out laughing, because there simply wasn't any other answer in Boolean logic. She didn't get it, but I got the lemon, which I wanted anyway. Later, I realized a quantum computer could have offered another answer: Schroedinger's Lemon! ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On 06/22/11 05:33 PM, Richard L. Hamilton wrote: > And a point against it: UFS root support is also said to be going > away. UFS root is long gone - IPS only runs on a ZFS root, since it relies on ZFS snapshots/boot environments. No OpenSolaris release ever had UFS root support, nor does OpenIndiana or Solaris 11 Express. -- -Alan Coopersmith-alan.coopersm...@oracle.com Oracle Solaris Platform Engineering: X Window System ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
FWIW, Mac OS X Lion will only support x64 as well. IMHO, this is a good move for modern operating systems since there are always going to be alternatives for those still using i386 architecture. How long has Solaris/SPARC been 64-bit? At least ten years if not more... -Gary ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On 06/22/11 06:03 PM, Richard L. Hamilton wrote: > Xsun: biggie for those of us with older SPARC hardware, lacking > Xorg graphics driver support > > I may be wrong, but I thing I've seen things implying that at least > a modest effort may be made to see that Xsun runs in a branded zone > with the proper setup, for those on such old hardware. Of course, that > wouldn't be open source (mostly) or redistributable (at all). I'm not sure who would be doing that - I released most of the Xsun source as open source last year after multiple people said they'd use it to provide continued support for the old graphics, and so far no one seems to have done anything with it, so there seems to be very little actual interest in doing more than talking about how nice it would be if someone else did the work to support old graphics. -- -Alan Coopersmith-alan.coopersm...@oracle.com Oracle Solaris Platform Engineering: X Window System ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Wow. They killed a lot of stuff. Not only 32-bit x86 support but tons of other stuff too. SPARC Workstation support has been killed off (no more UltraSparc I/II/III/IV support, no more Xsun and no more hardware accelerated OpenGL for SPARC) and a lot of legacy peripheral support for both x86 and SPARC is gone. Hopefully, OI won't be following along this path. ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Closer to twenty. And my Ultra 1 is doing fine. Sent from my iPad On Jun 22, 2011, at 9:10 PM, Gary Driggs wrote: > FWIW, Mac OS X Lion will only support x64 as well. IMHO, this is a good move > for modern operating systems since there are always going to be alternatives > for those still using i386 architecture. How long has Solaris/SPARC been > 64-bit? At least ten years if not more... > > -Gary > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On 23/06/11 11:52 AM, Michael Kerpan wrote: Wow. They killed a lot of stuff. Not only 32-bit x86 support but tons of other stuff too. SPARC Workstation support has been killed off (no more UltraSparc I/II/III/IV support, no more Xsun and no more hardware accelerated OpenGL for SPARC) and a lot of legacy peripheral support for both x86 and SPARC is gone. Hopefully, OI won't be following along this path. what, you still want to run parallel scsi 32bit-only drivers like ncrs, which are poor (very very poor) cousins of glm? McB. ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On Wed, Jun 22, 2011 at 10:03 PM, McBofh wrote: > On 23/06/11 11:52 AM, Michael Kerpan wrote: >> >> Wow. They killed a lot of stuff. Not only 32-bit x86 support but tons >> of other stuff too. SPARC Workstation support has been killed off (no >> more UltraSparc I/II/III/IV support, no more Xsun and no more hardware >> accelerated OpenGL for SPARC) and a lot of legacy peripheral support >> for both x86 and SPARC is gone. >> >> Hopefully, OI won't be following along this path. > > > what, you still want to run parallel scsi 32bit-only drivers like > ncrs, which are poor (very very poor) cousins of glm? > > McB. I can almost see dumping 32-bit x86. but dumping 64-bit US-III/IV? Clearly, Oracle is going to use this as a stick to force customers to upgrade to new hardware, or away from Solaris/Oracle completely while they overcharge for "legacy Solaris 10 support". Good business strategy. :-p Lines from "Other People's Money" keep rolling through my mind ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On Jun 22, 2011, at 7:19 PM, Ben Taylor wrote: > I can almost see dumping 32-bit x86. > but dumping 64-bit US-III/IV? Use a kill-a-watt or a smart PDU to compare the power draw for these older systems. Do you really want them in production? Solaris 10 isn't going away if you do. q.v. several BSD & Linux flavors. > Clearly, Oracle is going to use this as a stick to force customers to > upgrade to new hardware Nobody forces me to upgrade anything -- hardware or OS. But if in the long term it will cost more in electricity alone, why would I want to hang on to older hardware? I can understand "vintage" desktop users but I don't know many folks that wish to run vintage servers unless they're tech museums. -Gary ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
You are absolutely correct. No one is going to put a 32 bit x86 system into production. And no one is going to put an old 280R or V240 UltraSparc III system into production either. But that's not the point. Jr. admins and hobbyist pick these boxes up. They come from other hobbyist. They get picked up off eBay. And they come from companies that are decommissioning them and give them to employee's, because it is easier to give them away, vs paying someone to come pick them up and haul them off. That is where new growth comes from. A jr. admin can play with stuff at home at little to no cost. And when they are comfortable with an OS at home, then they are ready and confident to take it to work. Playing/learning stuff at home is typically fine on older/slower equipment, but you want to be able to learn on the latest/greatest software. Sun learned this lesson the hard way with the secret 6 when they initially refused to release Solaris 9 on the x86 platform. http://www.computerworld.com/s/article/70339/_Secret_Six_push_Sun_to_keep_Solaris_on_Intel_?taxonomyId=068 Sun finally figured it out, and released Solaris on the x86 platform again. And did very well with Solaris 10 on x86. I don't think that the "secret six" tactics would work on Oracle. Oracle will require some other method to re-establish support for newly unsupported systems. And I have no idea what that might be. OpenIndiana may have a great opportunity to establish a foothold, if we can support, or at least state that OpenIndiana runs on those boxes that Oracle said no to. I have a SunBlade 2000 on my desktop at work running Solaris 11 express. It has S11Express loaded so I can play with ZFS encrypted file systems. If the system runs now, that means that Oracle will physically be pulling drivers, etc out of the code to ensure that my system will not run Solaris 11 GA. I believe that there is a lot of opportunity available for OpenIndiana, by just not pulling out code that is known to work. Jerry On 06/22/11 22:42, Gary Driggs wrote: > On Jun 22, 2011, at 7:19 PM, Ben Taylor wrote: > >> I can almost see dumping 32-bit x86. >> but dumping 64-bit US-III/IV? > > Use a kill-a-watt or a smart PDU to compare the power draw for these > older systems. Do you really want them in production? Solaris 10 > isn't going away if you do. q.v. several BSD & Linux flavors. ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Strategical decisions with not so large boundaries on the technical side, I mean, they are not really saving much effort! ZFS feeds on memoy like chupacabra does on blood, but still there is still some life to those old 32 bit devices on the networking side of the game. We dont trash servers so easily over here... Admins are still creative people, an operating system should be flexible enough to suport unconventional projects not all IT problems should be solved with a credit card in the hand!. Old management tricks are not apreciated, vendor lockup is not apreciated, the good old "pay or trash" pushup isnt the place where I wanna be. What eslse to say... On Thu, Jun 23, 2011 at 7:28 AM, Jerry Kemp wrote: > You are absolutely correct. No one is going to put a 32 bit x86 system > into production. And no one is going to put an old 280R or V240 > UltraSparc III system into production either. > > But that's not the point. Jr. admins and hobbyist pick these boxes up. > They come from other hobbyist. They get picked up off eBay. And they > come from companies that are decommissioning them and give them to > employee's, because it is easier to give them away, vs paying someone to > come pick them up and haul them off. > > That is where new growth comes from. A jr. admin can play with stuff at > home at little to no cost. And when they are comfortable with an OS at > home, then they are ready and confident to take it to work. > Playing/learning stuff at home is typically fine on older/slower > equipment, but you want to be able to learn on the latest/greatest > software. > > Sun learned this lesson the hard way with the secret 6 when they > initially refused to release Solaris 9 on the x86 platform. > > > http://www.computerworld.com/s/article/70339/_Secret_Six_push_Sun_to_keep_Solaris_on_Intel_?taxonomyId=068 > > Sun finally figured it out, and released Solaris on the x86 platform > again. And did very well with Solaris 10 on x86. > > I don't think that the "secret six" tactics would work on Oracle. > Oracle will require some other method to re-establish support for newly > unsupported systems. And I have no idea what that might be. > > OpenIndiana may have a great opportunity to establish a foothold, if we > can support, or at least state that OpenIndiana runs on those boxes that > Oracle said no to. I have a SunBlade 2000 on my desktop at work running > Solaris 11 express. It has S11Express loaded so I can play with ZFS > encrypted file systems. > > If the system runs now, that means that Oracle will physically be > pulling drivers, etc out of the code to ensure that my system will not > run Solaris 11 GA. > > I believe that there is a lot of opportunity available for OpenIndiana, > by just not pulling out code that is known to work. > > Jerry > > > > On 06/22/11 22:42, Gary Driggs wrote: > > On Jun 22, 2011, at 7:19 PM, Ben Taylor wrote: > > > >> I can almost see dumping 32-bit x86. > >> but dumping 64-bit US-III/IV? > > > > Use a kill-a-watt or a smart PDU to compare the power draw for these > > older systems. Do you really want them in production? Solaris 10 > > isn't going away if you do. q.v. several BSD & Linux flavors. > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On Wed, Jun 22, 2011 at 11:42 PM, Gary Driggs wrote: > On Jun 22, 2011, at 7:19 PM, Ben Taylor wrote: > >> I can almost see dumping 32-bit x86. >> but dumping 64-bit US-III/IV? > > Use a kill-a-watt or a smart PDU to compare the power draw for these older > systems. Do you really want them in production? Solaris 10 isn't going away > if you do. q.v. several BSD & Linux flavors. Lets see. I still have a 280R in production, as well as some v490s. Do I care about Power consumption in my data center? No, because rack space/circuits is what I pay for, not a per Kw charge, so I don't care. Even if the draw is high, and I want to run a Solaris on Sparc (Yeah, T2000 makes such a *good* desktop) at home, do you think I care? Wow. > >> Clearly, Oracle is going to use this as a stick to force customers to >> upgrade to new hardware > > Nobody forces me to upgrade anything -- hardware or OS. But if in the long > term it will cost more in electricity alone, why would I want to hang on to > older hardware? I can understand "vintage" desktop users but I don't know > many folks that wish to run vintage servers unless they're tech museums. you're worried about power cost? That's the last thing we're worried about. The cost of *migrating* is magnitudes larger (exponentially) than the power cost (which again, we don't concern ourselves with. Personally, I wouldn't have signed up for a Kw based pricing scheme which you apparently did) And ignoring the rest of my comment, well, the point being is that Oracle can't make someone upgrade. But the Solaris support cost went up this year for us, so we're dumping it as fast as we can. Good business strategy. Ben ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Windows made the shift last server release (2008r2 is x64 only). So it's only the OSS server families which support 32bit, likely because both BSD and Linux support lots of platforms outside of x86. Deano -Original Message- From: Gary Driggs [mailto:gdri...@gmail.com] Sent: 23 June 2011 02:10 To: Discussion list for OpenIndiana Subject: Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same? FWIW, Mac OS X Lion will only support x64 as well. IMHO, this is a good move for modern operating systems since there are always going to be alternatives for those still using i386 architecture. How long has Solaris/SPARC been 64-bit? At least ten years if not more... -Gary ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
While we are talking about 32 | 64 bit processes; Which one is better? Faster? More efficient? Mike On Thu, 2011-06-23 at 13:59 +0100, Deano wrote: > Windows made the shift last server release (2008r2 is x64 only). > > So it's only the OSS server families which support 32bit, likely because > both BSD and Linux support lots of platforms outside of x86. > > Deano > > -Original Message- > From: Gary Driggs [mailto:gdri...@gmail.com] > Sent: 23 June 2011 02:10 > To: Discussion list for OpenIndiana > Subject: Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for > solaris 11 will OI do same? > > FWIW, Mac OS X Lion will only support x64 as well. IMHO, this is a good move > for modern operating systems since there are always going to be alternatives > for those still using i386 architecture. How long has Solaris/SPARC been > 64-bit? At least ten years if not more... > > -Gary > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On Jun 23, 2011, Ben Taylor wrote: > Personally, I wouldn't have signed up for a Kw based pricing scheme which you > apparently did) I didn't as it's one of many data centers that our company built and maintains. I just prefer not to be an ass about resources. -Gary ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
The answer really is "it depends". On x86-64, the big advantage to 64-bit performance-wise is the additional CPU registers that it enables. The downside is the additional memory usage due to bigger pointer sizes. From what I understand, this increased size can have negative effects on CPU cache hit rates. -Dustin On Fri, Jun 24, 2011 at 7:02 AM, Michael Stapleton wrote: > While we are talking about 32 | 64 bit processes; > Which one is better? > Faster? > More efficient? > > Mike > > > On Thu, 2011-06-23 at 13:59 +0100, Deano wrote: > >> Windows made the shift last server release (2008r2 is x64 only). >> >> So it's only the OSS server families which support 32bit, likely because >> both BSD and Linux support lots of platforms outside of x86. >> >> Deano >> >> -Original Message- >> From: Gary Driggs [mailto:gdri...@gmail.com] >> Sent: 23 June 2011 02:10 >> To: Discussion list for OpenIndiana >> Subject: Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for >> solaris 11 will OI do same? >> >> FWIW, Mac OS X Lion will only support x64 as well. IMHO, this is a good move >> for modern operating systems since there are always going to be alternatives >> for those still using i386 architecture. How long has Solaris/SPARC been >> 64-bit? At least ten years if not more... >> >> -Gary >> ___ >> OpenIndiana-discuss mailing list >> OpenIndiana-discuss@openindiana.org >> http://openindiana.org/mailman/listinfo/openindiana-discuss >> >> >> ___ >> OpenIndiana-discuss mailing list >> OpenIndiana-discuss@openindiana.org >> http://openindiana.org/mailman/listinfo/openindiana-discuss > > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
For Intel CPUs, 32 bit code is certainly more compact , and in some cases arguably faster than 64 bit code. (say, comparing the same code on the same machine compiled 32 and 64 bit) But, newer cpu silicon tends to make performance improvements in many ways (e.g locating more supporting circuity on the cpu's silicon, increasing L1 /L2 cache sizes, etc) Newer CPUs also tend to be more energy efficient. Intel made great strides towards energy efficiency. E.g.: idling the cpu when not in use ( deep C states etc. of gating off any circuitry that is not in use, modulating the cpu clock rate ( SpeedStep). So performance and energy efficiency is more dependent on which generation of cpu core design we have, rather than on just the the bitness . The primary advantage of "64 bit" per se ( ie running a given cpu in 64 bit mode) is the increased addressable memory space. The current hardware limit set by the manufacturers is at 48 address bits (256 terabytes theoretical limit) Actual OS support cuts this in half, or less. Motherboard limitations further curtail this, but 48G motherboards are now commonplace. On 32 bit Intel (Amd) you are typically limited to 4G, which is split between kernel and userland depending on the OS and configuration. (E.g.: 1G kernel and 3G userland) Steve - "Michael Stapleton" wrote: While we are talking about 32 | 64 bit processes; Which one is better? Faster? More efficient? Mike ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
The main difference is (in)ability to use large amount of RAM. 32-bit systems use 32-bit pointers, and maximum value of unsigned 32-bit integer is 2^32, or 4Gb. In practice many 32-bit OSes are able to address only 2Gb of RAM. 64-bit pointers can address much more RAM (no hardware has reached 2^64 RAM limit yet). Nowadays even laptops may have 4 or 8 Gb of RAM, not to mention servers, which should have much more. So, 32-bit OS is a past day OS, for legacy hardware only. Unfortunately not everyone can afford a new hardware. My web server has 1 GB RAM and 32-bit processor :( Happily running OSol b134 :) Regards, Dmitry. While we are talking about 32 | 64 bit processes; Which one is better? Faster? More efficient? ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
So I guess it would be fair to say that the best OS is the one that support both at the same time, and leaves the option to the developer for each individual application. My understanding is that Solaris is more like 4G per process/kernel, rather than 4GB total. Multiple 32 bit processes could use more than 4GB total; just not individually. Mike On Fri, 2011-06-24 at 15:58 +, Steve Gonczi wrote: > For Intel CPUs, 32 bit code is certainly more compact , and in some cases > arguably faster than 64 bit code. (say, comparing the same code on the same > machine > compiled 32 and 64 bit) > > But, newer cpu silicon tends to make performance improvements > in many ways (e.g locating more supporting circuity on the cpu's silicon, > increasing L1 /L2 > cache sizes, etc) > > Newer CPUs also tend to be more energy efficient. > Intel made great strides towards energy efficiency. > E.g.: idling the cpu when not in use ( deep C states etc. > of gating off any circuitry that is not in use, modulating the cpu clock rate > ( SpeedStep). > > So performance and energy efficiency is more dependent on > which generation of cpu core design we have, rather than on > just the the bitness . > > > The primary advantage of "64 bit" per se ( ie running a given cpu in 64 bit > mode) > is the increased addressable memory space. > The current hardware limit set by the manufacturers is at 48 address bits > (256 terabytes theoretical limit) Actual OS support cuts this in half, or > less. > Motherboard limitations further curtail this, but 48G motherboards are now > commonplace. > > On 32 bit Intel (Amd) you are typically limited to 4G, which is split between > kernel and userland > depending on the OS and configuration. (E.g.: 1G kernel and 3G userland) > > Steve > > - "Michael Stapleton" wrote: > > > While we are talking about 32 | 64 bit processes; > Which one is better? > Faster? > More efficient? > > Mike > > > > > > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
I really can not make a case for 32 bit except for a legacy binary where you do Not have a choice Do we need a 32 bit kernel ? Probably not. Do we need the ability To run a 32 bit binary?I think so -:::-sG-:::- On Jun 24, 2011, at 12:17, Michael Stapleton wrote: > So I guess it would be fair to say that the best OS is the one that > support both at the same time, and leaves the option to the developer > for each individual application. > > My understanding is that Solaris is more like 4G per process/kernel, > rather than 4GB total. > Multiple 32 bit processes could use more than 4GB total; just not > individually. > > Mike > > > On Fri, 2011-06-24 at 15:58 +, Steve Gonczi wrote: > >> For Intel CPUs, 32 bit code is certainly more compact , and in some cases >> arguably faster than 64 bit code. (say, comparing the same code on the same >> machine >> compiled 32 and 64 bit) >> >> But, newer cpu silicon tends to make performance improvements >> in many ways (e.g locating more supporting circuity on the cpu's silicon, >> increasing L1 /L2 >> cache sizes, etc) >> >> Newer CPUs also tend to be more energy efficient. >> Intel made great strides towards energy efficiency. >> E.g.: idling the cpu when not in use ( deep C states etc. >> of gating off any circuitry that is not in use, modulating the cpu clock >> rate >> ( SpeedStep). >> >> So performance and energy efficiency is more dependent on >> which generation of cpu core design we have, rather than on >> just the the bitness . >> >> >> The primary advantage of "64 bit" per se ( ie running a given cpu in 64 bit >> mode) >> is the increased addressable memory space. >> The current hardware limit set by the manufacturers is at 48 address bits >> (256 terabytes theoretical limit) Actual OS support cuts this in half, or >> less. >> Motherboard limitations further curtail this, but 48G motherboards are now >> commonplace. >> >> On 32 bit Intel (Amd) you are typically limited to 4G, which is split >> between kernel and userland >> depending on the OS and configuration. (E.g.: 1G kernel and 3G userland) >> >> Steve >> >> - "Michael Stapleton" wrote: >> >> >> While we are talking about 32 | 64 bit processes; >> Which one is better? >> Faster? >> More efficient? >> >> Mike >> >> >> >> >> >> >> ___ >> OpenIndiana-discuss mailing list >> OpenIndiana-discuss@openindiana.org >> http://openindiana.org/mailman/listinfo/openindiana-discuss >> >> >> ___ >> OpenIndiana-discuss mailing list >> OpenIndiana-discuss@openindiana.org >> http://openindiana.org/mailman/listinfo/openindiana-discuss > > > ___ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Steve Gonczi wrote: I really can not make a case for 32 bit except for a legacy binary where you do Not have a choice Do we need a 32 bit kernel ? Probably not. Do we need the ability To run a 32 bit binary?I think so Well, can't speak to opensolaris, but for 64-bit linux and windows, you can run 32-bit executables just fine, so that argument is not compelling, IMO... ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On 06/24/11 08:58 AM, Steve Gonczi wrote: > For Intel CPUs, 32 bit code is certainly more compact , and in some cases > arguably faster than 64 bit code. (say, comparing the same code on the same > machine > compiled 32 and 64 bit) But 64-bit code has access to more registers, which can make a measurable difference for some code. SPARC is much more like you say, since the registers and ISA's are basically the same between the two modes, leaving it down to things like accessing a greater range of memory vs. doubling the memory used by every pointer address & long int. -- -Alan Coopersmith-alan.coopersm...@oracle.com Oracle Solaris Platform Engineering: X Window System ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
> So I guess it would be fair to say that the best OS is the one that > support both at the same time Yes, and that's OSol and OI do. On 24.06.2011 22:17, Michael Stapleton wrote: So I guess it would be fair to say that the best OS is the one that support both at the same time, and leaves the option to the developer for each individual application. My understanding is that Solaris is more like 4G per process/kernel, rather than 4GB total. Multiple 32 bit processes could use more than 4GB total; just not individually. Mike On Fri, 2011-06-24 at 15:58 +, Steve Gonczi wrote: For Intel CPUs, 32 bit code is certainly more compact , and in some cases arguably faster than 64 bit code. (say, comparing the same code on the same machine compiled 32 and 64 bit) But, newer cpu silicon tends to make performance improvements in many ways (e.g locating more supporting circuity on the cpu's silicon, increasing L1 /L2 cache sizes, etc) Newer CPUs also tend to be more energy efficient. Intel made great strides towards energy efficiency. E.g.: idling the cpu when not in use ( deep C states etc. of gating off any circuitry that is not in use, modulating the cpu clock rate ( SpeedStep). So performance and energy efficiency is more dependent on which generation of cpu core design we have, rather than on just the the bitness . The primary advantage of "64 bit" per se ( ie running a given cpu in 64 bit mode) is the increased addressable memory space. The current hardware limit set by the manufacturers is at 48 address bits (256 terabytes theoretical limit) Actual OS support cuts this in half, or less. Motherboard limitations further curtail this, but 48G motherboards are now commonplace. On 32 bit Intel (Amd) you are typically limited to 4G, which is split between kernel and userland depending on the OS and configuration. (E.g.: 1G kernel and 3G userland) Steve - "Michael Stapleton" wrote: While we are talking about 32 | 64 bit processes; Which one is better? Faster? More efficient? Mike ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
On Jun 24, 2011, at 4:17 PM, Dmitry Kozhinov wrote: > > So I guess it would be fair to say that the best OS is the one that > > support both at the same time > > Yes, and that's OSol and OI do. Processes running under 32-bit Solaris can only be 32-bit. Processes running under 64-bit Solaris can be either. However, drivers, etc must match the kernel. And for all practical purposes, debugging a process (or the kernel) is a whole lot easier when done by a process with the same bitness as what it's debugging. For instance, it may not be feasible for a 32-bit process to control a 64-bit process (although the reverse may be possible, if more difficult than if they were the same). So the loss of a 32-bit kernel option limits mainly the hardware you can run on (CPU, but maybe also old drivers that were never ported to 64-bit). Other OS's may have somewhat different rules. OS X for instance can run 64-bit processes on a 32-bit kernel given 64-bit capable hardware. I expect that very few OS's would go to the trouble of supporting 32-bit drivers on a 64-bit OS, for example; not perhaps 100% impossible depending on the driver-to-OS interface, but certainly extra complexity where it's most dangerous. ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] oracle removes 32bit x86 cpu support for solaris 11 will OI do same?
Michael Stapleton wrote: While we are talking about 32 | 64 bit processes; Which one is better? Faster? More efficient? Initially, assuming a 32 verses 64 bit build doesn't change any algorithms... On x86, a 64 bit build of the same program will typically run ~50% faster if it's CPU-bound, because more registers are available for the compiler/optimizer to use. There's a wide variance depending what the program does (I have an example which gets much better than 50% gain). If it's not CPU-bound (and most things aren't), it makes no difference. However, if the larger pointers and data items push the 64 bit program's working set size over what fits in the CPU cache whereas the 32 bit version does fit in the cache, then you can in theory see the 32 bit version winning. On sparc, a 64 bit build of the same program does not benefit from any more registers like on x86, but it does pay the price for a larger working set size, and I typically see a 10-14% performance reduction for a CPU-bound program which has been just rebuilt 64bit. However, if you can use the 64 bit address space to change the algorithms used by your app, such as mmaping files rather than doing loads of lseek/read/write ops, then you may see additional gains from this, and on sparc that will often more than cancel out the reduction in CPU performance by some way. I wouldn't personally bother changing anything much which is shipped with the OS (very rarely is the performance of things in /usr/bin an issue). However, I would suggest taking these factors into account when building the key applications your system is going to run, if you are CPU-bound. -- Andrew ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss