[cctalk] Re: Intel's i860, Cray-On-A-Chip
On 9/28/2022 20:18, Paul Koning via cctalk wrote: FWIW, I just tried building GCC 2.95.2 on my Linux system (Fedora Core 32, GCC 10.3.1. It almost built, ran into an argument mismatch error message in something called "chill". So if you want something that old it looks like you'll have to start by building a less ancient version, say 4.8 or so, and then use that to build the dinosaur. Which switches/options did you use? Last weekend I spent some cycles to try ... I can build the binutils around 2.32, emitting elf No luck at all, getting gcc to compile. Support for the i860 was removed very early after 4.0, going back to 3.x versions fails as they can't compile on 64 bit systems. (and also never used ELF, but COFF?) So stuck a little at the moment, I guess I have to install some linux 32 bit version, and try again ...
[cctalk] Re: Intel's i860, Cray-On-A-Chip
> On Oct 13, 2022, at 11:02 AM, emanuel stiebler wrote: > > On 9/28/2022 20:18, Paul Koning via cctalk wrote: >> FWIW, I just tried building GCC 2.95.2 on my Linux system (Fedora Core 32, >> GCC 10.3.1. It almost built, ran into an argument mismatch error message in >> something called "chill". So if you want something that old it looks like >> you'll have to start by building a less ancient version, say 4.8 or so, and >> then use that to build the dinosaur. > > Which switches/options did you use? > Last weekend I spent some cycles to try ... I cheated a bit. My test was "can GCC V.old build at all". So I tried a native build, not a cross-build. Cross-builds have their own set of issues and it's been long enough that I no longer remember them well. That's what I would suggest when dealing with old versions: first try a native (or i386 if you have x86_64) build to see what, if any, issues you need to handle just for the generic old GCC. Given that you have that working, you can then do the cross-build. For example, if you need a GCC V.medium to build the GCC V.ancient, discovering that first makes things easier. Yes, you'd have to start by finding an old enough binutils that supports the target platform. With that in place, you can then build gcc for that platform. paul > I can build the binutils around 2.32, emitting elf > No luck at all, getting gcc to compile. Support for the i860 was removed very > early after 4.0, going back to 3.x versions fails as they can't compile on 64 > bit systems. (and also never used ELF, but COFF?) > > So stuck a little at the moment, I guess I have to install some linux 32 bit > version, and try again ... >
[cctalk] Re: Intel's i860, Cray-On-A-Chip
> On Oct 6, 2022, at 12:51 AM, Rodney Brown via cctalk > wrote: > > According to gcc-9.1.0/NEWS Intel i860 was an architecture declared obsolete > in GCC 4.0 (and previously in GCC 3.1). GCC does this in two steps. One is that it's marked as obsolete but can still be built, you just get a warning. The second is that the bits are actually removed. If you want to run something old, it's probably still good until that final removal. But given the lack of maintenance that triggers removal, it is certainly possible that there are some "bit rot" bugs before that point, possibly even before the first warning stage. paul
[cctalk] Re: Intel's i860, Cray-On-A-Chip
According to gcc-9.1.0/NEWS Intel i860 was an architecture declared obsolete in GCC 4.0 (and previously in GCC 3.1). So you could check the configuration files are there in 3.4.6 (2006-03-10) or in 4.0.4 (2007-01-31). The binutils/gas/i860 configuration files were removed in 2018-04-11, so binutils 2.30 (2018-01-28) may still be able to build cross binutils assemblers and linkers for the architecture. I don't have an unpacked gdb lying around to check. The announce mailing list archives https://sourceware.org/pipermail/gdb-announce/ may be worth checking if the config isn't in gdb.
[cctalk] Re: Intel's i860, Cray-On-A-Chip
> On Sep 27, 2022, at 7:43 PM, Paul Koning via cctalk > wrote: > > > >> On Sep 27, 2022, at 5:28 PM, Stefan Skoglund via cctalk >> wrote: >> >> fre 2022-09-23 klockan 10:30 -0400 skrev emanuel stiebler via cctalk: >>> Hi all, >>> anybody has some GCC or any other tool chain for the above? >>> Or some pointers, which was the last version of the GCC tool chain >>> which >>> supported the i860, and would be still compile-able on this days >>> tools/OS's? >>> >>> Anything? >>> >>> Thanks in advance! >> >> pick together something like a sunos 5 or sunos4 machine and build the >> tool-chain on that ? > > I don't know how old a GCC you can build on current machines. The other way > around, yes. But it's hard to see why old on new would be a problem. You > can grab any GCC version you want and try it yourself. FWIW, I just tried building GCC 2.95.2 on my Linux system (Fedora Core 32, GCC 10.3.1. It almost built, ran into an argument mismatch error message in something called "chill". So if you want something that old it looks like you'll have to start by building a less ancient version, say 4.8 or so, and then use that to build the dinosaur. paul
[cctalk] Re: Intel's i860, Cray-On-A-Chip
> On Sep 27, 2022, at 5:28 PM, Stefan Skoglund via cctalk > wrote: > > fre 2022-09-23 klockan 10:30 -0400 skrev emanuel stiebler via cctalk: >> Hi all, >> anybody has some GCC or any other tool chain for the above? >> Or some pointers, which was the last version of the GCC tool chain >> which >> supported the i860, and would be still compile-able on this days >> tools/OS's? >> >> Anything? >> >> Thanks in advance! > > pick together something like a sunos 5 or sunos4 machine and build the > tool-chain on that ? I don't know how old a GCC you can build on current machines. The other way around, yes. But it's hard to see why old on new would be a problem. You can grab any GCC version you want and try it yourself. As for when i860 disappeared, no idea. The release notes would say, so you could search those. Or the ChangeLogs. A quick scan says that i860 support was removed in 2002, put back in 2003. The newest reference is in 2005. So if you grab a 2004 release of GCC it's likely to have the support. Similarly, FWIW, I see references to i960 in the 2003 ChangeLog. paul
[cctalk] Re: Intel's i860, Cray-On-A-Chip
fre 2022-09-23 klockan 10:30 -0400 skrev emanuel stiebler via cctalk: > Hi all, > anybody has some GCC or any other tool chain for the above? > Or some pointers, which was the last version of the GCC tool chain > which > supported the i860, and would be still compile-able on this days > tools/OS's? > > Anything? > > Thanks in advance! pick together something like a sunos 5 or sunos4 machine and build the tool-chain on that ? i believe it is enough building binutils (which has gas) and gcc late 90s version. The trouble is mainly finding a compiled compiler for sunos5. Sun's compiler in sunos4 is part of the OS, not unbundled. NetBSD or FreeBSD is another possibility.
[cctalk] Re: Intel's i860, Cray-On-A-Chip
> I think there was a unix/unix-like OS for them, but I imagine context > switching > was slow... There were a couple *nix workstations based on it. The Oki 7300 series comes to mind. I think someone exhibited at that VCF pre-COVID. -- personal: http://www.cameronkaiser.com/ -- Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com -- Everything is permissible, but not everything is expedient. -- 1 Cor 6:12 --
[cctalk] Re: Intel's i860, Cray-On-A-Chip
Hi Emanuel, On 9/23/22 16:30, emanuel stiebler via cctalk wrote: Hi all, anybody has some GCC or any other tool chain for the above? Or some pointers, which was the last version of the GCC tool chain which supported the i860, and would be still compile-able on this days tools/OS's? I've got a PC with an Hauppauge 4860 motherboard. There's a 80486 and a 80860 CPU on the board. I have some things online for this machine on ftp://ftp.groessler.org/pub/chris/i860. I've put it online there long ago. I think the "i860tools-linux.tar.bz2" file could contain a gcc version for i860. But I found no source code for the compiler/toolchain when looking at the contents of this file. I had a gcc version in source code (work-in-progress), created by Jason Eckhardt, at some point in time (in the later 200X time frame). I could try to dig it out. I've also got a DOS version of (I think) the Portland Group C compiler for the i860. regards, chris
[cctalk] Re: Intel's i860, Cray-On-A-Chip
On Fri, 23 Sept 2022 at 23:57, Chuck Guzis via cctalk wrote: > > I believe (I'll have to check) that in the Osborne-McGraw-Hill/Intel > i860 book there's a quote from BillG saying that Microsoft was committed > to developing for the 860 as a personal computer CPU. > > I think that never happened...it would have been interesting, however. The closest I know of is that the early (late-1980s) versions of what was then OS/2 NT were built on i860 boards, codenamed "Razzle". The Smithsonian has one: https://www.si.edu/object/microsoft-windows-nt-development-board-pcr1-rev1-intel-i860-processor%3Anmah_742558 There are a handful of mentions of them here and there: https://www.betaarchive.com/forum/viewtopic.php?t=32510 https://devblogs.microsoft.com/oldnewthing/20181224-00/?p=100545 The codename of the first version of the i860 was "N-Ten" which is where the "NT" product name originated: https://web.archive.org/web/20110720042038/http://www.winsupersite.com/article/windows-server/windows-server-2003-the-road-to-gold-part-one-the-early-years-127432 -- Liam Proven ~ Profile: https://about.me/liamproven Email: lpro...@cix.co.uk ~ gMail/gTalk/FB: lpro...@gmail.com Twitter/LinkedIn: lproven ~ Skype: liamproven UK: (+44) 7939-087884 ~ Czech [+ WhatsApp/Telegram/Signal]: (+420) 702-829-053
[cctalk] Re: Intel's i860, Cray-On-A-Chip
> > IIRC the Intel IPSC (Inter Personal Super Computer) put a ton of these in parallel. It ran some kind of Unix and there has to have been a gcc port. https://www.vaxbarn.com/42-repair/756-ipsc-860-repair > > I believe (I'll have to check) that in the Osborne-McGraw-Hill/Intel > i860 book there's a quote from BillG saying that Microsoft was committed > to developing for the 860 as a personal computer CPU. > > I think that never happened...it would have been interesting, however. > > -- -Jon +44 7792 149029
[cctalk] Re: Intel's i860, Cray-On-A-Chip
On 9/23/22 11:12, Gordon Henderson via cctalk wrote: > On Fri, 23 Sep 2022, emanuel stiebler via cctalk wrote: > >> Hi all, >> anybody has some GCC or any other tool chain for the above? >> Or some pointers, which was the last version of the GCC tool chain >> which supported the i860, and would be still compile-able on this days >> tools/OS's? >> >> Anything? > > I can't help with your quest, however I wish you both luck and the > energy to preserve your old grey cells when dealing with these beasts. > > Back in the day I did some work on the i860. Fascinating chip. Very > capable of doing high speed vector math and multiply + add in a single > cycle. I believe (I'll have to check) that in the Osborne-McGraw-Hill/Intel i860 book there's a quote from BillG saying that Microsoft was committed to developing for the 860 as a personal computer CPU. I think that never happened...it would have been interesting, however. --Chuck
[cctalk] Re: Intel's i860, Cray-On-A-Chip
On Fri, 23 Sep 2022, emanuel stiebler via cctalk wrote: Hi all, anybody has some GCC or any other tool chain for the above? Or some pointers, which was the last version of the GCC tool chain which supported the i860, and would be still compile-able on this days tools/OS's? Anything? I can't help with your quest, however I wish you both luck and the energy to preserve your old grey cells when dealing with these beasts. Back in the day I did some work on the i860. Fascinating chip. Very capable of doing high speed vector math and multiply + add in a single cycle. As long as you were prepared to sacrifice the old grey cells to it's dual-instruction mode programming. I didn't do anything in C that I recall, all assembler (I was involved mostly in test/diagnostics and low-level accesses - we had them on a shared memory system with Transputers being relegated (by then) to nothing more than "smart" comms chips... (Late 80s) I think we used the Portland Group C and FORTRAN compilers, but even then we had large librarys hand-coded in assembler to achieve that fabled 3 instructions per cycle quote - you have to effectively pump the floating point pipeline by hand so you ran operations that assembled to a 64-bit word which was one 32-bit instructions for the integer unit (say a load from RAM into FPU pipeline) and one for the FPU (say, a multiply and add instruction). So some 8 cycles later you'd actually start to get results out as it took 8 cycles to actually do the multiply and add - however it was one cycle after that - good for vector work, but sub-optimal for a single number multiply. And lets hope you never, ever, have to take an interrupt during an FP operation. It's 100's of cycles to save/restore the pipeline. I think there was a unix/unix-like OS for them, but I imagine context switching was slow... Cheers, Gordon
