I was briefly a sysadmin for a fleet of these things. I got to use them in my government internship for a while. Seems incredible how well they aged, and concepts like the x terminal are so underutilized today. And yet... Here we are in exaflop country. Seems like a dream.
My dad was a computer programmer for his entire career, and he developed on Sun, SGI, HP, IBM Unix machines. I remember going into his office in the early 90s (I would have been about 10 years old) and being absolutely amazed by these computers with their large monitors and network logins. He always told me that PCs and macs were toys, and Unix was incredibly powerful. I never got to see them boot because they rarely crashed. Would love to hear stories about working for Sun. They were a remarkable company.
Unfortunately, I worked for SUN during their twilight, about 2 years before the Oracle acquisition. I was an open-source advocate in their sales organization.
PCs were "toys" until Linux/Windows NT and the Pentium changed things. Linux and NT were the first "serious" options for PCs, yeah UNIX was around for the 386/486, but it was expensive and not for general use. The Pentium (and AMD's Pentium-class chips) were the first CPUs to offer workstation-level performance on x86, and they did so by internally translating x86 to something that could run on a more RISC-like architecture. Then you started to get PC graphics cards that could rival SGI's offerings, and was all over for the UNIX workstation makers, a PC could do the same thing on affordable hardware with a totally free (if you ran Linux) or fully Windows-compatible OS. As for Macs... well, they were also toys when compared to UNIX workstations... until Apple started building workstation-class hardware with PowerPC RISC chips, then made them actual UNIX workstations with OS X! Fruity colorful workstations (just like SGI) but Apple was a bit ahead of the curve on that one.
If you haven’t seen Serial Experiments Lain, I bet you’d get a kick out of the sequence in the first episode of her dad booting his monster rig with, like, 6 or 7 CRTs. It’s traditional animation, so the backlit cels look, just, (chefs kiss) for the glow of the lights and monitors. The sound design is great too. (Heck, for all I know that sequence might have its own video on RUclips!)
@@coyote_den said "and NT were the first "serious" options for PCs" A serious option that installs Solitaire BY DEFAULT is NOT a serious option. It's just you're used to scammers
We used SparcStation 10s for workgroup servers. Drop duo or quad Ross CPU modules in them, load them up with RAM, and hook them up to a SCSI cabinet for some storage and they worked very well.
Your host ID should be a combination of the "machine type" (in your case 72 for an SS20) and the last 3 octets of your ethernet address. So in your case the host ID is 7220539f. Really hope to see this machine booting! I've got a soft spot for non-Intel based hardware (2 Sun workstations, 2 Sun servers, Amiga 500, and Atari 520ST).
It's a bit of a nod to Amiga development. Early AmigaOS development was done on SUN workstations (68K-based SUN 3 machines), and there are even references to SunOS commands in the early AmigaOS 1.x developer documentation.
In retrospect, they were probably doing earliest development on SUN 2 machines, since the SUN 3 line was introduced in 1985...same year as the original Amiga.
I believe that’s just the PID not the system ID. Also chuckling at how many folk don’t know how tty’s work, or that delete and backspace aren’t the same. Grin! Did a lot of work on Sun boxes back in the day…
Well ahead of its time. If linux existed back then, the Amiga would surely have been linux based. As it is, it was a fully preemptive multitasking OS with many similarities to Unixes of the day, except designed for single-user. I loved being able to swipe the screen down and reveal my Macintosh VM running at full speed in the background whilst I was playing Lemmings.
I have an SGI Indy and it has exactly the same issue (battery backed ram storing the mac address and clock) It died a long time ago and now every time I use it I have to manually enter the mac address upon bootup. On the Indy the address is printed in plain text on a sticker on the back so wasn't as tricky as this, but still. Some day I might get around to replacing the battery in the Dallas chip.... Some day ;) Great video btw. /Magnus
Thanks! I did not know that about SGI machines....so SUN didn't do something completely unique with the MAC Addresses! I had an SGI Indy for a short time, but gave it away....kind of regret that now.
This was interesting. I like seeing systems based on alternative architectures to x86, especially those that are no longer in production. It also reminds me why I don't like using modern Linux that much... it's reliance on cryptic, arcane commands in the terminal and scripts, all inherited from UNIX. It was bad enough learning how to comfigure MSDOS back in the day, and Windows 3.1 etc. 😂
SPARC processors, or rather their 64-bit descendants the UltraSPARC line, are still in production, albeit only in server form. Both Oracle and Fujitsu produce UltraSPARC servers and Oracle is still updating Solaris. How long this will last is anyone's guess, though....they're generally considered "Legacy systems" by most customers, and I suspect neither Oracle or Fujitsu have grown their SPARC business much in the last decade.
@@TheStefanskoglund1for my common tasks, and htop, I prefer it too. You can have the task done in the time it takes a GUI application to open, even on modern SSDs (but it was far more of a time saver in 2010). Though for having to look-up lesser-used commands, often the time taken is similar to the overhead from a GUI utility. Though even then, the added flexibility to 7z and ffmpeg are worth it for me versus GUI 7zip and Handbrake.
I had an old 1995 IPC 25 MHz with 48 Mo ram, external drive, external cdrom, external tape, flat screen and optic mouse. Loved this thing. I used it running NetBSD and WindowMaker. I've always consider Solaris being a pain, while AIX or Hp-UX were easy to use. Let's not talk about Linux or any BSD ... Or maybe I was not enough experimented with Solaris :) We also had a few modern Sun Blade, I was not impressed. Even if they used Gnome instead of CDE :) Damned those things were slow and boring .... I gave the IPC to a friend when I moved from France to Canada, with the promise never to sell it and take care :) No idea what he did with it later ...
I remember the old IPC and IPX "lunchbox" machines. I always wanted one, but never had one. I think a lot of people kept those running SunOS 4 (i.e. the BSD-based predecessor to Solaris) rather than upgrade to Solaris. SUN kept supporting SunOS 4 until 2003! I ran NetBSD or Debian on my older SUN machines, because Solaris was definitely slow....my main desktop for a while was an Ultra 30 and it could run Solaris 9 reasonably well...the Ultra 1 it replaced, not so much. I remember Solaris feeling a lot more complicated that NetBSD and Linux back in the day. Although, Solaris 10 was definitely ahead of it's time with smf (like systemd in modern Linux), zones (a precursor to containers), and zfs (now ported to Linux).
Thanks for the fantastic tips. I have my SS10 fully working but not really a Unix person so went through the Dallas battery ram replacement and config rebuilt using the same proceedures as most others I suppose same MAC address! I am suprised that people hack to bits the Dallas chips as I just bought one (about 5 years ago) from a reputable electronics supply retailer. Wish I had done more of the install with the serial term as cutting and pasting is of course so much easier and the web installer slows matters down.
Dremeling the Dallas chips (or in this case an SG Thompson Timekeeper) is not worth the effort, IMO. Identical and/or compatible replacements are still available at $25-$30. Not cheap, but not super expensive either. The only downside is when they go out of stock. It looks like I will have to wait until Nov/Dec to procure a replacement according to Mouser's restock estimate. Apparently, the original Ultra (the 1, 2, 5, 10, 30, 60 and 80) series' clock battery chip _is_ discontinued with no suitable replacement. I'm hopeful that one day, someone will create a full plug-in, no hack replacement using a lithium cell.
What an odd design of a PC; Reminds me of some early server gear. Pretty cool using the old Amiga to give it some love; I believe you can run modern Linux on that sucker.
The Linux kernel team has been stripping out support for old architectures. The last Debian release to support 32 bit SPARC machines was about 9 years ago. See: www.phoronix.com/news/Linux-Drop-Most-SPARC32-v2 NetBSD support is still pretty strong, IIRC.
@@ClarkHaleI had a long-running occasional-work attempt to get Debian to behave on 32-bit PPC, but even when technically supported it was as a second or third class arch with all sorts of bugs. So years ago, I came to the conclusion that the BSD options looked more promising. Even though it involves more compiling, but I’m not afraid of that. Somewhat validating to see you say it’s similar for SPARC.
very very very interesting - I have a sparc station 20 too, with pretty much the same issues. Although, I want to convert it into a NeXT OpenStep 4.2 Machine. I would ike to see how the story continues. When will you release the next episode?
Think the MAC address is on a sticker both on the rear AND inside ! Sun intentionally chose to use the host MAC address on ALL ethernet if in a machine - they observed that in the domain there a MAC address is used ie a network segment, the same address wont be used by more than one machine. I wonder, believe IBM RS/6000 does the same thing. It is the PC (and Mac) side which went with card manufacturer specific and provided addresses.
No, UEFI does not have Forth as part of the standard. You may be confusing UEFI with OpenFirmware, which was used by SUN, IBM (for their POWER-based systems), Apple (on some/all PowerPC-based Macintoshes), and some others. OpenFirmware is based around Forth. I don't believe this SS20 is technically an OpenFirmware machine, but the standard was largely inspired from SUN's firmware, so it's very close. UEFI was originally created for Itanium systems, and then later used for x86/x86-64 and ARM. UEFI was developed after OpenFirmware, and Intel had several beefs with how OpenFirmware was designed, including the use of Forth and the device tree. There's some discussion of this on a HackerNews thread which also has a link to an Intel whitepaper: news.ycombinator.com/item?id=15730382
@@ClarkHale Ooh, sorry. Looking at the current standard, I must have confused EBC (EFI Byte Code) with a FORTH. During the Itanium bring-up, there was a lot of interest from intel to move away from ISA-specific option ROMs and toward something portable, so they baked a preference for portable drivers into EFI.
I had a pizza box in the late aughts that I probably got for next to nothing. I then gave it away as part of a giant cleanout, kind of regret that now. Ordered a Blade 100 the other day, going to give getting Solaris 9 running a go on it...
With regards to MAC addresses bound to a system not the card, RARP uses that, basically for each IP net (before using multiple ip addresses in one ethernet network) you configured a RARP server to tell hosts which want's to know their assigned IP address while only knowing their MAC address the IP address they should use. The IP addresses is in environment adminstratively assigned to ONE machine. The client could then use their assigned IP address to broadcast for a bootp server. This design was common in diskless workstation environments and for X terminals. RARP and BOOTP is partly replaced with DHCP (though at least the DHCP servers in Debian is BRAIN-DEAD, they basically can only be depended on to do IP address assignment and giving out DNS information ie domain and server addresses, NOT doing what BOOTP was able to do.)
This is what I did for most of my SPARC machines. Solaris had a system called Jumpstart that allowed you to define the configuration for a machine and then install that machine over the network. It relied on RARP+BOOTP and was much faster than trying to install over a 3x or 6x CD-ROM drive!
@@ClarkHale and the design was easily able to support 200-1000-2000 machines from ONE jump-start server. Big-patch day ? Force all machines to reinstall from the jumpstart machine. 4 hours later - done.
SUN also produced x86 in the very late 80's machines, although they weren't fully IBM PC-compatible, and were a market failure. Later, in the 2000's, they produced true standard x86 workstations and servers based around AMD processors. I used to have an Ultra 20 that had a AMD Opteron processor. SUN also ported Solaris to x86 systems and fully open sourced it as OpenSolaris. Unfortunately, when Oracle bought out SUN, they eventually closed source Solaris again. However, a fork of OpenSolaris lives on the OpenIndiana project.
@@ClarkHale they did two generations of x86 machines - the Sun i386 machines of mid 80s (which was killed a few years post the SPARC launch) and the AMD thumper machines in different versions.
Sun did the first PCI66 implementation (PCI 33 /66 MHz and 32 or 64 bit wide) for the Ultra 30 (Sbus har lower latency than 32 bit PCI so is mostly faster over all even with it's slower clock so Sun to get a marketable improvement in performance compared with Sbus had to design a working 64 bit PCI implementation well before the PC crowd....)
I was briefly a sysadmin for a fleet of these things. I got to use them in my government internship for a while. Seems incredible how well they aged, and concepts like the x terminal are so underutilized today. And yet... Here we are in exaflop country. Seems like a dream.
Man! That stty erase tip is something I've been looking for eons! Thanks 🙏😅
Typing "halt" without first typing "sync" three times would have resulted in a slap on the head from my boss back in the day. :)
I had a professor in college who had the same habit. He said it was from his PDP-11 days :D
@@ClarkHale You realize that now, in a matter of thinking, we're them. :)
"sync:sync:sync:halt"
I recall using this machine back in the day. Dual CPU 50 mhz. Nice.
My dad was a computer programmer for his entire career, and he developed on Sun, SGI, HP, IBM Unix machines. I remember going into his office in the early 90s (I would have been about 10 years old) and being absolutely amazed by these computers with their large monitors and network logins. He always told me that PCs and macs were toys, and Unix was incredibly powerful. I never got to see them boot because they rarely crashed. Would love to hear stories about working for Sun. They were a remarkable company.
Unfortunately, I worked for SUN during their twilight, about 2 years before the Oracle acquisition. I was an open-source advocate in their sales organization.
PCs were "toys" until Linux/Windows NT and the Pentium changed things.
Linux and NT were the first "serious" options for PCs, yeah UNIX was around for the 386/486, but it was expensive and not for general use. The Pentium (and AMD's Pentium-class chips) were the first CPUs to offer workstation-level performance on x86, and they did so by internally translating x86 to something that could run on a more RISC-like architecture. Then you started to get PC graphics cards that could rival SGI's offerings, and was all over for the UNIX workstation makers, a PC could do the same thing on affordable hardware with a totally free (if you ran Linux) or fully Windows-compatible OS.
As for Macs... well, they were also toys when compared to UNIX workstations... until Apple started building workstation-class hardware with PowerPC RISC chips, then made them actual UNIX workstations with OS X! Fruity colorful workstations (just like SGI) but Apple was a bit ahead of the curve on that one.
If you haven’t seen Serial Experiments Lain, I bet you’d get a kick out of the sequence in the first episode of her dad booting his monster rig with, like, 6 or 7 CRTs. It’s traditional animation, so the backlit cels look, just, (chefs kiss) for the glow of the lights and monitors. The sound design is great too.
(Heck, for all I know that sequence might have its own video on RUclips!)
@@coyote_den said "and NT were the first "serious" options for PCs"
A serious option that installs Solitaire BY DEFAULT is NOT a serious option. It's just you're used to scammers
@@alastorgdla little solitaire never hurt anyone
Oh. Now I got a craving to go down in my basement and get my Sparcstation 5. Haven’t started it in like 15 years 😀
We used SparcStation 10s for workgroup servers. Drop duo or quad Ross CPU modules in them, load them up with RAM, and hook them up to a SCSI cabinet for some storage and they worked very well.
Thank you, I have this exact problem/project on my backlog, so I'm bookmarking to watch in the entirety later.
Woohoo glad I stumbled on your channel I wanted to get my hands on a sun machine
Your host ID should be a combination of the "machine type" (in your case 72 for an SS20) and the last 3 octets of your ethernet address. So in your case the host ID is 7220539f. Really hope to see this machine booting! I've got a soft spot for non-Intel based hardware (2 Sun workstations, 2 Sun servers, Amiga 500, and Atari 520ST).
Nice.
Had an SS20 many years ago with two 150MHz Hypersparcs where I installed Gentoo from stage 1. It took a while...
Using an Amiga to assist in recovering a SPARCstation isn't something you see often. Definitely a unique video.
It's a bit of a nod to Amiga development. Early AmigaOS development was done on SUN workstations (68K-based SUN 3 machines), and there are even references to SunOS commands in the early AmigaOS 1.x developer documentation.
In retrospect, they were probably doing earliest development on SUN 2 machines, since the SUN 3 line was introduced in 1985...same year as the original Amiga.
Wow. The memories…
Oh, I forgot it's still _Suntember._
Thanks for the backspace trick! I really needed that.
Glad it was helpful. Was one of the first thing I learned when I started working with SUN and Solaris machines. I'm not sure why SUN never fixed this.
I believe that’s just the PID not the system ID.
Also chuckling at how many folk don’t know how tty’s work, or that delete and backspace aren’t the same. Grin!
Did a lot of work on Sun boxes back in the day…
I particularly love the fact you fixed that sparc machine using an Amiga... My bro had an Amiga 500... Probably the best computers ever...
Well ahead of its time. If linux existed back then, the Amiga would surely have been linux based. As it is, it was a fully preemptive multitasking OS with many similarities to Unixes of the day, except designed for single-user. I loved being able to swipe the screen down and reveal my Macintosh VM running at full speed in the background whilst I was playing Lemmings.
I have an SGI Indy and it has exactly the same issue (battery backed ram storing the mac address and clock)
It died a long time ago and now every time I use it I have to manually enter the mac address upon bootup.
On the Indy the address is printed in plain text on a sticker on the back so wasn't as tricky as this, but still.
Some day I might get around to replacing the battery in the Dallas chip.... Some day ;)
Great video btw.
/Magnus
Thanks!
I did not know that about SGI machines....so SUN didn't do something completely unique with the MAC Addresses! I had an SGI Indy for a short time, but gave it away....kind of regret that now.
This was interesting. I like seeing systems based on alternative architectures to x86, especially those that are no longer in production. It also reminds me why I don't like using modern Linux that much... it's reliance on cryptic, arcane commands in the terminal and scripts, all inherited from UNIX. It was bad enough learning how to comfigure MSDOS back in the day, and Windows 3.1 etc. 😂
SPARC processors, or rather their 64-bit descendants the UltraSPARC line, are still in production, albeit only in server form. Both Oracle and Fujitsu produce UltraSPARC servers and Oracle is still updating Solaris. How long this will last is anyone's guess, though....they're generally considered "Legacy systems" by most customers, and I suspect neither Oracle or Fujitsu have grown their SPARC business much in the last decade.
Funny, i like using the shell for most stuff (but i have a been UNIX user since 1990....)
@@TheStefanskoglund1for my common tasks, and htop, I prefer it too.
You can have the task done in the time it takes a GUI application to open, even on modern SSDs (but it was far more of a time saver in 2010). Though for having to look-up lesser-used commands, often the time taken is similar to the overhead from a GUI utility. Though even then, the added flexibility to 7z and ffmpeg are worth it for me versus GUI 7zip and Handbrake.
I had an old 1995 IPC 25 MHz with 48 Mo ram, external drive, external cdrom, external tape, flat screen and optic mouse.
Loved this thing.
I used it running NetBSD and WindowMaker.
I've always consider Solaris being a pain, while AIX or Hp-UX were easy to use. Let's not talk about Linux or any BSD ...
Or maybe I was not enough experimented with Solaris :)
We also had a few modern Sun Blade, I was not impressed.
Even if they used Gnome instead of CDE :)
Damned those things were slow and boring ....
I gave the IPC to a friend when I moved from France to Canada, with the promise never to sell it and take care :)
No idea what he did with it later ...
I remember the old IPC and IPX "lunchbox" machines. I always wanted one, but never had one.
I think a lot of people kept those running SunOS 4 (i.e. the BSD-based predecessor to Solaris) rather than upgrade to Solaris. SUN kept supporting SunOS 4 until 2003!
I ran NetBSD or Debian on my older SUN machines, because Solaris was definitely slow....my main desktop for a while was an Ultra 30 and it could run Solaris 9 reasonably well...the Ultra 1 it replaced, not so much.
I remember Solaris feeling a lot more complicated that NetBSD and Linux back in the day. Although, Solaris 10 was definitely ahead of it's time with smf (like systemd in modern Linux), zones (a precursor to containers), and zfs (now ported to Linux).
Thanks for the fantastic tips. I have my SS10 fully working but not really a Unix person so went through the Dallas battery ram replacement and config rebuilt using the same proceedures as most others I suppose same MAC address! I am suprised that people hack to bits the Dallas chips as I just bought one (about 5 years ago) from a reputable electronics supply retailer. Wish I had done more of the install with the serial term as cutting and pasting is of course so much easier and the web installer slows matters down.
Dremeling the Dallas chips (or in this case an SG Thompson Timekeeper) is not worth the effort, IMO. Identical and/or compatible replacements are still available at $25-$30. Not cheap, but not super expensive either.
The only downside is when they go out of stock. It looks like I will have to wait until Nov/Dec to procure a replacement according to Mouser's restock estimate.
Apparently, the original Ultra (the 1, 2, 5, 10, 30, 60 and 80) series' clock battery chip _is_ discontinued with no suitable replacement.
I'm hopeful that one day, someone will create a full plug-in, no hack replacement using a lithium cell.
What an odd design of a PC; Reminds me of some early server gear. Pretty cool using the old Amiga to give it some love; I believe you can run modern Linux on that sucker.
The Linux kernel team has been stripping out support for old architectures. The last Debian release to support 32 bit SPARC machines was about 9 years ago.
See: www.phoronix.com/news/Linux-Drop-Most-SPARC32-v2
NetBSD support is still pretty strong, IIRC.
@@ClarkHale shout out to @renerebe as he is been doing some good with his own T2 distro; Worth a watch mate!
@@TradieTrev will definitely check his channel out!
@@ClarkHale He has a live channel also and the man is a machine how he codes! If you'll get support it'll be from a person like him!
@@ClarkHaleI had a long-running occasional-work attempt to get Debian to behave on 32-bit PPC, but even when technically supported it was as a second or third class arch with all sorts of bugs. So years ago, I came to the conclusion that the BSD options looked more promising. Even though it involves more compiling, but I’m not afraid of that. Somewhat validating to see you say it’s similar for SPARC.
very very very interesting - I have a sparc station 20 too, with pretty much the same issues. Although, I want to convert it into a NeXT OpenStep 4.2 Machine. I would ike to see how the story continues. When will you release the next episode?
The part that I'm waiting is out of stock until mid-Nov :(
Think the MAC address is on a sticker both on the rear AND inside !
Sun intentionally chose to use the host MAC address on ALL ethernet if in a machine - they observed that in the domain there a MAC address is used ie a network segment, the same address wont be used by more than one machine.
I wonder, believe IBM RS/6000 does the same thing. It is the PC (and Mac) side which went with card manufacturer specific and provided addresses.
I always liked these machines. Wonder how well a maxed out one with an SSD would run linux... guess the 32bit cpu might be a bit limiting these days
"Sun systems boot into FORTH" modern machines do, too - EFI is a FORTH interpreter.
No, UEFI does not have Forth as part of the standard.
You may be confusing UEFI with OpenFirmware, which was used by SUN, IBM (for their POWER-based systems), Apple (on some/all PowerPC-based Macintoshes), and some others. OpenFirmware is based around Forth. I don't believe this SS20 is technically an OpenFirmware machine, but the standard was largely inspired from SUN's firmware, so it's very close.
UEFI was originally created for Itanium systems, and then later used for x86/x86-64 and ARM. UEFI was developed after OpenFirmware, and Intel had several beefs with how OpenFirmware was designed, including the use of Forth and the device tree. There's some discussion of this on a HackerNews thread which also has a link to an Intel whitepaper:
news.ycombinator.com/item?id=15730382
@@ClarkHale Ooh, sorry. Looking at the current standard, I must have confused EBC (EFI Byte Code) with a FORTH. During the Itanium bring-up, there was a lot of interest from intel to move away from ISA-specific option ROMs and toward something portable, so they baked a preference for portable drivers into EFI.
"The network is the computer" 😁
I had a pizza box in the late aughts that I probably got for next to nothing. I then gave it away as part of a giant cleanout, kind of regret that now. Ordered a Blade 100 the other day, going to give getting Solaris 9 running a go on it...
i have repaired lots of crt monitors and vintage color and b/w tv's
With regards to MAC addresses bound to a system not the card, RARP uses that, basically for each IP net (before using multiple ip addresses in one ethernet network) you configured a RARP server to tell hosts which want's to know their assigned IP address while only knowing their MAC address the IP address they should use.
The IP addresses is in environment adminstratively assigned to ONE machine.
The client could then use their assigned IP address to broadcast for a bootp server.
This design was common in diskless workstation environments and for X terminals.
RARP and BOOTP is partly replaced with DHCP (though at least the DHCP servers in Debian is BRAIN-DEAD, they basically can only be depended on to do IP address assignment and
giving out DNS information ie domain and server addresses, NOT doing what BOOTP was able to do.)
This is what I did for most of my SPARC machines. Solaris had a system called Jumpstart that allowed you to define the configuration for a machine and then install that machine over the network. It relied on RARP+BOOTP and was much faster than trying to install over a 3x or 6x CD-ROM drive!
@@ClarkHale and the design was easily able to support 200-1000-2000 machines from ONE jump-start server.
Big-patch day ? Force all machines to reinstall from the jumpstart machine. 4 hours later - done.
Wow, true retro... But a im x86 fan :)
Well, someone has to like x86, you might as well be the one. 😂
SUN also produced x86 in the very late 80's machines, although they weren't fully IBM PC-compatible, and were a market failure.
Later, in the 2000's, they produced true standard x86 workstations and servers based around AMD processors. I used to have an Ultra 20 that had a AMD Opteron processor.
SUN also ported Solaris to x86 systems and fully open sourced it as OpenSolaris. Unfortunately, when Oracle bought out SUN, they eventually closed source Solaris again. However, a fork of OpenSolaris lives on the OpenIndiana project.
@@ClarkHale I didn't know that.
@@ClarkHale they did two generations of x86 machines - the Sun i386 machines of mid 80s (which was killed a few years post the SPARC launch) and the AMD thumper machines in different versions.
Sun did the first PCI66 implementation (PCI 33 /66 MHz and 32 or 64 bit wide) for the Ultra 30 (Sbus har lower latency than 32 bit PCI so is mostly faster over all even with it's slower clock so Sun to get a marketable improvement in performance compared with Sbus had to design a working 64 bit PCI implementation well before the PC crowd....)
i got one in my basement .. come get it
Where you at?
@@ClarkHale Rockford IL