Yeah, in the 1994 - 1997 era these WinNT RISC systems started to show up which became very popular with the Lightwave crowd. Not only would this be loads faster than the last model Amiga offered in 1994, the company went out of business. These were 100mHz+ compared to MC68040 models offered at nominal 25mHz with faster 3rd party add-on boards, but not this kind of speed for a while. The MC68060 accelerators were only good for 50-66mHz or so. Most of the folks in the circles I was familiar with opted for the DEC Alpha based RISC systems as they were very cost effective, also ran WinNT, but DEC developed a rather miraculous software-based emulation system for x86 compatibility called FX32. After the first couple times launching a new piece of software the system would instead run a copy made of new native code. It was so good it was even compatible with some x86 hardware drivers. I was able to do Premiere editing with a Personal Animation Recorder, which was a hardware based DDR, using a dedicated encoder/decoder board with attached SCSI drive, letting me do realtime broadcast quality video while any other PC would struggle to play anything higher resolution than 320x240 in full color with something like MPEG or Sorensen compression. These workstations had at least the perceived coolness of running the same chip as in an SGI, which was what I had at work during this period, until '99.
@@gregorymalchuk272 by '99 SGI were starting to lose their edge. But that year I had two on my desk while at DD. I had an Indigo2 Extreme, which was getting very long in the tooth but was reliable, and I had one of SGI's dual-Xeon WinNT systems, which was fast but a less than optimal environment for production. We did the best we could to make the WinNT system feel Unix-like, using Cygwin and its shell environment, since very little production work made use of the desktop environment or metaphor. And in '99 I spent a good chunk of the year working on the first X-Men, where I implemented image-based rendering for the first time in a feature film. It was also the first time we had systems fast enough, and a renderer fast enough, to raytrace (VEX Mantra). And later in the year I started working on Dracula 2000 at GMFX.
ntalpha was based primaly for tv-stations. they needed some kind of useable os in buisness and microsoft said "ok we will do the work" many videoediting machines based (not really later) with specialized hardware in the video editing branch with lower pricing from more than 250.000 dollar those went down to 50.000dollar. " the market makes the price " - More production in one generation and the engeneering costs get down. whak whak whigg horraaaaaa - no magic. but it seems to be that some parasites went into mass market and sucking all the blood from global states 😁
One of the big things that “saved” CISC was the decoupling of the architectural instruction set from the actual internal micro-architecture. Most CPUs today translate from one ISA to another on the fly. This gives your chip designers huge flexibility in designing the processor’s internals without breaking compatibility.
6:26 The BBC Micro did not have an ARM CPU, it was a MOS Technology 6502. The first ARM CPU was in the Acorn Archimedes line of computers beginning in 1987, which was sold under the BBC licensed brand mainly in the educational market. Those BBC branded models had red function keys (a throwback to its BBC Micro lineage,) the models intended for the home market had green function keys, and business market machines had gray function keys.
Adrian Black the Archimedes has a very similar feel to an amiga. The OS even had a similar feel. It didn't get near the amount of support that Commodore computers got, and still gets though. I have not tried RISC OS for the raspberry pi, just because there isn't near the amount of software that there is for the debian based distros. I still think it's interesting that RICS OS is available for a cheap and readily available platform for people to experience and experiment with.
I think that the confusion may be due to the fact that the BBC Micro branding was carried over into various models of the Archimedes up to the A30x0 series including the A3000 and, as RMC correctly states, the first ARM chip was produced as a second processor system for the BBC Micro connecting through the "Tube". By the time the Acorn RISC PC 600/700/A7000 hit the market, the whole BBC Micro branding had been discontinued.
and here is my Raspberry Pi running RiscOS 5.20+ with a modernising #ChooseOS drive image. and full supporting mobility hardware, next to my still used 1996 Rpc workstation, with notes. It boots in about 10 seconds and blazes on 700mhz... >> imgur.com/a/9jJjn1g
Really nice tour thank you. The most famous home computer here in the UK with RISC was the Acorn (of ARM fame) Archimedes. A really great little range of machines which went into most schools
BilisNegra sure, it was an interesting time with BBCs in schools and ZX Spectrums and other 8bit micros in our homes, which led to many programmers and IT professionals here. There has been a noticable dip in numbers as access to easily programmable machines and school lessons vanished. The Raspberry Pi was created to try and recreate this and get more kids interested again, and the creator Eben Upton was also part of the team who created the BBC originally.
The Archimedes was the first ARM home computer, if we ignore the ARM1 second processor. Released for the BBC Micro during #RISC development. HOWEVER... this machine may have a RISC badge on it but it is not a RISC PC - of any kind, sorry. Acorn never ported RISC-OS to MIps/PowerPC - They only built machines on the ARM610 to StrongARM chipset. Those ran Windows on a second CPU, plugged into the tandom socket in the machine. Mine has a blue x586 card at 300mhz. There is a long list of corrections I could post about the history, code impact and Risc-ness in this video, But I won't brow-beat you, not when you've made so much effort on it. Still, this isn't a RISC PC - or an Acorn-clone product (such as they ever existed) and cannot run Acorn operating systems. The Mips/PPC platform isn't the forefather of ARM #Cortex, as seen on all smallish devices on earth now. It's a different chipset, one that was never fully adopter due the Wintel pushing CISC to the bitter death... as it still continues to do. Research #RISC5 (HUGE money 'open source') and #Sifive custom SOC systems. For where Windows is trying to go now. In my opinion this is an attempt to force-adopt a less able (and largely copied) risc standard to avoid paying the tiny ARM licence in production. Having repeatedly failed to port any Win version fully to ARM - Wiot = failed - WinRT = failed and currenting WOA = also failing. Microsoft are a big and very competent company. If they wanted a full library on WinARM64. They'd have done it... In comparison, RISC5 ported 85% of the #Debian library to 'new risc' in under 6 moths. Linus Tech Tips #LTT has a video on this but it is all badly ill-informed and mostly advertising for Sifive. Disregarding 30 years of optimisation, when modern ARM SOC's often have Decacores (10) and run a total of 15ghz or more, by cycle Sorry there is a lot, I hope breaking it up helps. Here (link).. is a Risc PC as last seen in production by Acorn. >> www.retro-kit.co.uk/page.cfm/content/Acorn-RiscPC
I used to have an old pre-release version of Windows NT5 (before they changed the name to 2000 (I mean, NT5 was supposed to ship in 1997, but you know, it's going to be easier to just ship in 1999 and call it 2000 so it sounds like it's from the future)) that would still install on alpha/mips(/and some other RISC, I think) machines. They could have left in support in the actual release, but I think they felt like it was time to let the totally awesome future pass them by.
as an old os/2 user i used it for a multiline bbs - there was multitasking (for real and it was MULTITASKING OS) inside with such kind of 486sx25 most games and hardware run faster as in dos, wfw311 AND win95 or if you want nt 3.51 and nt4 it is on your own choice to believe a open hardware plattform is the better choice such like "non optimised" windows or you get a specialized, hardware dependent OS. _" now i must get to work "_ is my comupter screaming
If it was my machine I'd probably look into installing a BSD unix. It's been a while since I used BSD, but it has an extensive "port"-system which basically means that all apps (many thousands!) can be automatically downloaded and compiled (from source-code) on your system; the latter being important as you probably won't find many R4600 executables out there. That should allow you to run alot of software on it. Thanx for the video, used to drool over these things.
BSD, or check for a Linux port to MIPS. Not that there's much difference these days. BSD has a Linux (system calls, architecture neutral) emulation subsystem. You should start by looking at the old 32 bit OSes for MIPS support. In most cases you then look for good C/C++ compiler(s) to get programs compiled from source. That BSD ports software sounds like a real time saver. Arch Linux also compiles programs it installs from source, I have no actual experience with either Arch or BSD.
thanks for the great video! would you be interested in installing OpenBSD? it was forked from NetBSD, as i recall, but each iteration (whole number version) is not backwards compatible or updatable because of its global architecture, the legacy code is deprecated to keep the code pure, small, simple and improved. i have not used linux, except for Ubuntu Studio and not any other OS except MacOS
it's a plain copy of nt4 - he must slipstream sp3 with extra ide drivers or mabe sp4 and sp6 for mips to support larger harddrives remind limitations in the past which going on nowadays
Don't forget the terminator on the SCSI card :) Your videos bring back a lot of memories. EISA cards shared with IBM Microchannel that they could be configured in BIOS and not with jumper settings. What a bonus in the day!
If you haven't formatted the scsi drive yet, then you can easily get the my administrator password as well as all other account names and passwords.. you'll want to run either LCP or ophcrack against the SAM database. There are actually dozens of tools to do this but the two I mentioned above work wonderfully. Let us see it in all of its natively setup glory! Thanks and good luck.
I would've suggested ntpwedit, a simple tool that can just wipe the admin password with one click. So many tools to get this done. Windows passwords are a joke. This was like looking at a safe held together with a luggage lock and saying "well, I guess we'll never get in there, let's build a new safe". :(
That brought back memories. I used to work for IDT, a MIPS partner that produced the processors, and part of my job was to help developers port applications to Windows NT on MIPS. I had a pool of Deskstation machines that I would loan out to developers. They were R4000 based not R4600 however. The R4600 was just coming out at the time and we had some R4600 machines in house.
The fan issue is likely due to a bad power supply. It can also happen due to devices inside the computer with bad caps, and when devices have bad caps, they can sometimes can load a power supply quite a bit. So be sure to check the voltage getting to the fan with a scope and or multimeter.
We had a couple DeskStation Raptor Reflex towers at my job in the late 90s, and they were mainly used for LightWave and Photoshop. What's interesting was that they had an x86 emulator called FX!32 from Digital, which was how we could run Photoshop. There's very little online about it but it not only emulated an x86, it would find ways to improve the performance over time.
The Raptor Reflex models were Alpha-based, as I understand it. DeskStation seemed to shadow Digital (DEC) and initially planned to ride the ACE bandwagon, even wanting to offer Ultrix when the SCO Unix product for MIPS was delayed. When DEC abandoned MIPS and moved to Alpha, DeskStation followed suit. FX!32 was, of course, DEC's compatibility product for NT on Alpha.
Love seeing videos on these old PCs. I was a kid when this was new so I never really got a chance to work on them myself, although I did get to work on a bunch of old Sparc systems later on.
@@Anvilshock a terminator is not a dongle, dodoid when the software say ding the time must say dong a terminator is only a resistor in the end of a chain. it's unnessessary to be used without any devicechain
@@loschwahn723 I see you're the reason why the saying exists, "It's better to be silent and let people suspect you an idiot than to speak up and remove all doubt".
Great video! I have already heard about RISC processors, but never seen one being used. Nice piece of computer history. You sir have earned a new subscriber! Cheers.
The x86 architecture we use today is in fact a RISC processor disguised as CISC for backwards compatibility. Since Pentium, Intel used a microcode translation layer to convert the legacy CISC instructions to internal RISC ones, and you can also use the RISC instructions directly since the newer (SSE) revisions of the instruction set.
Very cool look back in time. 64mb ram in 4 sticks means they are 16mb a piece, I remember buying 1 16mb simm for a 486 in 92 or 93 and it was almost 200bucks. Thanks for the upload!!
LIke many others intelligently commenting (am I still on YT?), this appeared as a recommended video and I'm glad I took the chance on that click. Brought me back to the late 90's when I worked "desktop" as it was known back then and we marveled at the DEC Alpha and it's "crazy" 1 gigahertz speed. My side business was building PCs and I'd tear my hair out getting SCSI, Soundblaster, IDE, etc. cards to work nicely with the rest of the system and getting the OS operational. Queue Obi-wan meme with "VESA LOCAL BUS? THAT IS A NAME I HAVE NOT HEARD FOR A LONG, LONG TIME." Nice job with the video and nice job cleaning up those cards before filming :)
Vesa Local Bus. That takes me back to 1995. First PC ever 486 DX-2 80 MHz with a 1 mb VLB graphics card. 4 mb of RAM (later upgraded to 8 mb) and a spacious 512 mb hard disk. Fun times.
Probably the smartest thing Microsoft ever did was to hire Dave Cutler away from DEC to be the architect of NT. The NT name is long gone from their OS, but it is still NT, nonetheless.
Funny you mention it. I remember when reading through the NT API in the 90s going "Hey! This ain't half bad! There is a proper design here and not just a stinkin' pile of ad hoc bolt-on features. Guess I should cut MS some slack.". Then I learned that a DEC guy did it ;) ...probably also a good idea to get Anders Hejlsberg from Borland to design C# and so on.
Very cool. I was aware there was a MIPS release of Windows NT (there was an Alpha release too I think) but never saw MIPS chips outside of SGI machines. Have had a few of those but all running Irix of course. SGI did a Windows NT machine (Visual Workstation) later but it was a strange machine with an Intel chip. This machine is interesting, presumably the maker really thought RISC was the future most of us thought it was at the time. Looks very 'off the shelf' compared to the SGI machines of the time!
MIPS architecture CPUs were commonly used on many 2000's PDA and Pocket PCs, and in some game consoles (Sega Saturn, Dreamcast, Sony PlayStation, Nintendo 64...). www.mips.com/blog/five-most-iconic-devices-to-use-mips-cpus/
I was doing a training course at Pyramid Computers in early 1994 and Bull computers were also there, testing their prototype MIPS NT desktop PC's - they looked IDENTICAL to this one with the slanted top so I wonder if the Bull ones were actually Tyne manufactured (or Tyne re-branded them?). Pyramid were a manufacturer of MIPS powered mini computers, using MIPS R4400 CPU's and similar running UNIX (SVR4 - DC/OSx) in their big 42-U Nile and the MIPS R10000 CPU in their Reliant RM1000 machines (A couple of these RM1000's ran Link cashpoints and BT had one for many years in the UK - I saw the demo machine they'd built!). I don't think Bull ever sold their MIPS Workstations, and Pyramid got bought by Siemens Nixdorf ultimately.
Fun fact: NT ran on MIPS before it ran on x86. ARC was also the firmware interface that the newer SGI systems used, and was the NT flavor of firmware/PALcode for the DEC Alpha. NetBSD runs on this machine (NetBSD/arc), so there's an option if you want a contemporary OS for it.
Wow, that BIOS looks so awesome when compared to the modern "splashscreen" ones that hide all the tech details so as not to hurt the poor user's feelings. All the blinky tables, delicious!
There are some ARM-based Windows mini-PCs out there (usually found on AliExpress). They're basically Surface RT hardware with some USB and HDMI in place of the touchscreen. So, in a way, we've already got desktop PCs with RISC back on the market.
The Mips version of lightwave was a bit of a pain and the HIIP image loaders were such a pain as the 3rd party company that made those never got the libraries to work reliably on MIPS NT. Lightwave had been ported to MIPS for screamernet (The lightwave render farm software) and at one point I had a box of those motherboards that was used for the Screamer box. I believe they were actually desk station boards. I do have a copy of the full MIPS lightwave install around here that I could make an iso of if you want it. Lightwave at that time was ported to MIPS and Dec alpha and I believe Deskstation did make an alpha workstation as well but it's been a few decades and my memory isn't what it was. If I dig up any documentation on deskstation I'll post it.
I do have a few dongles I can't remember if they would work cross platform I do have some of the older lightwave dongle programming software that would be about the right vintage. I'll dig about and see if any of that stuff works. It was the 3rd party image loaders that were problematic but more a pain for the devs at the time. I believe when it hit release it was at least working.
If you want to replace the clock chip and make it look nice and have a removable battery, you could look into the alt12887 project where you build a new 12887 using an SOIC-24 DS12885 or TI/Benchmarq BQ3285 , a 6pf 32.768khz oscillator and a cr1220 battery holder and two rows of 12pin headers . Or for cheaper you can get a bunch of ds12885s from Aliexpress in DIP-24 encapsulation bend up pins 2 and 3 and solder a 6pf 32.768khz crystal between them and then glue whatever type of battery holder you want on top of the dip-24 chip and then solder the battery positive to pin 20 and the battery negative to pin 16. Their are many posibilities of different battery configurations, you could use a cr12xx or cr16xx holder which would fit nicely on the top of the chip,cr20xx or larger holder would stick of the side of the chip a lot, a cr2 photo battery holder would probably fit nicely. a cr123a holder would be a bit bigger than the chip, you could also get a 2xAA battery holder and a couple energizer ultimate lithium batteries and bend pins 16 and 20 up on the chip and use a 5 position DuPont connector with the outer position populated and connected to the battery holder.
MIPS is quite common for consumer wireless routers, and DD-WRT is among the most common (aftermarket, at least) Linux-based... firmware? Operating system? on them. While I'm not suggesting trying to run DD-WRT on it, it's possible that you can get a MIPS-based Linux distribution to run on it.
Generally, speaking you should be able to use ARCLoad to get a Linux kernel running though my experience with Debian/mipsel is it's pretty hit or miss. NetBSD on the other hand supports ARC fairly well though it lists the DeskStation as unstable: wiki.netbsd.org/ports/arc/.
I love how the bios shows everything getting detected, tested and started. Intersting machine, it feels strange because it's exactly like a classic x86 pc but it isn't on x86 arch. Speaking fo the dallas RTC+Battery chip, I've the chills thinking about it because of my old IBM PS/1 model 2011: if the battery is depleted the floppy drive doesn't work!
Adrian Black yup, I know about this mod, sadly I never had time to mess with it and I preferred to find on eBay new rtc chips... They didn't last long as expected XD
ARM1 was completed in 1985 and put to use in the original Acorn Archimedes line of computers released in 1987. AMR2 was completed in 1989 and put to use in the Acorn Archimedes A3000 computer, the last Acorn microcomputer released that bore the "BBC Microcomputer" nomenclature. The BBC Micro was released in 1981, using the MOS 6502 CPU as used in the Commodore PET / VIC20, and Apple II. As a sidenote, the BBC Micro was derived from an Acorn prototype platform internally named "Proton" which was supposed to be based around a 16 Bit CPU with the 6502 and associated I/O devices only acting as an "Intelligent" I/O platform for the 16 bit host. This idea made it to launch in the BBC Micro in the form of the "Tube" bus, which allows second processors to be easily connected to the BBC micro and use the micro as simply an I/O Host. The ARM development was done on the BBC Micro using this bus after Acorn visited MOS technologies and felt their approach to designing a 16 bit extension of the 6502 was a bit too "Cottage industry" (their exact words) and Sophie Wilson felt her team could do better at home; and thus the ARM architecture was born. Even back during the prototyping days of ARM in the mid-80s, it was an incredibly power saving chip. Their original prototype board provided enough power to run the processor through the protection diodes on the TUBE bus - which they discovered by accident when they connected it one time and forgot to connect the power supply to the ARM prototype board. It was working anyway. 34 years on, one can still connect a new ARM second processor to the TUBE bus on the BBC Micro. Yes, the Raspberry Pi can be interfaced with the BBC Micro via the TUBE bus allowing the Beeb to execute code on the 1GHz+ ARM on the Raspberry Pi. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-RLaeaoHXl4I.html And yes, someone's porting Doom to the 1981 platform and it's modern co-processor :) ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-V6HQrXCRZsw.html Honestly, growing up in a country and a time where Acorn Computers reigned supreme, at home and at school, gave me (and many others) great perspective into the rest of the computing world. Everything else just seemed so primitive. Toy like. Infantile. I'm very happy Acorn and their engineers live on in the ARM architecture to this day, I believe they've now licensed over 12 billion ARM cores.
The original Archimedes line, as well as the A3000 which was effectively a continuation of that line, all used ARM2 processors. The ARM1 was confined to Acorn's ARM evaluation products.
The section about RISC ARM CPUs was very interesting. I always wondered what made the Apples so different and incompatible. Now I know. Thanks for the history!
So easy to reset the password for the admin account on NT. You could have just booted a live Linux CD, mounted the windows drive and used any number of password tools to reset the account. Maybe next time! Great video!
You brought back some memories for me. In early 2000 I started working as a sysadmin for a university that was a big DEC shop. They had investigated NT on AlphaServers (I believe some 1000s ran it and a couple 4100s were bought for possible use) in the late 90s but didn't find it of great utility and the announcement that 2000 wouldn't be on that platform was the final nail. I only got to briefly play with them in the last week we ran it and the machines were re-purposed for VMS work, which they ran until sometime in the 2010s - I had left the place before that was decommissioned.
I did admin work on DEC Alphas, back in the day. All running OSF/1 aka Tru64 Unix. The amusing thing was, the motherboard battery on one of them died, so it forgot what OS it was supposed to boot. Guess what it defaulted to? Windows NT!
The magnet(s) inside the CPU fan may have depolarized and the motor will just exert a lower force. You can apparently see this in model trains too. Take with a grain of salt since I'm just repeating what others have said and not looked into it.
Nokia was concerned about the amount of data executable software for the ARM would take up. So ARM came up with the Thumb instruction set, which was a cut-down version where each instruction was 16 bits. The processor could run in thumb or normal (32-bit instruction) modes. C code compiled with this took up less room, which still mattered for phones. It did not run as fast however, esp. for hand-coded, so the 32-bit mode was still very useful, as well as for backwards compatibility. Because of this ARM got the deal which went on to make them a company worth billions. I used to write software in ARM assembler for the Archimedes and later the Game Boy Advance. I did some code for the PS1, but we mostly coded that in C, rather than MIPS.
Modern x86 CPU's are actually RISC, the CISC instructions are handled by microcode to translate them into a combination of RISC commands. AMD started this around the Athalon XP era, Intel implemented it in the Core line of CPU's.
> CISC instructions are handled by microcode to translate them into a combination of RISC > commands. AMD started this around the Athalon XP era, Intel implemented it in the > Core line of CPU's. Much earlier. Intel started doing this with the Pentium Pro 1995. AMD with the K5 in 1996.
I put together a small renderfarm of DEC Alpha-based machines running NT and Mental Ray. Unfortunately the amount of data that had to be shifted across the network for each frame meant that the speedup was relatively small for the 480i-res frames we were rendering. Upgrading the SGI workstations used for modelling to 100BT ethernet would have helped, but the cards were ~almost as expensive as the Alpha machines.
You should probably be able to run NetBSD on this. Together with its ports that should give you access to most open source stuff. Not sure if you can find a Linux version that can support this and still has repositories to install software from.
Oh the wonderful sound of multiple SCSI Hard Drives spinning up lol. PC's have become for the most part, silent if using SSD's. Thank you for your video. I enjoyed it. 👍🙂👍
4:54 - Intel introduced its EM64T (Expanded Memory 64 Technology) starting with Core 2 Duo processors to support 64-bit OSes. The Intel Xeon, was initially a native 64-bit processor, but later versions of Intel Xeon and Core CPUs are x86_64, meaning that they can run both 32-bit and 64-bit OSes and applications.
EM64T on Desktop began to be implemented with Pentium 4 Prescott under emulation, (depending on the model, it is not supported, and is only exclusive to socket 775) and for notebook since Merom. And as a curiosity, AMD was the first to implement the X86_64 architecture, but the first athlon 64 does not support CMPXCHG16B, which Intel implemented, then AMD implemented this feature in the K9. The most important change of the CISC architecture in Intel was with the introduction of P6 in 1995 with the Pentium Pro, this architecture is the first time it is based on a RISC CPU, and it was an architecture it was even superior to the new NetBurst architecture used in Pentium 4 (Netburst lived in the shade from birth until it was discontinued with Pentium D, the P6 architecture performed much more at lower consumption, the Pentium M at 2.26 GHz of 27w was equivalent to a Pentium 4 working at 4GHz at 140w of TDP, and then with the P6 architecture the core 2 was born that were improved versions of Merom), and currently Intel has continued to use the base of the P6 for current processors (realize that the meltdown bug affects up to the Pentium 2). And then a change in the P6 architecture that suffered, was the first time to implement an AMD instruction of 64 BITS, but that Intel improved by adding an additional instruction set called CMPXCHG16B, which I call EM64T. A while ago I made this image on the Intel CPU timeline: i.ibb.co/0nL0wwd/c429cb020e8c4dba6c6ff462a989c0b4.png
Got to this video through RU-vid recommendations. Sorry for reviving this old comments thread. It's also the first time I see Windows NT on a MIPS processor. I knew there was a MIPS port of NT, but I've never seen an actual machine that could run this port. I bet this thing can run a MIPS port of Debian Linux as well. Some corrections about Windows on ARM: Microsoft actually made an ARM port of Windows back in Windows 7 days. There was an experimental version of Windows 7 on ARM which was demoed but never actually released. Windows 8 had a fully functional ARM port in form of Windows RT (artificially limited to just Store or MS signed desktop apps) and Windows 10 has an ARM port as well.
Funny thing seeing this video now - I picked up a Surface tablet that unbeknownst to me is running an ARM64 Windows 10 (or, actually Windows 11 now) - Microsoft even seems to emulate x86 on this thing, cause it'll happily run Steam and apps from that, and it's doing pretty well. I keep using it to fall asleep to RU-vid without the power plugged in and the next morning it's usually STILL alive and playing something, so the battery life extension is no joke, either. It's actually really snappy. I put Windows Pro for Workstations on it and it's also pretty neat in the feature department. Makes for an impressive laptop alright. :) EDIT: oh and the MIPS and PPC did go strong in games consoles, the PS2 used an R3000, and everything in the PS3's generation used PPC. It took until the PS4 to see more x86, which the rumor mill has it is why they're so power hungry.
This goes back to the early days of Windows NT. There were "extraction layers" so that NT could support many different processors, and architectures. As the Intel processors became more powerful. they mostly eclipsed the other ones. MIPS, Alpha, and PowerPC, were just a few of what was available back then.
ARM chips were developed by Acorn and originally stood for Acorn Risc Machine and was available as an add on for the BBC to develop and test with, but it was never used in a BBC computer. The first computer it was used in was the Acorn Archimedes 300 series back in 1987. The ARM 1 chip was missing several key instructions and was never built into a computer and was used as a demomstrator at a number of computer shows to preview the technology. As for claims that the G3 was not as fast as an intel Pentium, Photoshop was demonstrated many times with a G3 PowerMac completing workflows faster with a G3 processor clocked at less than half the Windows machine next to it running the same workflows, again similar with AVID systems and many more. By the time of the G5 debacle where IBM and Motorola both failed to deliver decent performance, Intel had abandoned the P4 and was working on the core processors, which finally put Intel on a decent footing as they had been thrashed in floating point performance and general efficiency for years by the Athlon and were still embarrassed by having to adopt AMD's 64 bit extensions after the failure of itanium. Risc support was added in Windows 8 with Windows RT, not Windows 10, the latter only corrects the mistakes made earlier and paired with the right ARM based chips, can deliver better performance overall, never mind better performance per watt. With ARM based Windows 10 laptops available, well supported and well liked and ARM based Macbooks rumored for the end of the year, Intel should be worried, if not regretting selling off all their ARM portfolio to their competitors in the early 2000's
Technically you are correct, the first ARM, or more accurately the first simulated/emulated arm was indeed on the BBC micro. The first fully blown production ARM machine however was the Acorn Archimedes. "Acorn also employed the (BBC Micro) machine to simulate and develop the ARM architecture which, many years later, has become hugely successful for embedded systems, including tablets and mobile phones. In 2013, ARM was the most widely used 32-bit instruction set architecture." Source - Wikipedia - en.wikipedia.org/wiki/BBC_Micro
Should be able to run Microsoft Visual C++ on there, at least Version 4 had a dedicated MIPS version. Though it also sounds like version 6 might be able to do it, though I was only able to find a copy of 4. Would at least get you some large application to run, and possibly allow you to compile over some other applications.
What you need specifically is Microsoft Visual C++ for RISC Edition, which was made for 2-4, though there are reports of 5-6 as well. You need a RISC edition that explicately includes MIPS support as some only have Alpha, and others have Alpha, MIPS and PPC. It has to be installed on an x86 machine and then you can cross-compile and deploy to your MIPS one over the network (you can hook up standard ethernet to that network card with a AUI adapter from the net. NOTE: you may want to use the existing card vs. replacing it as I doubt you will find a card with ARCBIOS support).
Not sure if anyone else has pointed it out, but I'm pretty sure this Tyne lived its life as a rendering server alongside an Amiga - that Raptor folder that was on the SCSI drive is dropping some big hints. I've an old Amiga World in my collection that was raving about the Raptor rendering system, might have to dig up an issue number...
RISC was not a response to the backwards compatibility "bloat" of CISC, it was a project to create a simpler instruction set in general. The original intent of RISC/ARM was to be a simpler instruction set to even the 8086/8088 or 6502.
Nuh uh, ARM2 didn't show up until the Acorn Archimedes. It was not released on the BBC Micro, even though, development started on the BBC Micro, due to the Tube bus.
4:36 No, it was not Intel that took x86 to 64 bits. The credit for that goes to AMD. Intel figured that x86 would run out of steam by the time 64 bits was needed, so it developed an entirely new architecture, called “Itanium”. Which kind of fizzled in the marketplace, particularly after AMD came out with those 64-bit extensions to x86.
I understand where the confusion arose re: ARM/BBC Micro. The original ARM chip was designed and prototyped on a BBC Model B. A few years back, an expansion board was developed to allow a BBC Micro user to use a Raspberry Pi as a coprocessor. So, in a sense, some BBC Micros actually _do_ have ARM-based (co)processors in them! [cool machine btw... gotta love the 1990s RISC-based oddballs like the Tyne]
With a lasermaster rip box than drove a printer the computer used a form of nt3.51. The install cd was really like hardware. It formatted the drive then installed parts of nt3.51 then the lasermaster printer rip program. Nt3.51 had in its code looking for the cpu type. If it say a pentium the install would work. If the cpu was a pentium pro or higher the install would fail. Later on some folks found that one could hack and change lines of code to allow usage of a faster cpu. The original lasermaster printer was a custom bulk fed 36" inkjet with 4 bulk 500ml inks. The ink sets could be either water based or pigmented for outdoor signage. Each inkpack fed a custom hp26a cartridge with a tube filling it. Each new bulk ink cartridge had a color calibrator chip that would fine tune the color. The lasermaster system was a custom novajet 36" printer with a rip box and a wazoo calubrator that read the test patteens. The unit was 40k at first then for years in the 25 to 35k range
I don't know why RU-vid recommended this to me. I had nearly forgotten about using Windows NT. Back in the day I had Windows NT Workstation 3.5 on floppy disks. It ran on x86 architecture. Another part of the department had DEC Alpha machines, and were raving about X. I clearly remember the DEC stuff using coaxial Ethernet and those AIU (I think that is what they were) connectors. The building had coaxial cable running between floors to get 10 Mbit half duplex Ethernet connectivity. Eventually we got Windows NTAS, also on a tray of floppy disks. Around the same time we finally got fiber optic cable between buildings. All this was good stuff for the time, but I wouldn't bother going back to it. Just thinking of the power consumed by all that old technology, those long SCSI cables, the IDE and floppy cables, serial mouse, partitioning disks and then having to format before install. UGH!
In that era the PC and SC designations meant "Primary Cache" and "Secondary Cache" The PC chips were usually just the chip, and the SC chips were mounted on a card with additional cache, more what we might call L2 or L3 cache today.
On an SGI Indy faq, it recommends against using the 4600PC due to a lack of L2 cache. www.vogonswiki.com/index.php/SGI_Indy#R4600 So an interesting upgrade would be to acquire a 4600SC or 5000SC. The chips were all plug compatible across the 4400, 4600, and 5000. Additionally the 4600 was a more mainstream processor than the 4400. The 4400 was more geared towards floating point performance, and the 4600 was geared more towards integer performance. The 5000 added the MIPS IV instruction set and had larger on-chip caches.
A RISC has to execute more instructions to accomplish the same thing as a CISC, so instructions per second are not meaningful. Also, the 'E' in DEC stands for 'Equipment' not 'Electronics'. Also the company that produced SPARC is called SUN Microsystems, not SUN Workstations. Also, ARM is for Advanced RISC Machines, not Acorn RISC Machines.
BBC Micro models A & B were based on the MOS 6502 8 bit platform, Acorn build development machines with a second RISC chip for development of the ARM platform which went into the Archimedes, but these were never generally released.
