Obsolete processors, 286 vs. 386

It's possible I'm remembering incorrectly and it was actually a 12MHz 286.

On a sunny day (Fri, 13 Oct 2017 07:29:44 -0400) it happened bitrex wrote in :

My first PC (I bought) was a 486DX2-66 running win3.1 on DRDOS It replaced my home build Z80 system, but my Z80 was faster as it had a RAM disk that it loaded the floppies in, so you would always work from RAM disk (was I/O mapped).

For my work I designed among other things ISA cards with stuff for that old 286 and 386. I never liked the architecture, and still don't. We worked close with IBM, I remember boss called me in his office one day and showed the first 386, it WAS blazingly fast.. Games? I dunno, I had a 3D game with planes on that DX2-66, needed special shutter glasses, shutters driven from the parport. When the evil Widows from RatMond came with win 98 (probably in 1998) that thing they dared sell as an OS (was actually just to kill Digital Research DOS) no longer would run on DRDOS, so and then my 3Dgame no longer worked either. In the same year however I found Linux on some CD.. and that was the end for that Ratmond crap.

Now with raspberries and pads? who needs a PC? ;-)

Unfortunately there's still a lot of stuff that's not built for ARM.

The 80286 was just a stop gap between 8086 and iAPX432 when Intel realized that they could not get decent performance out of the 432.

The segment system of 286 was stupid compared to later x86 models. In addition, you could switch the 286 from normal mode to protected mode, but switching back required setting up the reset vector to a DOS program and performing a processor reset to run a DOS program.

I acquired a 10 MHz 286 when I started my company, mainly for sending invoices and using it as a terminal emulator with disk storage.

There was a semi-graphical Windows 2.x running on 286, but more or less useless for any real work. It was much more convenient to use plain MS-DOS.

In retrospect I was lucky, when I bought the next computer (486) and invested in a sufficient large memory for running Windows NT 3.51 instead of wafting for Win95 introduction later that year. After reading horror stories about Win9x unreliability, NT3.51 was as stable as RSX-11 or VMS that I had been used to in the previous decades.

The article is simply bogus: The first available 386 PC was a 16 MHz model from Compaq and cost a small fortune, but it was a huge step forward compared to the (overclocked) 286 machines that preceeded it.

Terje

bitrex wrote:

My first (Microsoft compatible) PC was a 33Mhz 386 with a 64K direct mapped cache. I was one of the first on the block to get one since I just started working for the company making the chip sets. It ran rings around anything else that was available at the time, but there were initial problems with getting the cache working, and until it did it was less than impressive. Prior to that job I worked with a Data General Eclipse MV/8000, and going to the new job felt like I had stepped 10 years back in time. It took a while to get used to the idea I couldn't just start a new thread running, or have a background task working on something. I think we could have been much further along much faster if Microsoft never existed.

On a sunny day (Fri, 13 Oct 2017 09:12:11 -0400) it happened bitrex wrote in :

It was a bit of making fun, while writing that I was using 3 raspies, 1 PC, 2 laptops. ;-)

But of course you need to compile from source, so all that closed source stuff is out.

."

My first PC was an IBM-XT (8088). I upgraded to the 286 and enjoyed a noti ceable performance increase (along with color graphics).

When the 386 first came out my roommate got one and although it was faster than my 286, it didn't inspire me to drop the cash on one. When the DX-40 version came out I finally got one, and it was indeed much faster than my r ommate's SX-16.

When the 486 came out, I made the plunge first and got the DX-33 (I knew to avoid the SX line). Later my roommate got the DX2-66 and I was endlessly jealous at how much faster it compiled code and ran games.

Windows didn't become useful to me until the 486 era.

That old 386SX of ours had a long service life. It was my main machine up until I went to college in 1997 and switched to a Pentium 166. It stayed home and IIRC Dad used it from time to time as a word processing/email machine (with AOL) running Win 3.1 until at least circa

I think the only service it ever needed was a replacement PSU sometime around 1994.

BTW I think we got it around Christmas of 1991.

I was in skirmish with Boca regarding the nature of the PC market. I would periodically post on internal formers, single unit prices from Sunday SJMN ... way below Boca predictions. Note late 80s, makers on the other side of the pacific built up huge inventory of 286 machines for the holiday market ... then intel announces 386sx (significantly reduces system cost with lots of stuff incorporated into 386 processor chip) ... and there are enormous 286 fire sales.

Then head of Boca does large contract with Dataquest (since bought by Gartner) to do detailed study and future of PC market ... that calls for several hour video tape of silicon valley experts. I know the person running the Dataquest study and am asked to be one of the silicon valley experts ... I clear it with my management and Dataquest promises to garble my bio/intro so Boca won't recognize me.

recent post with my last post on internal forum (july 1992), just days before leaving IBM

An accurate monday-morning-quarterback prediction in 1992 for circa 1997 regarding still using a top of the line 386: your friends would laugh at you and you wouldn't be able to give it away.

I got into the PC scene late because I went with the Amiga platform and stuck with it up to the end of the 90s. I'd been working with and on other people's PCs and had a collection of defective and discarded components. My first Windoze PC was built up entirely from those old parts.

The CPU was a 200MHz Pentium overclocked to 266MHz. The motherboard had a chip that tended to overheat and I rigged up a heatsink for it. The hard disk was a marriage of two, one with dead electronics and the other with a damaged platter. The CD-ROM drive was put together from three defective ones. The PSU and the

The ISA sound card had a built-in amplifier that drove two good quality passive speakers. The graphics card was a SiS 6215 that came with two 256KB RAM chips onboard. I added another two chips desoldered from another card and ended up with a 1MB card.

I didn't bother with a cabinet (case). The motherboard sat on one edge on my desk, leaning against a wall. The PSU lay near it, the CD-ROM drive sat on top of the PSU and the HDD squatted on the CD-ROM drive.

False. 386 SX was 32 bit, same as 386DX, but with 16 bit bus.

We had a PC running Windows 2.0 as the controller for a SATE test system at Microdyne. It was designed and built by Scientific Atlanta when they copied the design of one of our telemetry products, and put it into another case. The only real difference was that they used a LCD display instead of LEDs, but they copied several Microdyne patented components.

We came into work one Monday morning to discover that the IT characters had wiped the hard drive, and installed Win 95 in ts place. They didn't back up the custom software, or the OS. They caused us to lose a week's production on that product line. After that, they had to ask permission to touch any non networked computer in the company.

a 286 with a hot-rod coprocessor "cyrix?" could easily beat a bare 386 (witn no floating point co-processor) at floating point tasks. most games didn't use floating point back then, stuff like spreadsheets and spice did.

Not ALL of it, you can run the pi version of mathematica for free. The H.264 codec too.

."

I dimly recall that the 386sx was optimized for its 32-bit instruction performance, and (when running the old 16-bit software) used extra clock cycles. And, with a 16-bit bus interface, any of the

With no L1 cache to speak of, that 32-bit machine on a 16-bit bus wasn't the big win in performance that one might hope for.

It isn't false.

The only time an i386SX really could outperform an equivalent clocked i286 was running 16-bit code that did a lot of multiword (32-bit or wider) computation, or code that did a whole lot of segment switching [which was significantly faster on the i386 than on the i286].

But most people would not have discovered this - apart from some enormous WYSIWYG page design apps, there just weren't many 16-bit programs that would run faster on a i386SX unless the clock rate also was significantly faster.

And the i386SX was abysmal when running 32-bit code ... on average only ~40% the performance of a DX at the same clock speed.

George