Forgot your password?
typodupeerror
Intel Hardware

Intel's Pentium Chip Turns 20 Today 197

Posted by Soulskill
from the and-many-moore dept.
girlmad writes "Intel's Pentium processor was launched 20 years ago today, a move that led to the firm becoming the dominant supplier of computer chips across the globe. This article has some original iComp benchmark scores, rating the 66MHz Pentium at a heady 565, compared with 297 for the 66MHz 486DX2, which was the fastest chip available prior to the Pentium launch."
This discussion has been archived. No new comments can be posted.

Intel's Pentium Chip Turns 20 Today

Comments Filter:
  • The rest of us made do with 60MHz versions.

    It really had to hurt Intel to have to back down on clock speeds for once. They didn't do that again until NetBurst burst.

    • Re: (Score:3, Interesting)

      by Joce640k (829181)

      The rest of us made do with 60MHz versions.

      We couldn't afford the cooling systems for the 66MHz version?

      (Or didn't want to live in a wind tunnel...)

      • Also couldn't afford the RAM, IIRC.

        The RAM speed was tied to the CPU speed (FSB speed), and since the fast CPUs were expensive to buy, the RAM which was only needed for them was overpriced too even though it was only barely faster than the RAM for the 60MHz models.

        • I distinctly remember going out and buying a $500 upgrade. I got an additional 4MB of RAM for $250, and upgraded from a 33MHz processor to a 66MHz for $250. That must have been a 486, it was a Compaq IIRC.

          • The 486DX-2 66 wasn't really the fastest processor. Throughput-wise, it had a mere 33 MHz bus. If you were serious, you had a 486DX 50, with the 50 MHz bus. But then you had to deal with the faster bus, which had compatibility issues; since so many people poked along with the DX-2, support for the 50 MHz bus wasn't as robust.

    • by erice (13380) on Friday March 22, 2013 @02:14PM (#43249157) Homepage

      The rest of us made do with 60MHz versions.

      It really had to hurt Intel to have to back down on clock speeds for once. They didn't do that again until NetBurst burst.

      And they did it for the same reason. The 60Mhz Pentium was the end of the line for 5V CPU's. It suffered from overheating problems due to its exceptionally high power consumption. The P90, 486DX2 and later Pentiums were 3.3V.

      It is also questionable the the P66 dethroned the 486DX2. The 50Mhz 486DX was widely believed to be faster than the 66Mhz 486DX2.

      • by Joce640k (829181)

        >

        The 50Mhz 486DX was widely believed to be faster than the 66Mhz 486DX2.

        That was the theory: 50MHz bus beats 33MHz bus.

        In practice: The DX was much more expensive and the extra 16mHz of the DX2 kicked the DX's ass when you were playing Doom. Which you were.

        • by LSD-OBS (183415)

          I had a DX4-100 (33MHz x 3) which I overclocked to DX4-120 (40MHz x 3) and it tore the other 486's some new assholes

        • by rickb928 (945187)

          In practice, you were running AutoCAD. and you wanted the DX50.

          • by Joce640k (829181)

            Nah.... if you were running Autocad you wanted a 386DX40 with third party math coprocessor (eg. Cyrix/ULSI).

      • by SJHillman (1966756) on Friday March 22, 2013 @02:50PM (#43249663)

        I still have a P90 that I use for old games. It runs Windows 98 like a beast, although my Cyrix P166+ naturally blows it away.
        The P90 has a dual 3.5"/5.25" floppy drive and a 2X CD-ROM drive. The part that surprises most people is that there is no cooling fan on the processor heatsink or power supply, just a small one on the back of the case.

      • It is also questionable the the P66 dethroned the 486DX2. The 50Mhz 486DX was widely believed to be faster than the 66Mhz 486DX2.

        The 486DX2 used a 33MHz front side bus and a "doubled" processor speed, while the 50MHz 486DX used both a 50MHz chip and bus. However, the 50MHz also had significant heat problems (I heard rumors that it could melt the solder).

    • At work we still have a Gateway P5-60 running, albeit slowly, as a hardware test machine.
  • by Anonymous Coward on Friday March 22, 2013 @01:57PM (#43248917)

    fdiv bug

  • Ahh, Pentium. (Score:5, Interesting)

    by ZorinLynx (31751) on Friday March 22, 2013 @01:59PM (#43248943) Homepage

    The 66MHz original Pentium. What a beast.

    It ran on a full TTL +5V. So it sucked down power. Lots of power. I've disassembled first generation Pentium chips, removing the golden cover that protects the die beneath. The die is HUGE! Much bigger than any current production CPU.

    In fact, the early models produced so much heat that we boggled at the big fans needed to cool them! It was one of the first Intel x86 chips that REQUIRED a fan for cooling. We used to run our 486DX2/66 and below fanless and they worked great.

    All this for only less than twice the performance, at three times the cost.

    The vast majority of us skipped the first generation Pentium, instead going for more affordable chips as the i486DX4/100 and the Am5x86/133, which was RIDICULOUSLY popular for several years! In fact, the latter was faster than a Pentium 75MHz for anything that didn't require the FPU. And not much needed the FPU back then.

    Then of course we laughed our asses off when the FDIV flaw became known. Clearly the Pentium was the #0.9999999998855 processor on the market!

    Ahh, memories.

    • There were rumors that Intel actually looked into alternative cooling methods for Pentium before those big ass fans ended up being the norm. There was supposedly one system that actually used freon.

      Also, you're not kidding about the die being huge on those - in those days Intel would take the defective units and encase them in acrylic and give them away as keychains. Now, the actual chip is so small you can't do that anymore.

    • Yeah, I had a P60 machine for years that ran Smoothwall and acted as a firewall, router, hub, and file server, lol.

      I don't recall them being uneconomical though. There were plenty of reasons to ditch the 486, it really was a much more limited chip in some ways. You really HAD to have a pentium to do a number of things, and Linux was quite happy with them.

      The old AMD K5's were pretty good, but they invariably were paired with horrible pieces of shit Taiwanese Winbond chip sets and other such drek. I had ONE

      • Re:Ahh, Pentium. (Score:5, Informative)

        by ZorinLynx (31751) on Friday March 22, 2013 @02:14PM (#43249185) Homepage

        For quite a bit of time, Intel and AMD CPUs used the same motherboards and chipsets. You'd get the motherboard you want, and then decide whether you wanted an Intel or AMD CPU in there.

        In fact, the whole reason for "Slot 1" with the Pentium II was to put a stop to this. They patented the slot mechanism and locked AMD out. I'm not sure why they couldn't patent the socket type; I'm guessing there was a legal reason why the pin arrangements weren't patentable.

        • Re:Ahh, Pentium. (Score:5, Informative)

          by MBCook (132727) <foobarsoft@foobarsoft.com> on Friday March 22, 2013 @02:19PM (#43249273) Homepage
          No, Slot 1 was to allow them to put the cache on the same board as the processor so they could speed it up. It quickly became unnecessary as later Pentium IIs and all(?) Pentium IIIs put the cache on die, making the slot unnecessary and expensive.
          • by XanC (644172)

            IIRC, the first round or two of P3s also used the slot design. It was only the "flip chip" versions, I want to say starting with the 733MHz, that integrated the L2 cache on the die and went to Socket 370. This was the time period when the Celerons were often faster than the full-priced Pentiums: the Celerons had less L2 cache, but it was on the die in a Socket 370. Easily overclockable too. Put a couple of those in an Abit BP6 and you're really ahead of your time.

            • That was the "Coppermine". I believe they clocked as low as 500 (5x100) or 533 (4x133), actually. They did steal the Celeron (which already had onboard cache, albeit half as much) socket 370, though most required new chipsets (there were some mainboard manufacturers who setup older chipset equipped boards to run the newer chips; they were generally better also, the early Coppermine chipsets had some production and then other issues.
              • by MBCook (132727)
                Celeron. I'd forgotten that was how those started. They were just P2s without any external cache (and thus without any actual performance).
            • by MBCook (132727)
              You're right that there were slot based PIIIs using the Katamai core, I owned one. When I wanted to buy a second processor years later I had a terrible time finding a non-coppermine version that I could use in my dual slot motherboard. I don't know if the L2 cache was still off-die at that point or not. I think digitalsolo is right that they didn't go on-die until they went to socket 370. That was one of the best computers I ever owned.
              • by xhrit (915936)
                "The Katamai core, I owned one... That was one of the best computers I ever owned."

                Agreed. I absolutely loved mine. It was a p3 667, Dual Slot 1 OR840, 2gb rambus. I bought two matching machines; 2 motherboards, 4 cpus, 4 gigs of ram, 2 voodoo 5s (later upgrading to GeForce 5s). One for myself and one for my girlfriend. We ran the machines for almost 5 years, and did not upgrade until the dual core 3.0 mhz pentium D was released.

                One of the p3 machines is still running today, as a firewall.

                On a side
            • by xhrit (915936)
              You are correct. For the longest time I ran a dual 667mhz pentium 3's in an Intel OR840 Dual Slot 1 Workstation Board, with 2 gigs of rambus.

              That machine was such a beast - it benchmarked faster then 1.8 ghz pentium 4s. It was not until dual core pentium D's were released that I was willing to give up using it as my desktop. In fact the machine is still running right now as a firewall.
          • by Nimey (114278)

            Slot 1 P2s and P3s had the secondary cache running at 1/2 the processor's speed, but the Slot 1 Celerons had full-speed (smaller) cache, which was one reason why the Celeron 300A was such a stupendous overclocker.

            • I remember having two of those (in socket form) humming at around 500 (112FSB) in a BP6. God that thing was fast...

          • by yuhong (1378501)

            Yea, what they actually did was to patent what is now the FSB with the Pentium Pro.

          • by BitZtream (692029)

            No, that was never the reason. There were several excuses used but none of them were ever actually legitimate reasons for doing it. You just weren't smart enough to realize that from an engineering perspective, every single excuse they used was exactly the opposite of the way reality works. Those chips all already had onboard cache for instance. Hell, the 386 had on chip cache. They just added a new block, named it something different and you bought the excuse hook, line and sinker.

      • by Shinobi (19308)

        K5 was an absolute pile of dogshit, even without Winbond chipsets...

        Sure, if you got a system with a K5 to actually stay stable, it was fairly fast in Integer apps, as long as you didn't need to hit the caches too hard... And if you did some actual gaming(Quake, Descent etc), or tooled around with 3D graphics in Lightwave or such, you ran into the fact that the FPU performance was worse than crap... Oh, and tended to make the system really unstable too, no matter what chipset. Alas, it was the only PC I cou

        • Mine worked GREAT in Linux. Windows definitely hated them. Then again I had little use for Windows...

          • by BitZtream (692029)

            Sigh, you remember things differently than reality.

            Linux 'worked great' on them because at that stage, it didn't really do many useful things and it certainly wasn't optimized and taking advantage of all the quirks and tricks the hardware had to offer. You're referring to a time in Linux's infancy.

    • I bought 11th hour, still content with my DX2/66. It played fine, until I got an upgrade. Then it ran awesome. I didn't realize how slow the game was running until then. Friggin amazing. Windows 95 could do things. Not just grind away and sort of do things.
      • Re:Ahh, Pentium. (Score:5, Interesting)

        by ZorinLynx (31751) on Friday March 22, 2013 @02:19PM (#43249261) Homepage

        I wonder if you had a cacheless 486 system. These were very common in the early 90s! There were even "fake cache" chips that motherboard vendors would put in to make it look like you had cache when you didn't.

        I suffered with such a system for a long time before realizing that it had no cache. I always wondered why my friend's 486 system felt so much faster, then I finally read about the cache issue in a magazine! Those were different times, when you couldn't just use Google to get an instant answer as to why something sucks.

        Being a broke teenager, I suffered with that cacheless 486SX/25 (overclocked to 33) from 1993 until 1996 when I finally got a job and upgraded to a Pentium 166MHz. It was like getting out of slow computer prison. :)

        • by timeOday (582209)
          Those were amazing times. I got a Dell Pentium 90 mid-1995, and it was over $3000. (To this day it is still the most expensive thing I've ever bought besides a home and cars/motorcycles). But the amazing part is that within a year it was somewhat outdated. But it got me through my CS program and so, I think, repaid itself many times over :)
        • by Nethead (1563)

          I remember back then I had a 386/40 board (CPU soldered on) that would run circles around the 486SX/25.

          Almost miss those days, Herc video card, I/O card, HD/FDD card, modem card, and fighting jumper settings for I/O addresses and interrupts. You almost had to sit down with pen and paper to map out your settings before you added anything. Then spend half the day getting config.sys and autoexec.bat and QEMM working right. Bog help you if you wanted to add a network card!

          Now excuse me as I go yell at the da

          • by BitZtream (692029)

            Yea, I miss the days of working with crappy hardware, crappy software, and a time when developers thought it was perfectly acceptable that pretty much every program on the computer should be its own OS and have drivers for everything it needed directly.

      • I bought 11th hour, still content with my DX2/66. It played fine, until I got an upgrade. Then it ran awesome. I didn't realize how slow the game was running until then. Friggin amazing. Windows 95 could do things. Not just grind away and sort of do things.

        Wow, I remember that game all the time. It seems like once a month or so I'll hear something in a soundtrack in a movie, or a song, that is so close to the very distinctive music in that game.

        Of all the memories of my life that get triggered by sounds th

        • I loved the game as well, as it wasn't only good, but had a special place for me because it was the first game I bought myself. It also helped that CompUSA had mistagged half of the stock for $0.99
    • by jest3r (458429)
      Those were the good 'ole days. My CPU path was something like this:

      486 DX2/66 -> Pentium Overdrive -> Pentium 200 -> Pentium Celeron 300A (over clocked to 450) ......

      Video cards went something like:

      Matrox Millenium -> Diamond Monster 3D -> 3DFX Voodoo II x 2 ......
      • by RatBastard (949)

        Let's see... if we stay just with Intel: 8086 4.7Mhz, 80286 (forgot the speed), 80386SX 20, 80386 33, 80486 DX2/66, Pentium 133, 233, Celery 350 (2, one overclocked to 400), P3 500ish, and a slew of Core X and iX chips, and my Xeon-fueled Mac Pros.

        If we open it up to other CPUs, well, how much time have you got?

        • by RatBastard (949)

          Getting old sucks... I meant a Celery 333 overclocked to 450. And the P233 is obviously a P266.

          As you were.

          • by KZigurs (638781)

            Somehow I suspected that this thread will be rich in low UID's ;)

            (my first one was 5 digit one which I carelessly lost by forgetting the password)

        • by unixisc (2429386)
          The way Windows evolved meant staying with Intel. Microsoft could have taken real advantage of the MIPS and Alphas by making NT a 64-bit workstation OS on those platforms, while letting Windows 95 be a 32-bit OS on Intel. Instead, in the long term, NT ended up being x86 only. In the meantime, all the other OSs that attempted to make it big fell by the wayside - OS/2, BeOS, Copland, NEXTSTEP (not counting the Apple merger), and the story was even uglier for most Unixes, since the entire RISC Unixstation i
    • It ran on a full TTL +5V. So it sucked down power. Lots of power. I've disassembled first generation Pentium chips, removing the golden cover that protects the die beneath. The die is HUGE! Much bigger than any current production CPU.

      It may have run on a TTL +5V, but it was BiCMOS. Weighing in at 300mm2 [wikipedia.org], it's less than a Westmere Xeon's 500mm2 [wikipedia.org] and I think that's a pretty fair comparison of potential customers.

    • by dj245 (732906)

      The 66MHz original Pentium. What a beast.

      It ran on a full TTL +5V. So it sucked down power. Lots of power. I've disassembled first generation Pentium chips, removing the golden cover that protects the die beneath. The die is HUGE! Much bigger than any current production CPU.

      These old chips have a non-insignificant amount of gold in them. according to this page [ozcopper.com], the original Pentiums have about $20 worth of gold in them.

    • by sootman (158191)

      > I've disassembled first generation Pentium chips,
      > removing the golden cover that protects the die
      > beneath.

      Speaking of which, there's real gold [ozcopper.com] in them thar chips! About $20 worth, if you still have any original Pentiums around.

      Pentium Pros have almost $50 worth. And to think, I gave away (or sold very cheaply -- I forget) a dual-PPro monster all those years ago...

    • by Dadoo (899435)

      In fact, the latter was faster than a Pentium 75MHz for anything that didn't require the FPU.

      Or external bus access. A Pentium 90 could comfortably play Quake; an Am5x86/133 could not, because it only had a 32-bit external bus. Pentiums had a 64-bit external bus.

    • by evilviper (135110)

      The die is HUGE! Much bigger than any current production CPU.

      No comparison to the Pentium Pro CPUs that came on the scene just a few years later, and then became the Xeon line. Socket 8 was mind-bogglingly huge.

    • by yuhong (1378501)

      In fact, the latter was faster than a Pentium 75MHz for anything that didn't require the FPU.

      Which is where the P75 part of the Am5x86-P75 name came from.

    • "Am5x86/133, which was RIDICULOUSLY popular for several years"

      One of those is still humming along in an old DOS/98 Gaming rig (at 4*40), no fan, just an epoxyed northbridge heatsink for cooling. I just need to modify an ATX psu to power that motherboard...

  • by dingen (958134) on Friday March 22, 2013 @02:04PM (#43249009)

    This article has some original iComp benchmark scores, rating the 66MHz Pentium at a heady 565, compared with 297 for the 66MHz 486DX2, which was the fastest chip available prior to the Pentium launch.

    I'm amazed by these scores. I remember having a fairly fast 486 DX4 @ ~100 MHz (probably by Cyrix or AMD perhaps) at the time the Pentiums started to become popular. I got the impression that a Pentium 66 or 75 would actually be a downgrade for me, but maybe that hadn't been the case.

    I eventually switched when the Pentium Overdrive came out, so I could keep my 486 mainboard but still have a faster Pentium chip in my machine. That was a pretty sweet deal.

    I can't believe this is all 20 years ago, it feels like only yesterday.

    • by ZorinLynx (31751)

      The Pentium's biggest strength was its FPU. It completely outclassed the 486's (per clock cycle) by a ridiculous margin.

      The problem is back then very few applications actually used the FPU, because there were still so many systems on the market without them. The 486SX was an insanely popular chip, and it lacked an FPU. There were still 386s floating around, and competitor CPUs as well.

      Once games like Quake started coming out, which used the FPU heavily, the Pentium became a lot more alluring because it was

      • by jest3r (458429)
        Diamond Monster 3D pass through card.

        That's was the ticket to Quake Awesomeness.
    • by freeweed (309734)

      Your 486DX4-100 was most certainly faster than a Pentium 66, and on par with a 75 if not a bit better. At least for the vast majority of software out at the time. My DX4 lasted me well into 1997, but by that point the affordable Pentiums were into the 200Mhz+ range and MMX was all the rage, so it became a more obvious upgrade.

      The only people who ever ran 66s and 75s when they were current were those with money to burn.

      Man, I miss how simple things were back then. When clockspeed actually meant something, an

  • by Kemanorel (127835) on Friday March 22, 2013 @02:07PM (#43249055)

    From his royal Weirdness...

    All About the Pentiums [youtube.com]

  • I remember seeing them but I can't track down the official release.

  • Perspective (Score:5, Interesting)

    by Paperweight (865007) on Friday March 22, 2013 @03:24PM (#43250113)

    If you performed a calculation that took a week to complete on a modern Core i7 2600k, you'd still be waiting for your Pentium 1 to finish the same calculation even with a 20 year head start!

    Source [wikipedia.org]

  • I feel so old now :(

  • There's a nice New Yorker podcast from a couple of years ago that discusses what went into picking the name: http://www.newyorker.com/online/2011/10/03/111003on_audio_colapinto [newyorker.com] . It was done by Lexicon Branding, who actually write code to break up words into phonems and then remix those sounds into new words. The program spits out lists of candidates that are then vetted by the linguists at Lexicon. I found it a really interesting discussion.
  • by labnet (457441)

    Back in the day, most engineers wish IBM had gone with motorollas 68000 rather than the address hobbled 8086 series. Oh how we hated paging 640k hell.

    • by slew (2918)

      Back in the day, most engineers wish IBM had gone with motorollas 68000 rather than the address hobbled 8086 series. Oh how we hated paging 640k hell.

      That might have just traded one set of problems for another. For example, Apple went the 68K route with their Macs. The early versions of the 68K family only implemented 24/32 address bits, so some clever programmers (like those on their OS team) hid all sorts of pointer tags in the MSBs (like lock-bits). When Apple finally transitioned to the 68020, those clever pointer-tags hacks came back to bite them in a major way.

      Maybe the 8088 wasn't the best overall technical choice, but from a business perspectiv

    • by yuhong (1378501)

      Personally my favorite topic is the MS OS/2 2.0 fiasco that is about the 386, which is much much worse: http://yuhongbao.blogspot.ca/2012/12/about-ms-os2-20-fiasco-px00307-and-dr.html [blogspot.ca]

  • Was the calculation done on a Pentium?

    (Sorry, had to make the obvious joke.)

  • If 66 MHz was state of the art in 1993 and Moore's Law predicts a doubling of density (effectively clock rate) every 18 months, then 20 years/1.5 years = 13.3. So clock rates should be 2^13.3 times faster today, about 10,000x, or 66 MHz x 10,000 = 6.6 GHz, which is actually twice as fast as today's max of 3.2 GHz.

    Given that 3.2 GHz Pentiums arrived on the market about 10 years ago (2002) but haven't moved since, it seems Moore's Law was clearly pessimistic for the Pentium's first 10 years and wildly optimi

    • You've got it Amdahl's vs Moore and it has always been there. Density is roughly doubling as per Moore. Amdahl pretty much means we have a limited opportunity to find more work to parallelize so we need to find more different things to do. You won't be able to parallelize the heck out of everything if for no other reason than some logic is inherently serial and waiting on slower I/O. So They split off to multicore but still doesn't create more parallel work. At some point silly things like predictive branch

The person who's taking you to lunch has no intention of paying.

Working...