PC processors entered the Gigahertz era today in the year 2000 with AMD's Athlon — AMD hit marketing gold with its 1 GHz Athlon, beat Intel by a nose

News
By published

Consumer PCs have long abandoned the multi-GHz race for core count and NPU inflation.

AMD Athlon CPU for Slot A
(Image credit: Maddmaxstar)

Today, in the year 2000, AMD shipped an undisputable processor milestone, its 1 GHz Athlon CPU. Thus, the Gigahertz PC era was born. AMD scored marketing gold ahead of its powerful rival Intel. PC industry heavy hitters of the time, Compaq and Gateway, were key partners, and the first pre-built 1 GHz system deliveries began the following week.

Tom’s Hardware had previewed the new Athlon K7 processors back in August 1999 and reviewed a 1.1 GHz model in August 2000. Neither of these milestone chips made it into our five best AMD CPUs of all time feature, though.

It also managed to get industry analyst quotes comparing the 1 GHz Athlon launch to man’s first steps on the moon, the breaking of the four-minute-mile athletics record, and the conquering of Everest.

AMD Athlon CPU for Slot A. A 650 MHz sample.

(Image credit: Tullius)

Enough of the marketing bombast, what about the AMD Athlon 1 GHz specs? The first AMD Athlon processors would debut in June 1999. Over their production history, they would progress from 500 MHz to 1.4 GHz, FSB speeds from 100 to 133 MHz, and tech nodes from 250 nm to 180 nm. These K7 chips would also be made available in Slot A, Socket A, and Socket 563 platforms.

The specific 1 GHz barrier-breaking chip is thankfully cataloged by TechPowerUp. From the site's database, we can see it was a Slot A model with a bundled cooler, produced on the 180 nm process and packing 22 million transistors. Its clock speed was the magic 1,000 MHz, and it had a base clock of 100 MHz with a 10.0x multiplier. It drew 1.8V for a TDP of 65W. Of course, it was a single-core processor, before the days of hyperthreading, and came with a 128KB L1 Cache, and 512KB L2 cache. The first gigahertz Athlon's tray price at launch was $1,299.

Intel caught off guard

Intel was caught with its pants down by the AMD 1 GHz processor shipment announcement. The iconic PC chipmaker had been boasting about its breaking of the Gigahertz barrier for over a year, citing public demos of the 0.25 micron Pentium III processor pushing beyond this milestone.

AMD’s shipping announcement prompted Intel to paper launch its 1 GHz Pentium III chip (Tray price $990) two days later. However, it was plagued by supply issues for months. Contemporary reports suggest Intel planned to ramp volume in Q3 2000, which would give AMD quite a lot of time to make merry with its 1 GHz Athlon.

Google Preferred Source

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Mark Tyson
Mark Tyson
News Editor

Mark Tyson is a news editor at Tom's Hardware. He enjoys covering the full breadth of PC tech; from business and semiconductor design to products approaching the edge of reason.

17 Comments Comment from the forums
  • TechieTwo
    A very good day for AMD and consumers. Intel was stunned. History has repeated itself again in recent times and it's all good for consumers.
    Reply
  • Vanderlindemedia
    The beauty of these things where the overclocking options. Most known was the GFD, a Golden Finger Device.

    https://www.heise.de/select/ct/2019/27/1572616032266062/contentimages/ct2719AthlonOve_103836-chh-AthlonOver_nostA.jpg
    You had to crack open your casing in order to be able to install that thing onto the CPU board, no soldering or anything required, but after installation, you had a free set of multipliers to choose from including voltages.

    What was even better, where the often 500Mhz models or higher, simply rebranded 750Mhz chips. What it means was under the hood it was a downclocked 750Mhz model which was cheaper for AMD to produce.

    This gave so much variety in overclocking, that even the cheap boards supported FSB's could extract free performance. However, the whole Slot A never had any true multi-cpu support as the Intel's Slot did.
    Reply
  • EzzyB
    It was a big deal as far as marketing went. Intel could not get it's Pentium 4 to quite clock that high. This resulted in one of the most unusual CPU releases ever when, to get to 1Ghz, they released the Intel Tualatin processor. (Note that Tualatin was NOT Coppermine)

    It was technically a server chip, though the Xeon name hadn't been used at that time. I built a system around one that, I believe ran at 1.13 ghz and actually had hyperthreading. While it used the same socket as the P-III it needed a different chipset that enabled an additional pin in the socket.

    Strangely enough the first PC program that I used that was multi-thread aware was the Alpha/Beta test of Star Wars Galaxies that would use a second thread for terrain generation if it was available.

    Overall the chip ran quite well and compared to the Athlon and P-IV right up until you did something memory intensive (similar to Athlon) and then the higher bus/memory speeds of the P-IV would kick in and it would prevail in memory intensive stuff.
    Reply
  • Trake_17
    Fun article. Thanks for publishing it
    Reply
  • teckel12
    This was a big deal, I received a 1 GHz Athlon mouse pad as some advertising promotion. Meanwhile, I was running some slow 133 MHz Cyrix clone 6x86 (M1) CPU that constantly overheated trying to play Quake. But, I had the 1 GHz mouse pad.
    Reply
  • thestryker
    EzzyB said:
    This resulted in one of the most unusual CPU releases ever when, to get to 1Ghz, they released the Intel Tualatin processor. (Note that Tualatin was NOT Coppermine)
    Coppermine hit 1 GHz fine it was the 1.13 GHz model which didn't work right and had to be recalled. I believe they eventually released one or two above 1 GHz but it was largely irrelevant by that time.
    EzzyB said:
    It was technically a server chip, though the Xeon name hadn't been used at that time.
    There were multiple Tualatin CPUs and the server ones were named PIII-S. I have my dual PIII-S 1.4GHz system in storage as I have lots of fond memories.
    EzzyB said:
    actually had hyperthreading
    None had hyperthreading, but they did support dual socket configurations.
    EzzyB said:
    While it used the same socket as the P-III it needed a different chipset that enabled an additional pin in the socket.
    It wasn't an additional pin, but rather a shift of what the pin did. I want to say it was one of the power delivery pins holding a different value.
    EzzyB said:
    Overall the chip ran quite well and compared to the Athlon and P-IV right up until you did something memory intensive (similar to Athlon) and then the higher bus/memory speeds of the P-IV would kick in and it would prevail in memory intensive stuff.
    My system used a Via Apollo chipset with DDR support. While I picked it because I was using dual CPUs and they shared memory bandwidth it also had advantages on memory intensive software.
    Reply
  • warezme
    OMG, I had totally forgotten Slot A! Dang blast from the past. I remember having one of those.
    Reply
  • 80251
    DEC's Alpha was the first Ghz. CPU and it predated anything from Intel and/or AMD.

    RIP DEC.
    Reply
  • Thunder64
    thestryker said:
    Coppermine hit 1 GHz fine it was the 1.13 GHz model which didn't work right and had to be recalled. I believe they eventually released one or two above 1 GHz but it was largely irrelevant by that time.

    There were multiple Tualatin CPUs and the server ones were named PIII-S. I have my dual PIII-S 1.4GHz system in storage as I have lots of fond memories.

    None had hyperthreading, but they did support dual socket configurations.

    It wasn't an additional pin, but rather a shift of what the pin did. I want to say it was one of the power delivery pins holding a different value.

    My system used a Via Apollo chipset with DDR support. While I picked it because I was using dual CPUs and they shared memory bandwidth it also had advantages on memory intensive software.

    You nailed just about everything I was going to say. I would only like to add that the Athlon/P-III didn't "choke" because of lack of memory bandwidth. The P4 wasn't really a bandwidth monster until the P4C anyway (excluding horrible RDRAM). The P4 simply needed that bandwidth to keep its long pipeline full especially after mispredicts.

    What the P4 did do very well eventually was SSE2 once software caught up. Athlon/P-III lacked it and it was a huge step forward from 3DNow!/SSE in many areas. Media transcoding comes to mind.
    Reply
  • razor512
    Due to pricing back then, I had to settle for an anemic AMD Duron CPU back then.

    The design did pretty well for bare die cooling.

    https://i.imgur.com/m0h5OVW.jpeg
    Reply