AMD – Everything you need to know – Catch Up!
AMD, the underdog in both the CPU and GPU markets, has been around for a long while now with an impressive history, and some amazing products, so in this video I’m going to catch you up on everything you need to know about AMD! But first, if you haven’t already, consider subscribing for more videos like this one every monday, wednesday and friday!
Like I said, AMD is hardly new – it was founded in 1969 – nice – by Jerry Sanders, who left the – at the time – the industrial giant Fairchild Semiconductor, who was frustrated with the management of the firm, and decided to poach 7 colleagues on his way out, to set up AMD. He was actually following the footsteps of Robert Noyce and Gordon Moore, the founders of Intel who had left the year before. Seems like Fairchild really shot themselves in the foot by losing Noyce, Moore and Sanders, as they are worth roughly $300 billion today. Ouch.
Now, AMD didn’t start out with their own CPUs, no Zen Architectures for us for another 50 years, see this was back in the early days of computing, when everything was shift registers and logic gates. Intel wouldn’t even make their first CPU until late 1971 with the 4004, so AMD’s strategy was to start as a secondary supplier for their old boss, Fairchild, as well as the other big name at the time, National Semiconductor. They specifically manufactured to US Military Standards to aid the apparently significant reliability issues that were incredibly common across this bleeding edge tech.
They then made their first product late in the same year, a 4 bit shift register, a kind of memory that can be configured in different ways, but generally a single serial input, to parallel outputs, so if you have a 4 bit string, say 1001, you can send 1, then 0, 0, then 1 on the input line, and each “flip flop”, i know, great name, stores each bit then can output it all at once from it’s now 4 output lanes.
They then branched out to different types of logic, counters, multipliers and the likes, and even manufactured their own RAM, or random access memory. RAM is special, as it’s pretty much always what’s called volatile, meaning as soon as the power goes out, all the data that was saved is lost, as it requires power to maintain the state in the logic, unlike, say SSDs today, which are called non-volatile, as you actively write to them.
AMD went public in late 1972, and even became a secondary supplier to Intel for some other their shift registers, but their business with Intel didn’t stop there. AMD cloned Intel’s 8080 CPU as the AM9080 and when Intel started using Microcode – a sort of in between layer between the programmer’s instructions and the actual instructions the CPU can execute – they both signed a cross licensing agreement in ‘76 so AMD could use both it, and their CPU designs.
That partnership became very useful for Intel when they started manufacturing x86 processors for IBM’s new “Personal Computer”, as IBM stipulated Intel must have a second source manufacturer for it’s CPUs if it was going to be their supplier. Intel signed a technology exchange agreement with AMD in 1981, a 10 year agreement that meant either company could become a second source manufacturer for the other’s products, and therefore would share designs for their products, namely the Intel 8086, 80186 and 80286 chips.
The partnership was beneficial for both, enough so that they extended the agreement to 1995, with an important addition, that either company would use arbitration should they disagree. Arbitration is the process where an independent 3rd party will decide the outcome of any disagreement, but it’s important to note that’s a private company, not the legal system.
The reason that’s important, is in 1986 Intel broke the agreement stopping AMD from manufacturing their x86 CPUs, and AMD filed for arbitration to resolve the issue. The arbitrator decided in AMD’s favour, but Intel wasn’t happy about that, and ended up going to the California supreme court in 1994 who upheld the decision.
And, that wasn’t the only time Intel tried to interfere with AMD, in 2005, thanks to the Japanese Trade Commision investigating, Intel was found to use secret rebates, discounts and threats to lock AMD out of the market. In 2009 Intel agreed to pay AMD $1.25bn and renew a 5 year cross licensing agreement to settle all outstanding legal battles.
That agreement still stands today, especially thanks to AMD’s development of x86-64, the 64 bit counterpart to Intel’s 32 bit x86 instruction set. Intel did try to make their own version in their Itanium architecture chips, but wasn’t compatible with existing x86 softwares so while it was the 4th most deployed architecture in 2008, that didn’t mean much as AMD’s x86-64 was much, much more popular, and is finally dead as of 2021.
Going back to their products, beyond the Intel cloned chips, they started making their own in 1996, with their K5, believe it or not, named after Kryptonite from Superman, as a jab at Intel’s market dominance, suggesting their new chip would be, well, the kryptonite to Intel’s super-powers. A few years later came the K6, then brand names they’ve actually carried on to current day, like Athlon. This is when the type of socket your CPU went into changed too, split between the brands. Intel used Slot 1 – add in cards that carried your processor and heatsink onboard. AMD used the remarkably similar “Slot A” which used the same physical connector meaning motherboard manufacturers could easily adapt to AMD’s chips if they wanted to.
AMD’s next set of chips, using the x86-64 instruction set too came in the form of Opteron server CPUs, and soon after Athlon 64 chips for the PC market. It didn’t take long for AMD to then release a dual core CPU, the Athlon X2. These were incredibly impressive for their time, and while up to 2.4GHz on 2 cores doesn’t sound like much, remember that this is what games looked like in 2005.
AMD ramped up the core counts over the next 4 years, up to quad core with it’s Athlon X4s, then launched their Phenom II series in 2009. They started as dual, triple and quad core parts, although they all used the same die with cores disabled for the triple and dual core chips, meaning you could sometimes get lucky and unlock the cores in the bios to get free performance – although they later started laser cutting dies so that practice stopped pretty quick.
In 2010, they launched a totally new part for the Phenom II lineup, the x6 – a 6 core with “AMD Turbo Core” technology, which basically meant it would run on 3 cores most of the time, but spin up the other 3 when more performance was needed. Similar to the frequency boosting we have today.
Sadly, even with the extra cores, AMD wasn’t overly competitive with Intel’s offerings at the time, like the i7-980 and the year after the infamous 2600k and 2700k chips. AMD tried to come back in late 2011 with their Bulldozer Architecture – my first gaming CPU believe it or not, and the first CPU I ever reviewed, badly. Not much changed then.
This is what you might call the “bad times” for AMD, as their Bulldozer, Steamroller, and Piledriver Architectures, well, kinda sucked. While they did beat Intel to the punch with an 8 core, it wasn’t actually a “real” 8 core, as it only had 4 floating point units and shared cache, this was actually the topic of a lawsuit against AMD they settled just last year to the tune of $12.5 million.
AMD generally dwindled in this time, with their stock price reaching a record low to below $2 per share in 2015. They managed to hold onto some value by getting themselves into both the Xbox One and Playstation 4 consoles, and are confirmed to be in the next generation Xbox Series X and Playstation 5 consoles too.
Happily though, they really pulled themselves out of the mud, because in 2017 they launched their Ryzen series of processors to impressive success. They opened with up to 8 (real) cores, and impressive benchmark results that really surprised Intel, who had been sitting pretty for the better part of a decade at a paltry 4 cores. With impressive performance, although not quite a match as single threaded workloads, they were an amazing value for both gaming and content creation.
To bring us to present, their 3rd generation Ryzen CPUs have much improved single threaded performance, closely rivaling Intel’s options, and are still an amazing value – even up to the 64 core Threadripper chip too.
So, that’s AMDs CPUs, but I think we are missing a key part of the picture here, their graphics cards. That story is a lot shorter, as they just out and out bought a company called ATI in 2006 for a total of around $5.4 billion. ATI was already a fairly successful GPU maker, and initially remained the brand name until 2010 when AMD became the favoured name – along with their brand Radeon. Their market share generally declined over the years, as AMD cards were often plagued with issues from excessive power draw in their R9 290X cards, to driver issues – including very recently too.
Their recent popular cards include the RX 400 and 500 series, like the RX 480, and more recent ones in the RX 5700XT too, which offer decent performance for their price, but currently lack any top-end options to compete with their main rival in that space, NVIDIA. It’s not looking as promising for AMD’s GPU division, by comparison.
So, that’s a brief history of AMD and their products. Of course, there is so much detail I haven’t been able to cover, so if there is something of particular interest to you leave it in the comments below so other people can learn!
