So, back in the 1990s and early 2000s, there was a lot of lore. So, at the time AMD was not an underdog.. it was just dogshit(i mean not technically, a lot of their chips were still good for budget stuff, but like tbh if you were building a PC, Intel was the default unless you know what you are doing) . Which is crazy when you realized AMD is like the default for so many people now. At that time, if you wanted a CPU , Intel was the go to. Now, at that time computers were a relatively new concept and most of the people did not know literally anything about them. So to make it easier for the consumers companies did something STUPID AS HELL. They were like, monkey know number, monkey know big number = better, so we will have the big number a GHz. In all fairness, yes if you were a consumer at that time, it would've looked like the objective truth because the clock speeds back then were barely 1Ghz.
But then AMD shows up with Athlon , the first consumer CPU that could push 1GHz, to picture it, your mathematical genius could solve a problem in like what 2 seconds if you took advanced mental math classes, for basic athematic problems like the 985 time 954 which barely utilized the ALU and was solved in 1 clock tick, took the CPU literally 1 nano second WHICH IS ONE BILLIONTH OF A SECOND. This was really humiliating because AMD, the company that people literally didn't care about won the GHz race.
This literally jerked Intel's ego so bad that they had to rush within 48 hours and launched the Pentium III, The first intel chip to reach 1GHz.
Here is the twist, while Athlon was a legit product, Intel Pentium III was just a paper-launch. Let us take our drive thru example back, if you were a nerd back then and wanted the top of the line cheese(1Ghz processor) you could just walk up to a drive thru, get one in your backseat (The Athlon 1) and then go home. But Intel's processor was off the shelves for months and only a few selected people got their hands on samples while Athlon was production ready.
So, after being obliterated by an underdog the corporate egos literally came up to the mic and said burn everything, this led to a literal disaster like actually a disaster and Intel is still trolled for it. Their ego was so fat, they went up to a press meet with thousands of press, analysts, engineers, and the whole nerd squad and said listen fellas, we are reaching 10 Ghz, yes we are trust us. Now DO YOU REALIZE HOW RIDICULOUS IT IS. Lets put it this way, The Pentium 4 launches at 1.4/1.5 Ghz, but with an insanely low IPC that it got DESTROYED by the rushed Pentium 3 that boosted to 1Ghz but had reasonable IPC in many real world tasks. So you could picture how bad their egos were punctured that you bet a failing architecture that loses to it's previous generation, could boos to 10Ghz, fellas 10 Giga Hertz. NOW They did technically say 7-11Ghz but hey who doesn't like some drama. Just for the matter of fact, the Intel Core i9-14900KS which is the chip that has the highest factory boost clock gets 6.2Ghz at maximum which is 0.8Ghz less that what was promised 2 DECADED earlier. Crazy Stuff Man. Crazy Stuff.
And then project Tejas which btw is the code name for the new architecture that was supposedly going to reach the 7-11Ghz target, was basically scrapped. But here is the interesting lore that is actually pretty inspiring at least to me. So let us look at the story of a small Israeli Team and a CTO who cared.
Before we talk about the team I want to give a little backstory. So, back in (very old time), there was a prodigy who joined Intel at 18. Mind you, 18 is when most people enter college, while here Pat Gelsinger joined a multi-billion dollar semiconductor company at 18! Then he literally became the first ever CTO (Chief Technology Officer) of Intel, in 2001. Now although you might think he got plot armor, there was this one claim that we discussed earlier. He predicted the 7-11GHz trap, along with the entire Intel HQ in Santa Clara at that time.While the Intel Headquarters in Silicon Valley were busy working with their NetBurst Architecture that supposedly was to achieve the double digit GHz number, there was this ultra small team of around 30 engineers in Intel Israel and no one gave a flying shit about them. That was until they made their idea a reality. So, their approach to making faster chips was that why are we making the chips faster by trying to make them do work faster. This is stupid. What if we made them do more work every time they worked. That was over simplified but get the concept, imagine our CPU like a clock. Every time it ticks it does some work. If it ticks it does the work wrong, it drops it and does it again. The whole industry at that time thought if we just made it tick faster it would be faster overall. Here is the issue. Voltage scales roughly with frequency(the number of times the clock actually ticks). So making it faster would induce so much voltage that the math was scary. We know if you induce current you will induce heat. The Wattage / cm2 value, basically the heat density, got so high that the CPU got hotter than a nuclear reactor's surface! But even if we somehow got it to tick that fast then there is another massive problem. So when we make it do stuff "faster" we would have to break the work into multiple small pieces right, like obviously, if you are doing work fast then you would want it simpler because the more complex the problem the higher it will take for it to solve, the major issue arises when the CPU does it wrong. When the CPU does a problem wrong it has to clear all those work and then re do all of it. This took time, and in real world tasks it caused massive stuttering, imagine playing CS:GO, your doing a 180 aimed perfectly and boom your game just lags and you are back to lobby. So these Engineers looked at this problem and their Israeli genius brains were asking the right questions. The questions that changed the whole CPU MARKET. Why can't we make chips do more work every trip than making it tick faster? The answer is you absolutely can. That led to this little laptop chip called Banias that sipped power, was 1.6Ghz(which was crazy in the worst way possible back during the Great Gigahertz War Era), and somehow outperformed a Pentium 4 that consumed a ton of power and ran at double the clock speed. Now, imagine being at California in the HQ of INTEL! You are working for NetBurst the next big thing and your flagship is getting obliterated by a laptop chip while running at half the GHz like it absolutely did not matter(SPOILERRR) and sipping power and fitting in a battery. But when the whole HQ was like suppress them or force them into using NetBurst, Pat saw something. So he went to the CEO and literally told him "NetBurst is failing, if we do not listen to those engineers at Israel then our company is cooked, AMD would throw us off our freaking chairs and we wont be relevant anymore" now obv he didn't say it verbatim but you get the vibe. And because it was Pat himself who told it, the CEO listened and was like "Alright boys." and then : Killed NetBurst, made Banias an experiment into a legit product the Pentium - M Laptop chip, and then GAVE THE TEAM OF 30 HQ BUDGET AND A DIRECT ORDER TO MAKE DESKTOP AND SERVER CHIPS. Now that right there is crazy. So the Israeli team took the offer, made the legendary Intel Core Duo. THAT THING WAS PEAK. They added another core, did a lot of nerdy stuff, and made the CONROE architecture and literally made a chip that competed head to head with an AMD Athlon 64 and had a slower clock speed and sipped less power. The industry collectively went "WHAT THE ACTUAL -", all their assumptions for a whole decade just got proved wrong by 30 people with an idea. But here is the part that hurts, even though these smart asses have done so much that any modern x86 or x64 chip we use was depended on their question, you can't find their name on a single box, or a single demo. That is corporate reality man, "Intel designed the Core Duo" and that wraps the story...Now I really think I should speak more about Pat Gelsinger. So after the massive success of Intel Core Duo, you would expect Pat to be the next CEO, I mean he was the reason the team even got the funding in the first place, right, wait Intel... RIGHT....... So this is one of the most notable mess-ups of the Board of Directors. In 2009, when the Board Room decided a new CEO, It was not Pat, or any engineer instead it was a finance dude. This was so messed up that A. Pat left Intel, joined VMware, got the best CEO award in the United States in 2019 by Glassdoor, mind you 99% , aka out of 100 random employees 99 of them loved Pat Gelsinger B. Intel absolutely messed up their strategy and Ryzen was taking over so Intel asked Pat Gelsinger to come in as the new CEO and history repeated itself, 2021 Pat joined, 12th gen somehow survived and boom 13th gen was another mess up and 14th GEN IS REGARDED AS THE WORST CPU LAUNCH IN HALF A CENTURY! And the Board of Directors this time "forced" Gelsinger to stepdown aka corporate way of saying they fired him. Yeah tragic but hey peak entertainment.
If you read thru this, consider leaving a comment. Thanks.
- Sam