Comment by joecool1029 6 days ago
> There is nothing wrong with getting the size of the market wrong by that much - it happens too often for anyone to call it wrong. It isn't clear what was predicted, but marketing should have predicted a range of units sold (and various price points having different predicted ranges!).
Shout out to the Itanium sales forecast: https://upload.wikimedia.org/wikipedia/commons/8/88/Itanium_...
If they keep predicting that, eventually they’ll be right!
(It’s hard to harvest more power from a star than a Dyson sphere is capable of)
Reminds of something I heard: Of the 3 most recent recessions, analysts predicted 20.
Very soon we will produce more solar electricity than all of the word's consumption. A "problem" that is even more severe than it looks like, because we consume energy when the Sun is under the horizon too.
So, yeah, in a few years they'll be right. Even if for just a short time while the rest of the economy grows to keep up with the change.
Those 2 charts are amazing! At least the Itanium people adjusted their curves downward over time, looks like the IEA just carried on regardless!
It wasn't the Itanium people so much as the industry analysts who follow such things. And, yes, they (including myself) were spectacularly wrong early on but, hey, it was Intel after all and an AMD alternative wasn't even a blip on the radar and 64-bit chips were clearly needed. I'm not sure there was any industry analyst--and I probably bailed earlier than most--who was going this is going to be a flop from the earliest days.
an AMD alternative wasn't even a blip on the radar
It’s understandable why companies try and sometimes succeed at creating a reality distortion field about the future success of their products. Management is asking Wall Street to allow them to make this huge investment (in their own salaries and R&D empire), and they need to promise a corresponding huge return. Wall Street always opportunities to jack up profits in the short term, and management needs to tell a compelling story about ROI that is a few years in the future to convince them it’s worth waiting. Intel also wanted to encourage adoption by OEMs and software companies, and making them think that they need to support Itanium soon could have been a necessary condition to make that a reality.
I don’t know what factors would make IEA underestimate solar adoption.
There are trillions of dollars on the line in convincing people not to buy solar panels or other renewable sources.
Remember all the conspiracy theories about how someone invented a free energy machine and the government had to cover it up? Well they're actually true - with the caveat that the free energy machine only works in direct sunlight.
How often are they reality distortion fields vs leadership trying to put on a face to rally the troops and investors? How do you do the second without the first?
Something I ponder from time to time, while trying to figure out how to be less of a cynic and more of a leader.
> Management is asking Wall Street to allow them to make this huge investment (in their own salaries and R&D empire), and they need to promise a corresponding huge return. Wall Street always opportunities to jack up profits in the short term, and management needs to tell a compelling story about ROI that is a few years in the future to convince them it’s worth waiting
Explain Amazon, Uber, Spotify, Tesla, and other publicly listed businesses that had low or even negative profit margins for many years.
The idea that Wall Street only rewards short term profit margins is laughable considering who is at the top of the market cap rankings.
The section of my comment you quoted directly addresses this! Wall Street can be convinced by a compelling story.
Holy cow was that forecast bad!
It reminds me of a meeting long ago where the marketing team reported that oil was going to hit $400/bbl and that this would be great for business. I literally laughed out loud. At that price, gasoline would be about $18/gal and no one could afford to move anything except by ox cart.
> At that price, gasoline would be about $18/gal and no one could afford to move anything except by ox cart.
Just for some rough math here - I’m currently paying around $1.20/L for gas, and crude oil cost is roughly half of that, so if crude went up by 6x, I’d be looking at $5/L for gas. Gas is currently about 20% of my per-km cost of driving, so that price increase at the pump would increase my per-km cost by about 60%.
FWIW that’s roughly the same per-km cost increase that people have voluntarily taken on over the past decade in North America by buying more expensive cars.
(Though this does apply to personal transportation only, the math on e.g. transport trucks is different)
The issue isn't person transport it is shipping and home heating and agriculture
I drive electric so like to imagine myself sheltered from gas price increases but I know grocery costs would explode
Especially if you live were gas cost a buck twenty a liter
How can you possibly say that crude is half of the pump price? The economics are incredibly complex and murky, and the price of gas doesn't move with any sort of linear relation to crude except in very long timeframes. Regional refining capacity is way more important.
The price of gas isn't immediately and directly impacted by the price of crude because of futures contracts. This naturally means gas prices will move to match the price of crude over time. It's a feature of the current system, not an indication that the price of gas isn't heavily reliant on gas. Nobody is making gas with spot prices.
> How can you possibly say that crude is half of the pump price?
I googled for a couple sources on the breakdown of the price of gasoline, and they seemed to be in agreement that the raw cost of crude is somewhere around half. (And broke refining out separately.)
I'm sure it's not perfect, but it seems fairly reasonable. (And it can be off by quite a lot and still not make a huge difference to the cost-per-km of driving.)
> How can you possibly say that crude is half of the pump price?
Look at gas prices in your area. Look at the price of crude. Divide.
How could you possibly not be able to estimate the fraction?
And yeah ideally you use an average number over some months and you sample the crude earlier than the gas but those are minor tweaks.
Itanium needs a lot longer discussion than can be covered in an HN comment.
https://bitmason.blogspot.com/2024/02/the-sinking-of-itanic-...
I think Bob Colwell's account is the clearest short synopsis.
https://www.sigmicro.org/media/oralhistories/colwell.pdf
'And I finally put my hand up and said I just could not see how you're proposing to get to those kind of performance levels. And he said well we've got a simulation, and I thought Ah, ok. That shut me up for a little bit, but then something occurred to me and I interrupted him again. I said, wait I am sorry to derail this meeting. But how would you use a simulator if you don't have a compiler? He said, well that's true we don't have a compiler yet, so I hand assembled my simulations. I asked "How did you do thousands of line of code that way?" He said “No, I did 30 lines of code”. Flabbergasted, I said, "You're predicting the entire future of this architecture on 30 lines of hand generated code?" [chuckle], I said it just like that, I did not mean to be insulting but I was just thunderstruck. Andy Grove piped up and said "we are not here right now to reconsider the future of this effort, so let’s move on".'
I’m curious what kind of code his 30 lines were - I’m betting something FP-heavy based on the public focus benchmarks gave thst over branchy business logic. I still remember getting the pitch that you had to buy Intel’s compilers to get decent performance. I worked at a software vendor and later a computational research lab, and both times that torpedoed any interest in buying hardware because it boiled down to paying a couple of times more upfront and hoping you could optimize at least the equivalent gain back … or just buy an off-the-shelf system which performed well now and do literally anything else with your life.
One really interesting related angle is the rise of open source software in business IT which was happening contemporaneously. X86 compatibility mattered so much back then because people had tons of code they couldn’t easily modify whereas later switches like Apple’s PPC-x86 or x86-ARM and Microsoft’s recent ARM attempts seem to be a lot smoother because almost everyone is relying on many of the same open source libraries and compilers. I think Itanium would still have struggled to realize much of its peak performance but at least you wouldn’t have had so many frictional costs simply getting code to run correctly.
I think you're right. The combination of open source and public clouds has really tended to reduce the dominance of specific hardware/software ecosystems, especially Wintel. Especially with the decline of CMOS process scaling as a performance lever, I expect that we'll see more heterogeneous computing in the future.
This form versus substance issue is a really deeply embedded problem in our industry, and it is getting worse.
Time and again, I run into professionals who claim X, only to find out that the assertion was based only upon the flimsiest interpretation of what it took to accomplish the assertion. If I had to be less charitable, then I’d say fraudulent interpretations.
Promo Packet Princesses are especially prone to getting caught out doing this. And as the above story illustrates, you better catch and tear down these “interpretations” as the risks to the enterprise they are, well before they obtain visible executive sponsorship, or the political waters gets choppy.
IMHE, if you catch these in time, then estimate the risk along with a solution, it usually defuses them and “prices” their proposals more at a “market clearing rate” of the actual risk. They’re usually hoping to pass the hot potato to the poor suckers forced to handle sustaining work streams on their “brilliant vision” before anyone notices the emperor has no clothes.
I’d love to hear others’ experiences around this and how they defused the risk time bombs.
> “You're predicting the entire future of this architecture on 30 lines of hand generated code?"
It’s comforting to know that massively strategic decisions based on very little information that may not even be correct are made in other organizations and not just mine.
Everybody does it. Information only comes because you made your strategic decision, never before it.
I don’t think it is that simple. Itanium was for years supported for example by RHEL (including GCC working of course, if anybody cared enough they could invest into optimising that), it is not like the whole fiasco happened in one moment. No, Itanium was genuinely a bad design, which never got fixed, because it apparently couldn’t be.
Well, yes, the market didn't care all that much for various reasons. (There were reasons beyond technology.) RHEL/GCC supported but, while I wasn't there at the time, I'm not sure how much focus there was. Other companies were hedging their bets on Itanium at the time--e.g. Project Monterey. Aside from Sun, most of the majors were placing Itanium bets to some degree if only to hedge other projects.
Even HP dropped it eventually. And the former CEO of Intel (who was CTO during much of the time Itanium was active) said in a trade press interview that he wished they had just done a more enterprisey Xeon--which happened eventually anyway.
We're not living through this again at all with generative AI, right?
The plan was to artificially suppress x86-64 to leave customers with no real alternative to Itanium. The early sales projections made sense under that assumption.
I had heard that it wasn't suppression as much as just not making it a thing at all, and that AMD used the opportunity to extend x86 to 64-bit, and Intel was essentially forced to follow suit to avoid losing more of the market. It also explains why the shorthand "amd64" is used; Intel didn't actually design x86_64 itself.
There was apparently earlier Pentium 4s that supported some version of a 64bit isa, support for which was fused off before sending to customers in order to convince people to move to Itanium.
https://www.tomshardware.com/pc-components/cpus/former-intel...
I've still got a couple small business models along these lines that are over 20 years old now. Still running possibly because I always turn them fully off when not using them. No hibernation, sleep or other monkey business.
One Dell has an early 64-bit mainboard but only a 32-bit CPU in that socket, just fine for Windows XP and will also run W10 32-bit (slowly), mainly dual booting to Debian i386 now since it retired from office work. Puts out so much heat I would imagine there is a lot of bypassed silicon on the chip drawing power but not helping process. IIRC a 64-bit CPU for that socket was known to exist but was more or less "unobtanium".
Then a trusty HP tower with the Pentium D, which was supposedly a "double" with two x86 arch patterns on the same chip. This one runs everything x86 or AMD64, up until W11 24H2 where the roadblocks are unsurmountable.
To this day, I don't know if Intel thought Itanium was the legitimately better approach. There were certainly theoretical arguments for VLIW over carrying CISC forward--even if it had never been commercially successful in the past. But I at least suspect that getting away from x86 licensing entanglements was also a factor. I suspect it was a bit of both and different people at the company probably had different perspectives.
Internal inertia is a powerful thing. This was discussed at length on comp.arch in the late 1990's early 2000's by insiders like Andy Glew. When OoO started to dominate intel should have realized the risk, but they continued to cancel internal projects to extend x86 to 64-bits. Of which apparently there were multiple. Even then, the day that AMD announced 64-bit extensions and a product timeline it should have resulted in intel doing an internal about face and acknowledging what everyone knew (in the late 1990's) and quietly scuttling ia64 while pulling a backup x86 out of their pocket. But since they had killed them all, they were forced to scramble to follow AMD.
Intel has plenty of engineering talent, if the bean counters, politicians and board would just get out of the way they would come back. But instead you see patently stupid/poor execution like then still ongoing avx512 saga. Lakefield, is a prime example of WTFism showing up publicly. The lack of internal leadership is written as loud as possible on a product where no one had the political power to force the smaller core to emulate avx512 during the development cycle, or NAK a product where the two cores couldn't even execute the same instructions. Its an engineering POC probably being shopped to apple or someone else considering an arm big.little without understanding how to actually implement it in a meaningful way. Compared with the AMD approach which seems to even best the arm big.little by simply using the same cores process optimized differently to the same effect without having to deal with the problems of optimizing software for two different microarch.
Sophie Wilson (ARM instruction set designer) was very enthusiastic over her "Firepath" architecture that had VLIW aspects.
It was targeted at DSL modems, and I think the platform has faded and is now somewhat obscure.
https://royalsociety.org/people/sophie-wilson-12544/
https://old.hotchips.org/wp-content/uploads/hc_archives/hc14...
Intel and AMD have a cross licensing agreement where they pay each other the right to use various IP. One of the things Intel pays AMD for is x86_64.
And its inverse, the IEA solar energy forecast: https://en.wikipedia.org/wiki/File:Reality_versus_IEA_predic...
(This version of the graph is pretty old, but it's enough to get the flavor. The rate of new installations is still increasing exponentially, and the IEA continues to predict that it'll level off any day now...)