Why tecnology advances at this pace?

[IIIl]

Is it about resources? Why couldn't  we produce processors like today's, a few years ago for example?

[r2724r16]

[Arika]

research and development.

 

why aren't we already living on mars when we could be in x number of years?

[GoldenLag]

Research

Research

Money

Money

Competition

Power consumption

 

 

 

Thats roughly the reasons why

[Mbowen]

It's about profit. Look at intel's i7-2600k. Came out in 2011 and was easily able to hit 5GHz. It wasn't until recently that you saw that happening regularly in today's market. It wouldn't be profitable to make a HUGE leap forward and then release it all at once. If you were to say make a discovery that allowed you to clock your next chip at 7GHz easily, you could also easily kneecap that chip and only release a 30% increase in performance for the next 4-6 years until you hit that limit. You'd make more money and in the long run have more time to look into developing technology. 

Plus, we're also running into manufacturing constraints because of the conductivity of the material we're using. That's why we're going smaller and smaller. We're limited by the speed of conduction now too. That's why people are talking about... what was it? Graphite? Cobalt? It's been a while and I can't be bothered to look. Just know that there are people looking to discover the next big jump in performance through use of a different medium when manufacturing CPUs. 

Good question though. Funny how money and marketing often stifle innovation. 

[mariushm]

12 minutes ago, IIIl said:

Is it about resources? Why couldn't  we produce processors like today's, a few years ago for example?

It takes time to research new ways of producing processors, study different chemical formulas for substances used in fabrication, improve optics (used in the manufacturing, lasers are focused through lens on the silicon wafers and create patterns, transistors and stuff), improve existing lasers or create new lasers that work with the smaller structures and with the new optics, also software has to be developed to work with the new stuff invented and so on...

 

As the manufacturing processes shrink, some things react slightly different compared to previous processes, so engineers have to learn a lot of new things and adapt designs to account for these changes.

Right now we're at 5nm..7nm .... 3-5 years from now you could say "why can we make 2nm processors today but couldn't make 3-5 years ago?"

Well, right now we know some of the current chemicals can't be used for 2nm because it's just too small of a process for those substances to continue to work as they work now. Engineers are studying other chemicals and mixes of chemicals that promise better results compared to current ones... so in a few years the engineers may figure out the most optimum mix of substances to produce processors with a high degree of repeatability, quality, consistency and so on..

 

Also companies are for profit.... there's a lot of money invested in research and other things... if there's no competitor (like it was the case with Intel and AMD back when AMD only had the FX series), a company like Intel prefers to slow down a bit and stretch things out and make profit for their investors and recover some of those research money

 

 

[pizapower]

Peeps don't want to pay higher taxes/prices.
No money = No development

[Jurrunio]

5 minutes ago, Mbowen said:

It's about profit. Look at intel's i7-2600k. Came out in 2011 and was easily able to hit 5GHz. It wasn't until recently that you saw that happening regularly in today's market. It wouldn't be profitable to make a HUGE leap forward and then release it all at once. If you were to say make a discovery that allowed you to clock your next chip at 7GHz easily, you could also easily kneecap that chip and only release a 30% increase in performance for the next 4-6 years until you hit that limit. You'd make more money and in the long run have more time to look into developing technology. 

Yeah IPC's so worthless

 

16 minutes ago, IIIl said:

Is it about resources? Why couldn't  we produce processors like today's, a few years ago for example?

Why aren't babies allowed to drink alcohol when they can do it say, 20 years later? I think common sense should already answer the question.

[Lurick]

Why wasn't the polio vaccine invented years earlier?

Why didn't people just land on the moon faster?

Why does technology take time to progress and not just magically appear?

[Glorious]

9 hours ago, IIIl said:

Why couldn't  we produce processors like today's, a few years ago for example?

We do, the stuff they're working on now will come out usually in 5 years. 

[Trik'Stari]

9 hours ago, Arika S said:

research and development.

 

why aren't we already living on mars when we could be in x number of years?

Because some idiots would rather the governments spend money on handing out free shit, or military spending.

 

All I'm gonna say is:

 

1. Get a fucking job. Society owes you nothing unless you provide something for society.

And

2. There's oil (precious metals) in them thar asteroids.

[wasab]

9 hours ago, IIIl said:

Is it about resources? Why couldn't  we produce processors like today's, a few years ago for example?

Arm processor still seems to follow Morse law and so does gpu

[Donut417]

10 hours ago, IIIl said:

Is it about resources? Why couldn't  we produce processors like today's, a few years ago for example?

Well, in 2006 AMD bought ATI, which means they had to compete on CPU and GPU's. The issue is the company hit a rough patch. For like a decade AMD was not competitive. When there is no competition there is no innovation. Which is why Intel was complacent for sooooooo long. Now that AMD has chosen to dig it self out of the hole it was in, at least on the CPU side, they now force their competitor to be innovative again.

[mariushm]

11 hours ago, Donut417 said:

Well, in 2006 AMD bought ATI, which means they had to compete on CPU and GPU's. The issue is the company hit a rough patch. For like a decade AMD was not competitive. When there is no competition there is no innovation. Which is why Intel was complacent for sooooooo long. Now that AMD has chosen to dig it self out of the hole it was in, at least on the CPU side, they now force their competitor to be innovative again. 

It wasn't quite as simple as that.

They invested a lot of money in the purchase of ATI but then they had to sell their chip making plants so they had to sell what's now Global Foundries ... but the only way they could sell it was by promising that for a number of years they will make a minimum number of chips (well wafers, the company didn't care what AMD cuts out of the wafers), if not AMD would have paid huge fines and lose even more money.

So even when they realized the FX series was not as great as they thought, they still had to use that capacity at Global Foundries and make processors that didn't sell well. 

At the same time, people that bought Global Foundries from AMD promised them they'd make improvements and go faster from that 32nm and 28nm processes to something better, but just like Intel is screwing up now with the 10nm manufacturing process, Global Foundries had failures after failures in upgrading their lines - a few years ago they simply gave up in trying to make their own... they gave up and bought the 14nm process from Samsung (they licensed it)

AMD was hoping they use the wafers they have to buy anyway to produce video card chips at Global Foundries and designed chips for something like 20nm or what Global Foundries hoped to achieve and then they ended up having to spend millions redesigning chips again to be made on TSMC's lines and pay more to produce video cards, because GF couldn't get their shit together.

 

So AMD had to split between TSMC for their video card chips and some processors and using Global Foundries for a lot of CPUs just to consume their allotment of wafers and not be sued by Global Foundries...  they had to delay some video cards they planned for 16nm or 22nm or something like that because Global Foundries was still stuck on 28nm when they promised a new manufacturing process years ago.

 

Another thing that people are forgetting is that AMD lost a lot of money that could have been used for research and for producing better chips because of Intel's non competitive behavior ... basically Intel had money to burn and bribed various companies to not use AMD processors at all, or severely limit the use... they conditioned marketing money and promotions on companies agreeing not to use more than some percentage of AMD cpus in their lineup (like 5-10% or less AMD cpus or they'd lose money)

 

AMD won recently a lawsuit against Intel and they were awarded around 1.25 billion dollars and have a bunch of cross patent deals out of it making it even easier for them to make CPUs and video cards... see https://www.extremetech.com/computing/184323-intel-stuck-with-1-45-billion-fine-in-europe-for-unfair-and-damaging-practices-against-amd

Oh and this fine is on top of another 1.25 billion Intel paid in 2009 after some FTC complaints..

 

But 1.25 billions while a big amount of money... it's hardly enough considering the long period of time they were constantly choked by Intel with their throwing the money around.

 

Quoting from the Reuters link in the page: https://www.reuters.com/article/us-eu-intel-ruling-factbox/factbox-key-facts-on-eu-commissions-ruling-on-intel-idUSTRE54C3II20090513

Quote

EU FINDINGS:

* Intel gave wholly or partially hidden rebates to computer manufacturers on condition that they bought all, or almost all, their x86 CPUs processors from Intel.

* It also made direct payments to a major retailer on condition it stocked only computers with its x86 CPUs. This effectively prevented customers - and ultimately consumers - from choosing alternative products.

* Intel paid computer manufacturers to halt or delay the launch of specific products containing competitors’ x86 CPUs and to limit the sales channels available to these products.

* Computer manufacturers affected includes Acer, Dell, HP, Lenovo and NEC. The retailer concerned is Media Saturn Holding, owner of the MediaMarkt chain.

CONDITIONAL REBATES AND PAYMENTS

The Commission said Intel awarded major computer manufacturers rebates if they bought all or almost all of their supplies, at least in certain defined segments, from Intel:

* Intel gave rebates to computer manufacturer A from December 2002 to December 2005 conditional on this manufacturer purchasing exclusively Intel CPUs.

* Intel gave rebates to computer manufacturer B from November 2002 to May 2005 conditional on this manufacturer purchasing no less than 95 percent of its CPU needs for its business desktop computers from Intel (the remaining 5 percent that computer manufacturer B could purchase from rival chip maker AMD was then subject to further restrictive conditions).

* Intel gave rebates to computer manufacturer C from October 2002 to November 2005 conditional on this manufacturer purchasing no less than 80 percent of its CPU needs for its desktop and notebook computers from Intel.

* Intel gave rebates to computer manufacturer D in 2007 conditional on this manufacturer purchasing its CPU needs for its notebook computers exclusively from Intel.

PAYMENTS TO PREVENT SALES OF SPECIFIC RIVAL PRODUCTS

The Commission said it found the following specific cases:

* For the 5 percent of computer manufacturer B’s business that was not subject to the conditional rebate, Intel made further payments to computer manufacturer B provided that this manufacturer:

- sold AMD-based business desktops only to small and medium enterprises.

- sold AMD-based business desktops only via direct distribution channels (as opposed to through distributors)

- postponed the launch of its first AMD-based business desktop in Europe by 6 months.

* Intel made payments to computer manufacturer E provided that this manufacturer postponed the launch of an AMD-based notebook from September 2003 to January 2004.

* Before the conditional rebate to computer manufacturer D, Intel made payments to this manufacturer provided that it postponed the launch of AMD-based notebooks from September 2006 to the end of 2006.

 

 

[TetraSky]

No competition = no incentive to make a better product.

AMD didn't really compete for the better part of the last 10 years, until they launched their Zen lineup. (along with what @mariushm said above)

Hence why Intel didn't bother really making a CPU that was more than just a few % IPC better than the previous gen, why make the best thing you can, when you're alone at the top? That would cost a lot more in R&D to do so than simply increasing the IPC a little bit at a time...
Not to mention, they might be hit with "antitrust" bullshit for being "too good" so others "can't compete fairly".

Just like how Google is getting hit with them over their search engine, because their competition is f'ing garbage in comparison... Last I checked, Google wasn't forcing us all to use Google Search. We use it because it just works.

[Lathlaer]

Also some markets rely on annual releases and are fine with "good enough".

 

I imagine the R&D department for mobile CPU's and cameras is constantly working at 100% capacity. 

 

Then comes Samsung executive and goes "ok guys, what do you have for me" and they say "right now we have something that is 20% better than the last year's model but we are constantly working so if you come back in 6 months it will be 50% better"

 

But the exec is like "ok, we need a new Samsung phone every year so I will take this 20% now and return for the better one later for the next year model";.

 

Now, if Samsung were to release a new phone every 5 years, the boost would be obviously more substantial if they had been pumping that R&D department during those 5 years. Which basically means they would have to fund their phone R&D by revenue from their other venues. 

 

As such, yearly issues of new phones guarantee that the R&D gets constant fresh money injections.

 

Basically, the R&D required for today's tech was funded by tech of yesterday and the sales of todays tech fund the R&D for tech of tomorrow.

[Gmoneyinthebank]

On 7/5/2019 at 8:57 AM, IIIl said:

Is it about resources? Why couldn't  we produce processors like today's, a few years ago for example?

It's economics.

10 years ago they could have spent all their money working to make the stuff we have today.

But it would cost 10 times as much so no one could afford it.

No one would know how to write programs that could make use of ray-tracing, tensor cores, cuda cores, 28 thread CPUs, quad channel memory. So the power would go unused.

They would go out of business.

[Anomnomnomaly]

Part of the progression is down to research limitations, discovering new ways to make items more efficiently, faster and at lower costs... But a big part is also simply money... It's more profitable to release small incremental improvements each year than make big leaps.

 

AMD spent 7yrs flogging a crappy FX line of CPU's with no HT and terrible single core performance... Just releasing ever so slightly improved ones... It took 7yrs to design and get into production the first Ryzen CPU's to actually compete in the market again.

 

Then there's the backwards compatibility that's required... Making sure that your hardware will work on older products and software...

 

If you could eliminate the backwards compatibility side of the equation... You'd really see some ground breaking leaps... But it would require all new hardware, software, operating systems and so forth... Which makes the cost prohibitive until such a time as it gets decent market penetration.

 

So we have to put up with small incremental improvements... Which is why it's pointless upgrading a computer with each new increment... It's better to wait for at least 2 preferably 3 before upgrading.

 

Then you have companies that sit back and do very little to innovate because they have market dominance.. Why bother when there's no real competition to steal sales/revenue. This is why intel have literally been caught napping.. Lack of proper investment, issues with bring 10nm to market and complacency regarding AMD has led them to where they are now... They won't collapse or anything but they're in for a large revenue and market share drop.

[NineEyeRon]

Making one of something is easy.

 

Making lots of it and have them all work is hard.

[lewdicrous]

Meaningful improvements take a lot of time and money for research and development, these things don't happen overnight.

[Donut417]

8 hours ago, mariushm said:

invested a lot of money in the purchase of ATI

But it all was a result in purchasing ATI. Thats where they fucked up. If they didnt purchase ATI Im sure things would have been different in the CPU space. Intel would have never been complacent. Innovation would have happen sooner. 

[mariushm]

12 minutes ago, Donut417 said:

But it all was a result in purchasing ATI. Thats where they fucked up. If they didnt purchase ATI Im sure things would have been different in the CPU space. Intel would have never been complacent. Innovation would have happen sooner. 

You could call it a "strategic investment" ... think consoles, think laptops with integrated graphics (too bad Intel bribed their way and kept them for so long out of the profitable laptop market)... without it, maybe we would still have only Intel processors in consoles, with nVidia as video cards.

AMD needed ATI for their patents in video cards and their knowledge, they couldn't have made video cards without paying license fees to either ATI and/or nVidia, they needed integrated graphics in chipsets to counteract Intel chipsets... they couldn't rely on slowpoke SiS chipsets or nVidia integrated graphics..

Kind of same reason why Via bought the S3 graphics company, because they had the rights to make x86 processors and needed graphics to make their processors into a nice (embedded, low power etc) ...

 

AMD *had* to buy some graphics company, but maybe the timing wasn't right back then.

[mr moose]

Technology has been advancing on the same curve for at least as long as the industrial revolution.  Basically the more we know the more we can develop.  It is the same in every field of research be that medicine, computers, tooling/mechanical, etc.  this is why the curve is exponential and why moores law was a thing (it is a basic observation of a the mathematical function of development).

 

Interesting side note, it's also why people today think they are smarter than generations before or that some inventions/products are the single turning point in technology use, they haven't worked out they have the benefit of building on all the research and developments of the past.  

[wasab]

31 minutes ago, mr moose said:

Snip

So how long until we invent immortality by that curve? 

[mr moose]

1 minute ago, wasab said:

So long until we invent immortality by that curve? 

 

I wouldn't discount it as possible.  many times we have thought something impossible have become a thing.  I have already heard people saying that there are generations alive today who's average life expectancy will be over 100 by the time they get there. Sadly for me that will not be the case, however my doc is confident I will make to 90+ barring unexpected life threatening condiitons.