John Fruehe (JF-AMD) no longer with AMD

[ShintaiDK]

I do think however, that we can agree that outside Intel. Other foundries will have a hard time keeping the process node race up, unless they expand the timeframe to recover the ROI.

The increased cost is kicking players out one by one. And if Intel gets Atom to dominate smartphones. Then it will look really bad for independent foundries.

The endgame only got room for 1 foundry. And Intel is already 3-4 years ahead of TSMC, not to mention economics.

 

[nehalem256]

If you are going to attempt to speak technically, you should note that Bulldozer was made on a 32nm process instead of Phenom II's 45nm process.

Taking that into account, one can easily see how bloated Bulldozer is.

It seems like they tried to save die space by merging 2 cores together, but then doubled the L2 cache size to make up the difference.

Having more Cache while being smaller than last gen, i dont see whats the problem with that to call it bloated ?? I believe you are negative biassed towards anything called bulldozer.

Because the more cache did not result in more performance.

 

[Ajay]

I have the unfair advantage of having worked (1) as a process node development engineer, (2) at a foundry (Texas Instruments, foundry to Cyrix and SUN), and (3) while operating in the capacity of a customer of TSMC, UMC, and Chartered.

You can't walk in those shoes without coming to learn a thing or two when it comes to reality versus marketing And Nvidia is very good at marketing TSMC's industry-wide generic issues as being Nvidia's Achilles heel.

Hey, that's why we pay you the big bucks

That's a funny quote in that thread. Physics + statistics == NV Fail; if it wasn't an inside joke, some of NV's own engineers must have been cringing.

I had thought of going into the semi-con business at the end of my sophomore year, but my University didn't have an Applied Physics program that I could have switching into. I finally managed to get a professor who would do an independent study on solid state physics my senior year, but he treated it as a grad course - given the work load I already had with three other senior level physics classes a grad level class was too much (~ twice the work of my other senior classes), so I switch and took optics instead. Too bad, I didn't go into optics, but semi-con would have been a great career IMO.

 

[Ajay]

I do think however, that we can agree that outside Intel. Other foundries will have a hard time keeping the process node race up, unless they expand the timeframe to recover the ROI.

The increased cost is kicking players out one by one. And if Intel gets Atom to dominate smartphones. Then it will look really bad for independent foundries.

The endgame only got room for 1 foundry. And Intel is already 3-4 years ahead of TSMC, not to mention economics.

Well, that's why I think Glofo did a great job landing a contract with ARM (which will including tuning process and cell libraries, etc. for better outcomes for IHV's ARM SOCs).

TSMC's business model doesn't seem to be in jeopardy any time soon. Demand for 28nm fabbed parts is high enough that some companies like Qualcomm are using multiple fabs to source their SOCs. I expect the same to be the case for 20nm & 14nm parts as well. After that, we'll see.

With so much of the industry focused on ARM based mobile devices, Intel is facing an uphill battle, even if they have superior price/performance and performance/watt. This will be interesting battle to watch - no doubt about that.

I'm moving this thread way off topic Then again, that seem to be SOP at AT.

 

[Idontcare]

I do think however, that we can agree that outside Intel. Other foundries will have a hard time keeping the process node race up, unless they expand the timeframe to recover the ROI.

The increased cost is kicking players out one by one. And if Intel gets Atom to dominate smartphones. Then it will look really bad for independent foundries.

The endgame only got room for 1 foundry. And Intel is already 3-4 years ahead of TSMC, not to mention economics.

Heh, no question, come <10nm for MPU it'll be just 3 players - Intel, TSMC, and Samsung.

If GloFo is still in the running it will only be because their owners have bigger and grander plans for semiconductor manufacturing as it relates to a national strategic agenda. (one which I happen to fully agree with and support, I only wish my nation's government felt domestic semiconductor manufacturing was a national security concern as well )

On an unrelated note, but related to the thread's topic: Anand Mandapati & David Wang: Key Console and Desktop GPU Architects Leave AMD

The last time I remember seeing a company bleed so much talent and high-visibility people over the course of a year or two was when DEC was in its final days of implosion.

AMD's economics are nowhere as dire as DEC's were, and I doubt an imminent demise is in their future, but whatever Rory is up to he sure is causing a lot of churning within the ranks.

I suppose it would be more accurate for me to say this reminds me of the employee churn rate when Carly Fiorina took the helm of HP. It did not lead to the demise of HP as a business but it surely led to the demise of HP's corporate culture and a great many talented folks fled the company to escape that abysmal environment.

Could be good for AMD if the churning is removing people who are responsible for AMD's decline post K8. Clearly change was needed. But not so good if the churning is because smart people are smart enough to know they need to jump ship for the sake of preserving their own future and job security.

 

[Phynaz]

if glofo is still in the running it will only be because their owners have billions of dollars of oil money

ftfy :d

 

[Idontcare]

ftfy :d

Don't kid yourself, our government has billions in oil money that they could reinvest into propping up the domestic semiconductor industry. But our government's priorities are not aligned with that agenda. We are much more interested in investing our billions into maintaining a military complex.

Blew my mind when it was pointed out to me that we spend as much on our military as 50% of Russia's GDP And that the US's "defense" spending at the federal level is half of the entire global expenditures for military.

In other words we literally are spending at a rate as if we are at war with the entire rest of the world already, our $1 to every one of theirs.

I don't blame GloFo for being a national strategic investment for jobs and so on, better plows than guns IMO. It is simply a matter of national priority, ours is to maintain a military that occupies an economic footprint that is equivalent to that of all the militaries of all the other 170+ countries on this planet combined.

UAE's priority is to invest its financial windfall of today into something that stands a chance of providing domestic jobs tomorrow. What's not to like about that?

 

[ShintaiDK]

I wouldnt be surprised if China also threw money after it. State foundry business to secure national interests and security. Not to mention hightech jobs and revenue.

SOEs are already doing quite well there in overall to put it mildly.

 

[Ajay]

Don't kid yourself, our government has billions in oil money that they could reinvest into propping up the domestic semiconductor industry. But our government's priorities are not aligned with that agenda. We are much more interested in investing our billions into maintaining a military complex.

Blew my mind when it was pointed out to me that we spend as much on our military as 50% of Russia's GDP And that the US's "defense" spending at the federal level is half of the entire global expenditures for military.

In other words we literally are spending at a rate as if we are at war with the entire rest of the world already, our $1 to every one of theirs.

I don't blame GloFo for being a national strategic investment for jobs and so on, better plows than guns IMO. It is simply a matter of national priority, ours is to maintain a military that occupies an economic footprint that is equivalent to that of all the militaries of all the other 170+ countries on this planet combined.

UAE's priority is to invest its financial windfall of today into something that stands a chance of providing domestic jobs tomorrow. What's not to like about that?

And that spending didn't include the two wars we were in (and one still), since they were 'off the books' and not counted as defense spending :\

Folks I've spoken to in the military and retired say a phased 1/3 reduction in spending would have little impact on our ability to defend ourselves against existing threats and still retain the power to project our military might if needed. We are the world's 'police' force and keep larger conflicts from escalating, but only we pay for that world wide police force. I'm all for cutting the military budget, btw.

The other problem is how do we allocate those funds. Intel is doing a good job investing in the US and much more efficiently then a bureaucracy would. Should the US dump 10B on Glofo to expand production in New York? Is that fair to Intel? These are just questions. I largely agree with you, but I don't know what the best plan is.

 

[Vesku]

Yeah, it does seem like the people who have had the most exposure to Rory Reads ideas and vision for AMD have suddenly found better places to be. =S Perhaps the board recruited someone a little too focused on the low end of the market. I recall many SoC design positions popping up for AMD. If the focus is now almost entirely on low wattage designs, these departures aren't that surprising.

Heh, no question, come <10nm for MPU it'll be just 3 players - Intel, TSMC, and Samsung.

If GloFo is still in the running it will only be because their owners have bigger and grander plans for semiconductor manufacturing as it relates to a national strategic agenda. (one which I happen to fully agree with and support, I only wish my nation's government felt domestic semiconductor manufacturing was a national security concern as well )

On an unrelated note, but related to the thread's topic: Anand Mandapati & David Wang: Key Console and Desktop GPU Architects Leave AMD

The last time I remember seeing a company bleed so much talent and high-visibility people over the course of a year or two was when DEC was in its final days of implosion.

AMD's economics are nowhere as dire as DEC's were, and I doubt an imminent demise is in their future, but whatever Rory is up to he sure is causing a lot of churning within the ranks.

I suppose it would be more accurate for me to say this reminds me of the employee churn rate when Carly Fiorina took the helm of HP. It did not lead to the demise of HP as a business but it surely led to the demise of HP's corporate culture and a great many talented folks fled the company to escape that abysmal environment.

Could be good for AMD if the churning is removing people who are responsible for AMD's decline post K8. Clearly change was needed. But not so good if the churning is because smart people are smart enough to know they need to jump ship for the sake of preserving their own future and job security.

 

[CHADBOGA]

Blew my mind when it was pointed out to me that we spend as much on our military as 50% of Russia's GDP

The US spends around 4% of its GDP on Defence, are you saying that Russia's GDP is only 8% of the USA's?

 

[Ferzerp]

It's much closer to 5% of GDP, and Russia only has around 12% of the US's GDP, so while not exactly 50%, it's close. It's really more like 40%, but what he said wasn't unreasonable.

The USSR fell apart, remember?

 

[AnandThenMan]

The United States accounts for almost 50% of military spending for the entire planet.

 

[ctsoth]

Maybe the US government could prop up domestic fab industry..... Chips could be sold at a loss and the taxpayer covers the difference. Maybe it would be just as successful as propping up us auto makers.

http://nlpc.org/stories/2012/08/16/auto-bailout-legacy-gms-european-nightmare

 

[bononos]

The US spends around 4% of its GDP on Defence, are you saying that Russia's GDP is only 8% of the USA's?

More like 12%(nominal) of US GDP but what IDC said is still broadly correct since US spends roughly as much on its defense as the rest of the world put together. One overlooked reason for such high defense spending is that it is also a public jobs program on the sly for millions albeit dressed up as 'defending democracy'.

 

[bononos]

...........

I suppose it would be more accurate for me to say this reminds me of the employee churn rate when Carly Fiorina took the helm of HP. It did not lead to the demise of HP as a business but it surely led to the demise of HP's corporate culture and a great many talented folks fled the company to escape that abysmal environment.

Could be good for AMD if the churning is removing people who are responsible for AMD's decline post K8. Clearly change was needed. But not so good if the churning is because smart people are smart enough to know they need to jump ship for the sake of preserving their own future and job security.

Could you expand a little (or alot) on the demise of HP's (or DEC) abysmal corporate culture, its the kind of thing that I don't hear alot of in detail.

 

[CTho9305]

On an unrelated note, but related to the thread's topic: Anand Mandapati & David Wang: Key Console and Desktop GPU Architects Leave AMD

The last time I remember seeing a company bleed so much talent and high-visibility people over the course of a year or two was when DEC was in its final days of implosion.

Do people watch LinkedIn for larger-scale career moves? It seems like it wouldn't be difficult to get a sample of a few thousand people in the industry to figure out where people are moving from/to. Might make for some excitement for the rumor sites .

While I have no doubt that Nvidia has unquestionable access to pricing info on past nodes as well as price negotiations underway on nodes under development to form the basis of their graph above - I would point out that if we are to believe the data presented in this graph then we are to believe that 40nm never brought any cost savings over 55nm in its entire history of production and that is without question not true.

The data containing such an obvious error makes me question the assumption then that 20nm will not be lower in cost than 28nm.

However, there is one way that I can rationalize the data in the graph, and that involves this graph not representing cost/transistor but instead representing TSMC's willingness to attempt to balance supply with demand while maximizing their own gross margins.

<snip>

Customers looking to buy 20nm wafers versus 28nm wafers are going to be looking to buy those 20nm wafers because the resultant parametric properties of the chips themselves can command a higher resell price in the market - better power consumption, better clockspeeds. And TSMC knows this, so they would be the fools to not price accordingly.

<snip>

And that makes the story behind the graph a very different one indeed, it has nothing to do with new node production costs at TSMC and everything to do with gross margins at TSMC.

I'm less skeptical of this graph than you are. Of course, you've had access to more real numbers than I have (none), so take this with a grain of salt too...

Capitalism is likely *a* factor, but I don't doubt that costs are really going way up. The latest process nodes keep increasing the number of layers that require double-patterning to make, which reduces the throughput you get from your equipment.

High-performance CPUs have always used a tapered metal stack - basically, the first couple layers allow for very fine wires, but the next few layers can only print coarser wires, and so on, up the stack. This gives density for short, local routing, and low resistance for longer wires. It also helps with costs, since the coarser wires can be built with older technology. ASICs generally use a metal stack that has fine wires all the way up (ignoring 1-2 power distribution layers at the top), which allows for increased density (i.e. theoretically smaller, cheaper chips). Unfortunately, once the finest wires require double patterning to print, the cost of using them on every layer becomes prohibitive, and ASICs may need to switch to tapered metals. This also reduces density, so you don't get as much area scaling on the new process.

Before you ask how to trade off density vs. equipment cost... I don't know. Maybe because the CPUs historically moved to new nodes before the ASICs, they couldn't afford to use the thin metals for more layers, and since the ASICs always lagged in the past, they just waited for the cost to come down enough, but now the ASICs lag less behind (non-Intel) CPUs and experience more of that cost pressure?

Intel 65nm metal stack for high-performance CPUs:

This is apparently TSMC 90nm (looks like 7 thin layers + 2 thick):

This one showing TSMC 40nm appears to have a little tapering in it, but the first 6 are still all thin. 12 layers seems excessive, which makes me wonder what the "TSMC" image is really showing though:

Basically, the graph isn't inconsistent with the trends and complaining in the industry. The general attitude seems to be that the latest process nodes are only interesting for power savings; you're not saving much (if any) money, and you're not gaining much (if any) speed.

 

[ShintaiDK]

More like 12%(nominal) of US GDP but what IDC said is still broadly correct since US spends roughly as much on its defense as the rest of the world put together. One overlooked reason for such high defense spending is that it is also a public jobs program on the sly for millions albeit dressed up as 'defending democracy'.

Pretty bad exchange for jobs tho. US military is hugely inefficient and largely plagued by corruption. All it does is to channel more money into already rich people and companies that are all busy moving them to Cayman. And thats not making more jobs for average Joe.

Trim it and start using some of it on renewable energy that would make the US energy independent. Huge trade balance savings and 100000s of substainable self paying US jobs. Then move on from there with other sectors.

We are doing it in Denmark. I am sure you can do it in the US as well with even bigger oppotunities.

 

[Idontcare]

Capitalism is likely *a* factor, but I don't doubt that costs are really going way up. The latest process nodes keep increasing the number of layers that require double-patterning to make, which reduces the throughput you get from your equipment.

Sure, no question, costs per-wafer increase with every node, that has always been the case. Its not just the addition of more metal levels, it is also the additional processing time that comes with dealing with the required additional degree of precision necessary for smaller pitches.

The cleans take longer, the etches take longer, the depositions take longer. Not because they have to do more, they actually have to do less, but because they have to do less with even more precision than ever before and the way you achieve that is by extending processing times and reducing throughput.

But you can't have cycle-time increase, so you must buy more tools to keep up the aggregate fab throughput, and that is where things get expensive when rolling out your newly minted node. Same issue with double-patterning, buying extra 193nm immersion tools is expensive. Using them to print patterns of ever finer details requires a throughput hit on top of that because of overlay and alignment requirements. Usually you don't take this hit at the scanner, you take the hit in your rework cycle-time. Reworking wafers 2 and 3 times before you accept the printed patterning and commit them to etch.

That said, even still the costs generally never exceed 30% node-over-node (and are usually limited to 15-20% by design, as an engineering constraint set by management), leaving room for improvements in gross margins unless you are building 500mm^2 die every single node and not gaining any increase in die counts per wafer.

In that case your per-die costs are increasing 15-20% per node, but your per-xtor costs are still decreasing.

And the graph is normalized on a per-xtor basis. In order for 40nm to have provided no cost benefits over 55nm the wafer costs themselves would have to double. That did not happen. Wafer costs increased 20%, in-line with the traditional node cost increases.

That is what I disagree with in terms of what the graph means. It is implying node costs are increasing at a near doubling rate commensurate with the shrinking rate itself. That didn't happen at 40nm, but the graph would have you believe that. And it would also have you believe it will happen at 20nm, which I dispute as well.

But I have no doubt that TSMC is going to charge exactly the highest price-premium they can on 20nm wafers in those first 4 quarters when demand is high and supply is very limited. That is what they are in business for, making as much money as they can per wafer they sell. And Nvidia can make noise about that but they won't get any reprieve in the price-premiums TSMC is charging until TSMC feels competitive pressures from UMC and GloFo for those 20nm wafers.

And at the end of the day Nvidia is just going to pass on those cost-premiums on to us, the end-users. Its not like they are UNICEF doing charity work and worrying about their customers paying $2k for next-gen TESLA SKUs

 

[Idontcare]

The United States accounts for almost 50% of military spending for the entire planet.

When the aliens attack again, the rest of the planet can thank us for at least being as prepared as was humanly possible

Trim it and start using some of it on renewable energy that would make the US energy independent. Huge trade balance savings and 100000s of substainable self paying US jobs. Then move on from there with other sectors.

We are doing it in Denmark. I am sure you can do it in the US as well with even bigger oppotunities.

Word.

I got into the semiconductor industry because I was initially working on developing improved efficiency solar cells for NASA. And later on my grad research was based on developing compounds that could be added/dissolved in water which would then use sunlight to split water ino hydrogen gas and oxygen. (standard "harvesting the sun" themes here) Those applied science experiences were my gateway to a career at TI in semiconductor development.

There are big ways in which the USA could subsidize domestic semiconductor manufacturing. And I'm not talking about buying companies outright, but it all starts with education and a labor base.

Case in point, my own job. It was outsourced to Taiwan, TSMC to be exact. (TI shut down internal CMOS development in 2007, solely relying on the foundries for post-45nm) And the reasons were pretty obvious, we engineers cost TI a pretty penny. My salary was 6 figures, as was everyone I worked with. Why? Did I need 6 figures in salary to do my job? No. I could have been paid 50% of my salary and still made a decent living.

But why was my salary so high? Lack of competition. TI had very few options when it came to hiring someone in the US who could do what I did, and I made them pay dearly for it. To dearly, so they shut it down and outsourced the work the people in Asia who were willing to be paid less.

The reason there was such a limited pool to hire from is because the barrier to entry is ridiculously high. To be able to do my job I had to attend 10yrs of higher education at an upfront expense of nearly $150k. (and that was some time ago, surely that cost has doubled or tripled by now )

How many new college students want to take on such an expense, and such a timeline before seeing a paycheck at the end of the journey? Not many, which was to my personal benefit but not to the benefit of my employer.

So one thing the government could do if it wanted to promote domestic semiconductor manufacturing is to lower that barrier to entry to increase the labor supply. It is the last thing the current labor supply wants to have happen though, no engineer wants competition, no engineer wants their salary to be cut in half.

But the fact is you can live in this country on $60k-$70k per year, you do not need $120k-$140k per year. And the truth is no one can really afford $120k-$140k per year because it ultimately strangles the domestic market and drives the jobs off-shore.

But that all just entails trading off where dollars go as they get spent. We are still spending the same dollars. Pay your engineers 50% less and now they are spending 50% less, your starbucks jobs go away, restaurant jobs go away, tourism dollars go away, as all that excess disposable income is gone.

So simply reducing wages within any given business sector is not the answer at a macro-economic level.

You must get people working on things that improve productivity, improve the creation of wealth. Investments into developing a domestic alternative energy industry does exactly that. In the 1930's the US government put a lot of people to work building dams to generate hydroelectricity.

That cheap electricity fueled a lot of jobs, and the electricity itself fueled a lot of industries. Wasn't good for the fish, and we now recognize this, but the benefits we gained from that experiment are real and repeatable.

 

[mrmt]

Its pretty clear to me, and I'd think it is clear to most everyone else, that John was taking counsel (getting his info) from the same internal sources that were busy convincing executive management to funnel R&D money to them to create that which became Bulldozer in the first place.

John Fruehe was thoroughly questioned in every forum out there and he didn't back off. I think the most virulent and technically accurate questioning came from this forum, from Scali, that was banned for this btw. Others were banned too, OTOH almost all of the AMD shock troop that helped him to spread the FUD in the community and defended him is still around posting like nothing happened.

I'm pretty sure that he also lied to his AMD customers and in exchange he received the same or even deeper and technically strong questioning, and if he had the same posture that he had here, which was insult the people making questions and dismissing their claims as absurd per se then he was really burned with customers.

He wasn't hyping bulldozer for a week or two, he was hyping bulldozer for the tech community literally for years and I doubt that he started with the tech community. Customers should have been receiving his crap for even longer, and I wouldn't want to receive someone that lied to me for that long in my company. How can someone trust him in a negotiation? I wouldn't buy even a webhost server with him, much, much less build a datacenter

And he didn't back off. He lied until the last minute. Even when Sismark leaked the benchmarks one month before bulldozer launch he denied that those benchmarks were true, and there is no way he can deny that he knew that bulldozer at that time.

Remember that this wasn't the only thing he lied about. He lied about AMD TDP definition right on semiaccurate's forums, one thing that you can easily find on AMD documentation free on the web, and he lied about AMD then-current architecture at the time, and he didn't back off in both cases. Did he lie to his customers about that too?

So we are not dealing with an honest professional here, we are dealing with a recurrent liar. The problem now is that bulldozer was a lie too big to get ignored.

 

[jones377]

Nobody got banned from AT for disagreeing with JFAMD. Nobody. That is simply revisionist history. I disagreed with him fiercely, especially over the FPU part. I didn't get banned. I also turned out mostly right but that's another matter entirely

Some people are simply unable to disagree with another person without making it personal. THAT IS WHY they got banned or suspended.

 

[ShintaiDK]

When the aliens attack again, the rest of the planet can thank us for at least being as prepared as was humanly possible



Word.

I got into the semiconductor industry because I was initially working on developing improved efficiency solar cells for NASA. And later on my grad research was based on developing compounds that could be added/dissolved in water which would then use sunlight to split water ino hydrogen gas and oxygen. (standard "harvesting the sun" themes here) Those applied science experiences were my gateway to a career at TI in semiconductor development.

There are big ways in which the USA could subsidize domestic semiconductor manufacturing. And I'm not talking about buying companies outright, but it all starts with education and a labor base.

Case in point, my own job. It was outsourced to Taiwan, TSMC to be exact. (TI shut down internal CMOS development in 2007, solely relying on the foundries for post-45nm) And the reasons were pretty obvious, we engineers cost TI a pretty penny. My salary was 6 figures, as was everyone I worked with. Why? Did I need 6 figures in salary to do my job? No. I could have been paid 50% of my salary and still made a decent living.

But why was my salary so high? Lack of competition. TI had very few options when it came to hiring someone in the US who could do what I did, and I made them pay dearly for it. To dearly, so they shut it down and outsourced the work the people in Asia who were willing to be paid less.

The reason there was such a limited pool to hire from is because the barrier to entry is ridiculously high. To be able to do my job I had to attend 10yrs of higher education at an upfront expense of nearly $150k. (and that was some time ago, surely that cost has doubled or tripled by now )

How many new college students want to take on such an expense, and such a timeline before seeing a paycheck at the end of the journey? Not many, which was to my personal benefit but not to the benefit of my employer.

So one thing the government could do if it wanted to promote domestic semiconductor manufacturing is to lower that barrier to entry to increase the labor supply. It is the last thing the current labor supply wants to have happen though, no engineer wants competition, no engineer wants their salary to be cut in half.

But the fact is you can live in this country on $60k-$70k per year, you do not need $120k-$140k per year. And the truth is no one can really afford $120k-$140k per year because it ultimately strangles the domestic market and drives the jobs off-shore.

But that all just entails trading off where dollars go as they get spent. We are still spending the same dollars. Pay your engineers 50% less and now they are spending 50% less, your starbucks jobs go away, restaurant jobs go away, tourism dollars go away, as all that excess disposable income is gone.

So simply reducing wages within any given business sector is not the answer at a macro-economic level.

You must get people working on things that improve productivity, improve the creation of wealth. Investments into developing a domestic alternative energy industry does exactly that. In the 1930's the US government put a lot of people to work building dams to generate hydroelectricity.

That cheap electricity fueled a lot of jobs, and the electricity itself fueled a lot of industries. Wasn't good for the fish, and we now recognize this, but the benefits we gained from that experiment are real and repeatable.

Look at socialism so to say. Even tho I know its basicly the devil itself over there due to the witchhunts under mccarthy. Tax to prevent overabuse of products. Free education so you dont drop your talented people on the floor and get a higher more educated workforce. And we got alot higher wages than in the US, but we got huge trade balance surplus and green energy.

We live in an apartment (condo I guess the right name when we own it.) thats 103m2 here in Copenhagen. Thats 1110sqf or so for you americans.

The building itself is B1 energy rated and from 2007 and every healthy to live in with fresh air. So our heat consumption is extremely low and we use block heating. Had it been A1 marked, a candle could heat the entire apartment in winter when it was -10 below outside. Our household contains the usual fridge/freezer, stove, oven, dishwasher, tumbler, washer plus 55" LED TV, 4 PCs where one is a HTPC that runs 24/7. I work from home so the usage is also higher. We use ~3Kw/h a day or 95Kw/h a month without sacrificing anything. Just doing it smarter because there is a tax on electricity. We didnt get LED backlight and A+++ appliances because electricity was cheap, we got it because it was expensive. In the US you basicly just build more powerplants as a solution instead of fixing the root of the cause and making everyday life for americans cheaper and better with quality indoor climate.

Same goes for cars etc. You subsidize fuel in the US and people drive around in retarded big cars to say it gently. We tax fuel to get people to drive smart efficient cars.

All the options are there and you dont have to waste it away being cheap. Sugar tax and fat tax if people start to eat too crappy. Or if its too cheap. Again you know this problem to the extreme in the US as well. You even make cows eat subsidized corn instead of grass because its cheaper. Even tho you then have people dying from e.coli due to that. Simple grass would fix it, plus the cows dont have good of the corn. US solution is currently to wash the meat in amonia. Hightech solution for a lowtech problem. And the hightech fails obviously. When you try to get productivity increases you make people work more, even tho they produce less and create more errors this way. Not to mention their quality of life gets hit.

And it all goes hand in hand really. Regulations and consumption control dont hinder your freedom. It expands it.

Talking about solar cells:
http://epn.dk/brancher/energi/alternativ/article4811616.ece

Insurance company that used all their roofs for it. Work smart, not hard.

Key elements to success is efficiency, urbanization and regulation of usage.

I always get sad when I look at the US. It got everything and all the elements to easy success in one of the most beautiful landscapes there is. But you throw it all away on the floor.

 

[mrmt]

And he knew this was at stake back when he was talking up bulldozer, which means the only reason he'd put so much of his personal reputation on the line was that he had been thoroughly convinced it was all real by the same engineers he kept referencing as his internal sources for the IPC gains and areal die-size tradeoffs for bulldozer's CMT

I think you are underestimating the objectives of the hype in the first place. You hype a product to bring attention to it, to make people aware of it, and we cannot deny that Fruehe was *very* successful in that.

At the first moment he knew that he was making a claim based on the engineering assessment. Questioning from the community and from customers should have lit a red light for him and he should have stopped, or at least be more economical in the hyping, nobody would blame him for that but no, he couldn't, he had to go full speed ahead. He gambled.

The essence of his gambling was that regardless of achieving or not performance targets Bulldozer wouldn't be a failure of the magnitude it was. He then could just get away with it easily saying "engineering said this, but instead we got that, but it is still a fine product, not as good as we were expecting but still ok" and the product would still have a lot of attention due to the previous hype. But with bulldozer being a failure of such a magnitude there was no easy get away for him, the gamble didn't pay off and he was left in the cold to face criticism.

 

[georgec84]

What is this thread about?