800 Million Transistors go Missing

[badb0y]

AMD depresses me.

I hope they don't screw up on the GPU side though...

 

[The J]

What about that "void" we see between the modules in the die shots? Is it possible that stuff had 800 million transistors in it and AMD just isn't counting them anymore?

 

[iCyborg]

You quote a conversation revolving around a singular point (the 65nm development and Hector's public comments about intentionally slowing down R&D) only to cite something that is also true but no more relevant (that AMD spends a disproportionate amount of revenue on R&D compared to Intel) as if Point A is somehow rendered irrelevant or untrue because Point B exists.

No one is disagreeing with Point B (the data you cite), but the conversation about Point A is one worth engaging in a conversation like this IMO.

Perhaps I should have replied only to exdeath, but since you concurred with him, I replied to both which is why I included exdeath's original post which isn't specific to this singular point (Hector's comment).
I still think it's relevant even to that: Intel is at 15% R&D and has great margins, AMD at 30% with poor margins. If they were at par or better before Conroe, it doesn't seem so crazy for AMD to try to make some cash too by having their R&D budget more aligned with the rest of the industry. Sure Conroe caught them with their pants down, but that was probably the biggest performance jump in the last decade.

 

[drizek]

Conroe was not a major performance jump. Merom was only slightly faster than Yonah, no more than the difference between SB and Westmere.

AMD have only themselves to blame. Anybody could see Conroe coming from a mile away. Banias was clock for clock competitive with AMD and running at a fraction of hte TDP. Boost the TDP, boost the performance, kill the AMD advantage. It was only a matter of time.

 

[iCyborg]

http://www.anandtech.com/show/2045/19
Intel's Core 2 processors offer the sort of next-generation micro-architecture performance leap that we honestly haven't seen from Intel since the introduction of the P6.

But you're entitled to your opinion...

 

[drizek]

On the desktop, yes. Nobody disputes that the Pentium 4 was terrible. But the performance potential of Conroe was already visible in Yonah and Dothan and Banias.

Overall, Merom may not be as big of an upgrade to Yonah as Conroe was to NetBurst, but the bottom line is that you get equal or better performance in every test without increasing cost or decreasing battery life. Owners of Core Duo laptops really have no reason to worry about upgrading for now, and waiting for the Santa Rosa platform before your next laptop upgrade seems reasonable.

http://www.anandtech.com/show/2056/17


AMD should have known that once Intel put this thing in the desktop and allowed it to draw a hundred watts, they were going to be in trouble.

 

[Idontcare]

If they were at par or better before Conroe, it doesn't seem so crazy for AMD to try to make some cash too by having their R&D budget more aligned with the rest of the industry.

I disagree, it was crazy. The phrase "resting on your laurels" exists for a reason, and its use is rarely invoked as a means of bestowing accolades.

AMD's key decision makers - with Hector at the top - figured out how to literally buy ATI at its absolute peak price/premium while electing to intentionally slow down their R&D and allow for the very possibility that Intel would catch up and surpass them.

There's a word for that - arrogance.

It ran entirely counter to Intel's mantra - only the paranoid survive.

And if Hector's criminal involvements are any indication of the character of the man, he was clearly not paranoid enough given that he was caught out

Its not that Hector himself was inconsistent, on the contrary he was remarkably true to his nature, but its only in hindsight that we outsiders piece together the puzzle and then step back and realize what it was that Sanders unleashed on the company.

Consider the very reasons he torpedoed the AMD/NV merger - ego and not wanting to let Jensen take the helm of AMD despite the fact that Jensen had a much better track record of being able to successfully manage a company in this industry.

Rather than set his personal ego aside and do what was best for shareholders and employees, he forced everyone else but himself to accept second-best by going with the ATI deal and proceeded to literally throw billions of dollars away on the buyout.

Bah, at any rate the point is that excuses do not make up for lost business, and lots and lots of excuses do not make up for the loss of lots and lots of business. Regardless Hector's excuse for intentionally slowing the development of 65nm, it came as no surprise that it was a fatal mistake.

 

[piesquared]

Oh come on. lol Lot's of one liners in there. Hector was CEO in one of the most difficult times for AMD. As many companies have found, it is difficult to impossible to compete with a monopoly who's objective is to compete with the competition by abusing their market status.

Implying that it was soley Hector's decision to buy ATi instead of merging with NV, due to ego or arrogance, is ridiculous. I would imagine the BoD had some desicsion making input and obviously wasn't the direction AMD wanted to go. Clearly it was the right direction indicated by the success of Fusion. It's not hard to figure out that building a roadmap and implementing it for the future doesn't happen in a year.

There were projects in the pipeline that never met projections sure, that doesn't mean they were sitting idle doing nothing.

No, the disdain for Ruiz comes from intel/intel's investor wing being butthurt (there's your ego and arrogance) because he had no option but to sue intel, and got the ball rolling for countries around the world to follow suit.

 

[jhu]

http://www.anandtech.com/show/2045/19
Intel's Core 2 processors offer the sort of next-generation micro-architecture performance leap that we honestly haven't seen from Intel since the introduction of the P6.

But you're entitled to your opinion...

Perhaps if you're coming from Netburst, then yes Core 2 is pretty awesome. However, if you're coming from the Pentium M line (from which Core 2 was eventually derived from), then not so much.

 

[LoneNinja]

http://www.anandtech.com/show/2045/19
Intel's Core 2 processors offer the sort of next-generation micro-architecture performance leap that we honestly haven't seen from Intel since the introduction of the P6.

But you're entitled to your opinion...

Core2 was a small performance gain over Core. While Intel was running Netburst against K8 on the desktop and loosing horribly, the CoreDuo processors at much lower power consumption and clock speeds were beating most of AMDs and Intels desktop processors. AMD knew for some time that Intel had a vastly superior architecture and did nothing, than got crushed when Intel released the Core2 for desktop.

 

[iCyborg]

I disagree, it was crazy. The phrase "resting on your laurels" exists for a reason, and its use is rarely invoked as a means of bestowing accolades.

AMD's key decision makers - with Hector at the top - figured out how to literally buy ATI at its absolute peak price/premium while electing to intentionally slow down their R&D and allow for the very possibility that Intel would catch up and surpass them.

There's a word for that - arrogance.

It ran entirely counter to Intel's mantra - only the paranoid survive.

And if Hector's criminal involvements are any indication of the character of the man, he was clearly not paranoid enough given that he was caught out

Its not that Hector himself was inconsistent, on the contrary he was remarkably true to his nature, but its only in hindsight that we outsiders piece together the puzzle and then step back and realize what it was that Sanders unleashed on the company.

Consider the very reasons he torpedoed the AMD/NV merger - ego and not wanting to let Jensen take the helm of AMD despite the fact that Jensen had a much better track record of being able to successfully manage a company in this industry.

Rather than set his personal ego aside and do what was best for shareholders and employees, he forced everyone else but himself to accept second-best by going with the ATI deal and proceeded to literally throw billions of dollars away on the buyout.

Bah, at any rate the point is that excuses do not make up for lost business, and lots and lots of excuses do not make up for the loss of lots and lots of business. Regardless Hector's excuse for intentionally slowing the development of 65nm, it came as no surprise that it was a fatal mistake.

But how is throwing significantly more revenue into R&D than your competitors and losing money/breaking even eternally any better business strategy?
I'm not trying to defend Ruiz, I dislike him as much as anyone here, but the alternatives were not much brighter. And if BoD had the power to kick out Dirk for (allegedly) not moving into tablets/smartphones, they didn't have to stand still regarding Ruiz' handling of AMD/nV and ATI buyout. I'd say there are more people to share the blame.

There are 5 major soccer leagues in Europe that stand out money-wise from the rest: English, Spanish, Italian, German and French. In the last 15 years, only one club outside of these leagues won the Champion's league, and only one other made it to the finals. I don't think this is because players from other leagues aren't putting enough efforts or all the managers are doing a poor job. I know about the "correlation is not causation" adage, but frequently correlation is causation.

 

[iCyborg]

Core2 was a small performance gain over Core. While Intel was running Netburst against K8 on the desktop and loosing horribly, the CoreDuo processors at much lower power consumption and clock speeds were beating most of AMDs and Intels desktop processors. AMD knew for some time that Intel had a vastly superior architecture and did nothing, than got crushed when Intel released the Core2 for desktop.

So why did Intel develop Prescott instead of just doing a desktop Pentium M? It it was visible "from a mile away", Intel should have seen it from 2 miles away. Of course AMD should have expected that Intel won't like being behind performance-wise for long. The part where I disagree is that "not developing products competitive with someone who has 5-10x the resources years after years" == "doing nothing".

 

[jhu]

So why did Intel develop Prescott instead of just doing a desktop Pentium M?

Prescott was a processor arch designed by the Intel's marketing folks.

 

[Idontcare]

But how is throwing significantly more revenue into R&D than your competitors and losing money/breaking even eternally any better business strategy?

Its not, but if you intend to remain competitive with a competitor that has 6x your revenue then you have no choice.

The difference between AMD's reality and their desire is the same thing that made all the difference (in a negative way) for Cyrix/Via. Cyrix/Via made their choice, and they reaped what they sowed.

AMD sowed. Now they weep reap.

 

[TidusZ]

8 cores, 3.6 ghz, and 2 billion transistors looks nice on the box in the store or online. The fact that its 4 cores, 1.2 B transistors and slow as molasses going uphill are not things most people would bother researching about before buying unfortunately.

They are essentially lying about the number of cores so it shouldn't be a surprise they lied about transistor count.

 

[Arachnotronic]

I'm surprised that everyone focuses on AMD's weaker x86 performance and ignores their stronger iGPU performance versus Intel.

If I could have a Llano or a SNB and had to use the IGP, it'd pretty dang clear that I'd pick the Llano without batting an eye.

AMD will be fine. Fusion is a good strategy. It is the mainstream strategy, and mainstream works. Their main goal should be to focus on market penetration and keeping production high enough to satisfy demand. If they try to chase the high end, monster CPU market, Intel will crush them handily. They can't afford to play this game.

Not many people buy extreme edition Intel CPUs. In fact, I feel like SNB-E was almost as badly received as the FX 8150.

 

[maddie]

Its not, but if you intend to remain competitive with a competitor that has 6x your revenue then you have no choice.

The difference between AMD's reality and their desire is the same thing that made all the difference (in a negative way) for Cyrix/Via. Cyrix/Via made their choice, and they reaped what they sowed.

AMD sowed. Now they weep reap.

I 100% agree.

When smaller, you have to invest a bigger percentage to stay in the game, assuming both parties are competent. Over time, you work to become bigger.

Hector is a product of modern management, short term share price gains. Look at the present state of the world.