Intel has dropped a stink bomb into its ongoing dust-up with the US Federal Trade Commission over allegations of anticompetitive practices, revealing a 2004 communication in which AMD's then-top salesman referred to his own company as "pathetic" and said he "would never buy AMD for a personal system if I wasn't workin here." The …
Can't argue with that
AMD sat on their laurels while the Athlon 64 was kicking the Prescott P4s in the ghoulies, then Intel came back with better technology. AMD had nothing as they'd cancelled or delayed their next generation of chips, they're so far behind in performance now I can't see them catching up without another game changing chip like the Athlon 64.
Right on the mark
It seems AMD lost one of their most competent people in Mr Richard. Intel products have always been superior to AMDs processors. My guess is that even if they wanted, they simply could not match the scale and economics of Intel.
Even companies like IBM and HP are having big problems to keep up with Intel. The X86 killed the excellent Alpha of DEC and HP's PA Risc, which was also pretty strong. HP outsourced its Itanium CPU to Intel. SPARC appears to be the next victim of Intel. Only IBM still has something competitive with the Power CPU. But even they will be hard pressed to compete with the economics of Intel process technology and manufacturing the next couple of years.
Intel is just a natural monopoly (due to enormous R&D expenses) like Microsoft, Google and (earlier) IBM. Get used to it.
Opteron vs pre nehalem Xeons?
The opteron was WAY better than the older Xeons at crunching stuff in parallel processing (Hypertransport!!)
I got way more efficiency out of our opteron boxes than a theoretically more powerful Xeon box. Intel has now picked up that gauntlet, by essentially copying the hypertransport idea.
Some nuance I think is in order.
Netscape became a little naff because MS gave away the browser, so Netscape had to follow and if you are in a small business you don't put the effort into the item you sell for nothing. Netscape switched emphasis to the Server, but eventually Apache took them there.
Mozilla foundation, was a different approach, and Open Source from the beginning, which allowed them a better business model. Netscape just got caught on the hop, both by MS and Open Source.
It is really pathetic when someone in sales tries to be technical.
Having worked in the software side of an R&D company, I'll say that trusting a sales person on technical issues (or technical competency) is a bad idea. They will all too often promise customers what they have been told the company cannot deliver in the time stated. Then, when the technical folks say they can't do it, or are forced into doing some half-baked kludge, they get the blame for not producing what was "promised to the customer".
What has AMD done?
- produced and shipped true 64 bit X86 processors. Intel was forced to play catch-up and had to copy AMD's instruction extensions to be compatible.
- produced and shipped processors using integrated memory controllers, which was a big part of Athlon's performance advantage. Again, Intel had to catch up
- produced early true multi-core X86 chips. Intel responded by squishing two chips into one package, and only later had true dual-core chips
And remember, it was Intel that had the infamous floating point bug in their processors, not AMD. (Virtually all CPU chips have bugs, but you usually find out about them from the manufacturer, not by discovering incorrect results!)
don't forget Hypertransport->Quickpath
- produced and shipped processors using Hypertransport, which was a big part of AMD64's advantage for serious system builders. Again, Intel had to catch up, eventually introducing Quickpath, which as yet has failed to be Intel's promised "Common System Interconnect" (common between x86 and IA64 families).
Any other suggestions?
[yes I'm well aware that Hypertransport isn't entirely an AMD invention]
Don't forget Intel's achievements
- Late DVI: Intel stuck with analogue VGA when everyone else had DVI. A blurred picture on an LCD panel was a strong hint that it was connected to an Intel motherboard. The high speed A to D converters required by LCD panels to connect to Intel chipsets must have been a significant cost.
- RDRAM: High latency high power high cost memory covered in patents. RDRAM was intended to remove competitors to Intel's chipsets and had no advantages for customers.
- Memory Translation Hub: Rushed out to allow Intel CPU's to talk with SDRAM because of the limited supply of RDRAM. If it had worked, the latency would have been crippling.
- i945: Tolerable when it was first released, but stayed in the market long enough to make Atom look like a power hog. Somehow it got listed as a suitable chipset for Vista.
Doesn't matter, it's all about the execution
Yes, AMD got 64 bit chips and the integrated memory controller out first. Pity that AMD failed to capitalise on it then, isn't it? Over five years of memory controller lead and AMD threw it all away.
Multi core is completely irrelevant. In theory the AMD solution should have won - in practice Intel's performance was far better. The same pattern repeated itself for years - what appeared to be a technically superior solution had a subpar implementation.
It doesn't matter if your competitor's solution is technically inferior if their actual functionality is greater. The only exception is when your competitor paints themselves into a corner and needs to completely re-architect. Intel did that with P4 - AMD failed to respond. Nvidia now appear to be failing with endless rebrands and AMD (ATI) is for once reaping the benefits of a superior architecture.
I'd been an AMD fan for many, many years, starting with their 386 clone chips back in the early 90s. Up through the last new machine I bought 4 years ago I stayed with them; the only Intel chip I bought was in a laptop where I didn't have a lot of choice.
This year though, I decided it was time for a new machine, and there was clearly no choice - I went with a Core2Quad. The thing smokes, the CPU and chipset are rock solid and the price was very reasonable. Sadly for my old AMD-lovin' self, it's probably Intel for me in the future.
Fail because AMD seemed to just stop working hard and allowed Intel to just leave them in the dust.
But AMD made better CPUs than Intel in 2004!
I could understand if he'd made those comments in 1998 with K5/6 v P2/3 or in 2006 after the Core 2 was released, but in 2004 AMD were ahead of the game. The Athlon, Athlon XP and Athon 64/Opteron were a series of winning designs that out-performed Intel's rival CPUs and offered better value for money. People voted with their wallets and AMD steadily gained market share over those years, particularly in the gaming / enthusiast then later in the server markets where buyers had some technical insight. Whether this was due to AMD's brilliance or Intel's incompetence with the architectural disaster that was the Pentium 4 is debatable.
Today is much like the K5/6 era again, where the AMD product is slightly inferior but usually offers better value for money. But I doubt Intel will allow themselves to be eclipsed again as they were in the Pentium 4 era.
Relevant article: http://www.emulators.com/pentium4.htm
Wasn't 2004 the year Athlon64 embarrassed the P4
Mr. Richard should have been fired, as the Athlon 64 was released in 2004. As everyone knows, this product was superior in performance and technology to the Pentium 4s of the day that did little more than generate heat.
Why was he not canned immediately for ignorance of the competitive landscape? How can you sell a product if you don't know THAT it is or HOW it is superior to the competition? How can you be Head of Sales and speak with such disdain and ignorance about your employer and the products they pay you to sell?
I think what this illustrates is that AMD harmed their own competitive position by employing this moron, not that AMD wasn't producing a comparable or superior product.
Athlon 64 released in 2003
To correct an earlier comment I made: The Athlon 64 was actually released in 2003, making me question even more what Mr. Richard is basing his 2004 comments on.
AMD's 15 minutes on top
AMD did have its moment in the sun but couldn't hold it. With the Athlon XP processors, gamers were free to overclock the hell out of them and they were happy. Their chips were inexpensive so system builders were happy. I could also swear to you that Windows ran snappier compared to Intel's P4s due to different cache configurations. Intel wasn't fazed since it had brainwashed the world with "Intel inside" for a decade so people still considered AMD chips cheap knockoffs but the teenagers (read: gamers) who were asked by their family for "cheap but good" PCs steered the hoi-polloi towards AMD.
This gave AMD the economic breathing room to put more money into R&D and what we got was the Opteron and the x86 multicore-on-the-desktop "revolution" that is still playing out (though now to Intel's tune).
I remember how AMD was patting itself on the back as it was telling anyone who would listen how its Athlon (MP/XP) architecture was built from the ground-up to be multi-core while Intel's early entries were sort of slapped together extensions of the Pentium with no integrated memory controller and so on.
In the server space, Opterons gave people the alternative they were looking for when Intel tried to shove Itanic down everyone's throat thinking "What are they gonna do? Not buy Intel?". Too bad AMD ended up blundering the 3-core Opteron while Intel had already realized its mistakes, changed its course, was already throwing out 4-core chips. What was AMD doing? Eyeing ATI for an acquisition hoping to get some sort of performance synergy in gaming and design workstations by controlling the CPU and GPU. Still waiting for it.......
Can't say AMD was always crap. Just rested on its laurels and made bad business decisions...... and maybe some bad technical decisions too..
BUT If it wasn't for them, you'd all be speaking IA64 now!!!!!!!
I've been sick, so thoughts not as coherent and concise as they should be
Trust a sales person on technical Issues?
No sane person would.
I'm betting he wasn't allowed to offer the same kind of ridiculous palm-greasing money that Intel were willing to dish out so sales were harder, his commision was affected and he was pissed off about it. Enough evidence has been shown that is exactly what Intel did, they have been convicted and should just deal with it and move on.
As for AMD falling behind - if Intel hadn't carried out anti-competitive practices they would have made more money and been able to pump more into r&d. It could have been a completely different story. Who knows what AMD would be producing now?
This is the point...
"As for AMD falling behind - if Intel hadn't carried out anti-competitive practices they would have made more money and been able to pump more into r&d. It could have been a completely different story. Who knows what AMD would be producing now?"
This is the point. AMD were better than Intel at one point, but Intel were the dominant player, so they forced the hardware manufacturers not to buy AMD. Thuhs AMD were starved of the necessary profit to continue innovating. If anything, the fact that AMD were better for a while and still didn't gain massive amounts of market share proves that Intel were engaging in market manipulation, and they really shouldn't be shouting about it.....
I nearly wet myself...
IBM's POWER competitive? It doesn't even beat that 'toy' cpu from Sun as IBM sales weasels call Sunacles T2+ despite having nearly 3ghz of clock speed advantage. Its garbage mate. I work with Power6 everyday and its a staggering disappointment.
You see, max the workload on a POWER, and it flounders as the hypervisor eats the monster cache it needs to even work, but leave it too idle and it just slurps all the power in your data centre.
"Mr. Richard described AMD as "pathetic" for "selling processors rather than platforms [as Intel did]"
He has described his company's sales direction as "pathetic". Come on, show of hands. Who *hasn't* thought some management decisions are pathetic? The company I worked for back in the UK made some seriously pathetic decisions. Plenty of them. Of course, I wouldn't go shooting off my mouth to a competitor, but althesame...
Note also that he's the top sales guy. And their "pathetic" decision is a sales-motivated one (selling processors, not platforms). For all we know this bloke has a God-complex and wasn't happy because the company wasn't doing things HIS way?
I, for one, would not welcome nor trust the opinion of an (as in one, singular) EX employee blabbing to the competition.
FWIW, I had an Intel Celeron 466MHz machine that was trounced (nearly 2x) by a PII clocking 450MHz. Why? FSB. The "slower" processor's memory is 100MHz, while the faster machine has 66MHz. Add that to the "quality" of the interfacing logic, the hardware used, and it becomes quickly evident that the slower machine seriously outperforms the one that you'd think would be faster. This was both in benchmarks AND use (just timing a DVD-to-XviD gave the same expectation as the benchmarks implied). Thus, sometimes Intel were better and sometimes AMD were better, but in the end there's quite a bit more to it than which processor you use...
It seems it was no mistake letting that salesguy lave , they should have kicked him harder.
It does not matter which chip is better or more advanced, INTEL abused it's position, used it bigger moneypot ( which it stole from unsuspecting unknowledgeldeable customers ) 2 wrongs dont make a right.
AMD is still better in theprice/performace ratio, and you get what you buy, usually the ppl who complain are cannot keep their CPU cool to let it perform well.
So much for progress and Ex-employee hates ex-company ...
@"Ex-AMD exec called own company 'pathetic'"
... Ex-employee hates ex-company, news at 11. So I don't see how the US Federal Trade Commission could see his comments as anything but very biased against AMD.
As for this ... "The Complaint seeks to characterize Intel as a technological laggard in microprocessors"
From a technical perspective, Intel's X86 design is a bloated, old fashioned, inefficient design. This is highlighted by the power differences with say for example ARM processors where cycle for cycle they are far less power hungry. But is that Intel's fault. Partly no, partly yes. No, in that PC's are tied to x86, but yes, in that Intel have made a huge marketing push trying to convince everyone X86 is the future, and so they have not looked into any way to move away from x86, when everyone else sees the future in efficient multi-cores, which Intel cannot deliver as their old design is bloated with 32 years (!) of legacy support. In an ideal world, from a programmer perspective, 1 super-ultra-insanely-fast CPU is all we ever need. But from an electronics point of view, there is no way to achieve that (currently). So while less than ideal, multi-cores seem at least a way forward. However since we all went multi-core, Intel have been placed at a major disadvantage by their old x86 design. Thats because for each core, they have to replicate all their bloated design, which eats up a lot of chip space for each core (and more power).
As for AMD, they have to replicate that same x86 bloated disadvantage as they have to emulation true compatibility with x86, so they are tied down to the same millstone, so to speak. They also slowed up efforts when they were winning in the move to 64 bit. That was a costly mistake, they should have pushed ahead more to hurt Intel even more and stay ahead. But its always easier for non-technically minded bosses to sit back and not spend money when the money is puring into their company. They so often have a short sighted view that fails to take into account reinvesting in the future technological growth of their company.
So anyway, are Intel holding progress back. Partly yes. Intel's progress is mostly tied to ever more process component shrinks, which is barely innovative, plus they are running out of time with that tactic. But beyond that, the game will move towards 3D chips where each layer is again adding cores that way rather than with component shrinks (which is harder to do as each bloated core is so power hungry). So from here on out Intel's bloated design is going to get multiplied ever more copies of it in the same chip. Thats hugely inefficient in terms of both power wastage and wasting available chip space. Its therefore no wonder GPU's are putting x86 CPU's to shame. For example at 40nm I can have 4 cores via Intel, or 512 cores via NVidia Fermi, yet both are at 40nm. It shows its possible to put a lot more cores on a chip when the core design is more efficient than Intel's current bloated design.
So is Intel's design bloated, and holding back progress. The answer has to be absolutely yes. But AMD also hurt themselves as well. I've worked for technology based and technology driven companies that have to suffer non-technically minded bosses. Its always been a nightmare of mistakes and lack of vision. That explains both Intel and AMD. I bet both are filled with too many executives playing their manipulative office politics power games, who are also not technically minded. These power game executive types seems to be a common malaise that corporations are afflicted with.
So much for progress.
Regarding some of the comments here, I have to say AMD had dual core CPUs in the market place for quite a while before Intel (with it's significantly higher R&D budget) caught up, and isn't the x64 technology still licensed from AMD? It certainly was for a while. Don't get me wrong I'm not particularly an AMD fanboi, but let's be honest both companies have their time as the lead, same with graphics cards etc. Currently I would say Intel is back in front with i7 and corresponding Xeon but that doesn't mean it will stay that way indefinately.
AMD failed to capitalize?
When major OEMs refuse to carry your product because a competitor has given them money to do so, how can you compete?
You can compete even at a disadvantage if the playing field is level. Yet this is not the case in the CPU market.
AMDs possible major earner, Opterons were not carried by one of the major OEMs (Dell, yes look it up in Wiki and corroborate with product launch dates) and you'll see how AMD couldn't capitalize on the server market properly when Intel, with their crappy Xeons managed to hold most of the server market.
Despite having a cheaper, more efficient product, AMD couldn't move enough product to earn enough money to invest into R&D.
Now, if AMD wanted to hit Intel right where it hurts, they should develop a very tight integrated platform and sell it cheaper than Intel. You see, Intel earns so much because they have food on several plates. NICs, IGP, Chipsets, CPUs, WLAN controllers and whatnot. If AMD had a platform that could do all that, they could offer a very attractive offering to OEMs.
AMD just has really bad marketing
AMD has always had a component to match Intel's at a price point, not everyone buys the top CPU. AMD is only beginning to realise the marketing potential using their dragon and spider branding of system.
The problem really lies with marketting, Intel has their logo, their tune, their men in toxic suits. AMD has had ... their logo? and that's it, they've had sponsorship sure, but sponsors don't really work unless its for gaming, since sports don't really correlate with PCs.
The main issue is Mr Richards and all his marketing and sales has probably sunk the ship by doing near ineffective marketing. If AMD spent more cash into supporting TV adverts and radio adverts like Intel has bombarding geek and non-geek alike with how they're computer is so great then yes, AMD would at least be known as a competitor to Intel, I can ask my mum and she would know that Intel has something to do with computers, the same can't be said of AMD and my parents watch lots of TV.
Now that TV is starting to go out of fashion, and more people are becoming knowledgeable about computers AMD is more widely recognised than before, but it is too late for that. The marketing department at the end of the day is what sells products, how many countless times when companies with sure fire good products get shot down because they didn't market their product effectively, look at sega dreamcast vs sony (sony were just too strong with their marketing muscle), psion, cherry keyboards, olympus, etc. Most of these companies just didn't have the marketing nous or the power, thus relying on simple niche features which aren't provided by the opposition, or lower price points.
With smarter buyers lower price points work, but really how many of these are there really? it is a growing population, but most people will go with tried and tested, for every good ipod there's 100 competitors, there is no real competitor on tv? maybe sony but they had an inferior products and music system, and they marketted too many different products.
At the end of the day it was probably Richard's fault that AMD did so poorly, hardly any coverage of AMD in the UK, maybe a sponsored football game, where you will hardly ever see the bill boards.
Also Richard's an executive of course he wants the latest and greatest, not all of us are so lucky to afford the stuff he can afford. He probably marketted all AMD products at really glitzy events with the people least interested in AMD because they weren't well known and also weren't known to be good
So Intel is saying their predatory monopolistic practices were ok because the victim's former head of sales said they had a poor product?
Is that supposed to excuse bad actions on Intel's part, or just make them look all the sadder because they couldn't compete fairly against AMD even though AMD's products were allegedly awful?
Gee cut the crap lads
Never mind the quaint technology history lessons, we know that AMD is in serious debt.
This kind of limits what the company can do I would have thought.
I buy AMD because that's all I can afford and quite frankly I don't need the odd ten percent or so extra performance that an Intel chip might show up in a benchmark.
This is reality knocking....
@eclairz and all other Apologists
"AMD just has really bad marketing"
They surely had something better in stock at the time they licensed the 286 design from Intel. It was just that their mktg dept was too shy to sell it. Also, they are the true inventors of the microprocessor, not Intel. Definitely.
Face it, AMD fanbois, Intel is the natural monopoly and they led the industry for 14 of the last 15 years. They have the most advanced semiconductor base technology and they have the best CPU and optimizing compiler designs. Their factories produce more high-end CPUs than anybody else combined. They killed DEC's, HP's and Motorolas CPU's and soon they will have done the same with SUN/Oracle.
Apparently it does not make a big difference what instruction set you implement as long as you have the deepest budget, best semiconductor base technology and the best engineers of the industry. X86 is only outperformed by IBM Power and these are only a threat in the Unix business.
Power consumption is a different story, but if it truely where the critical factor for the market, Intel would surely turn out something at least on par with ARM. But most customers care just about performance, so Intel focuses on that.
You conveniently forgot the console market ;-)
It's such a shame
For a short while, AMD had the hot hands. Intel were just force-feeding the pubglic higher and higher p4 clocks, with deep pipelines that were criplled every time branch target prediction bubbled. Heat and power problems with Intel stuff just made the mediocre performance more obvious.
However, not only did Intel wake the hell up, but AMD got smug and lazy. Intel's big and ugly enough to have loads of different irons in the fire, and the CoreN stuff was being developed quietly in parallel elsewhere (wasn't it Intel Israel or somewhere?). They were able to to pull it out of a hat, just as AMD were running out of steam, getting fat and careless.
I loved the Athlons, had a first gen one, and an Athlon XP. However, by the time I was ready for a 64 bit X86, the Core 2 duo was faster and lower power than the Opterons out there, sadly. I was all for rooting for the underdog, but having seen the Phenom mass fail (late, buggy and horribly underperfoming).
I buy quite a lot of eight core Nahalem boxes of late, as I need to do a lot of reliable brute force CPU-based HD video coding (don't ask). Every now and then AMD claim that their new Phenom variant doesn't suck, and I try one- and always feel disappointed.
Sadly, they just feel like shoddy second-rate gear, now. I also can't buy their GPUs, partly because the default colour reproduction is so horrifyingly off (like TVs set up by a Dixons salesman, and I don't trust end users to manage their colour) and partly because their Linux support is unforgivably piss poor.
It's a shame- I used to root for AMD, but now they seem like a shoddy distraction. The only situation that I'd buy their stuff is if I needed a lot of cheap puny CPUs for massive high performance computer clusters. Sorry, AMD. Sort your CPUs out, write some decent fscking Linux support, and cut the bullshit, and maybe the story would be different.
The guy's a nutter!
He said this in 2004. Now, in 2004, AMD had released the Athlon 64 and Intel was pushing the Prescott P4. In fact, 2004-2006 is probably the time when AMD had the biggest performance and technology lead over Intel for quite some time.
This guy is effectively saying back when AMD had released an ass-kicking new product and Intel were pushing an old, hot, slow and inneficient design that "If you look at it, with an objective set of eyes, you would never buy AMD. I certainly would never buy AMD for a personal system if I wasn't workin here."
What a pile of shite! I looked at it objectively at the time, saw how much better the Athlon 64 was than the lousy P4 and promptly bought an Athlon 64 for my home PC.
Sure things have changed since then, Intel has kicked the P4 into touch and has the performance lead and I'm now running a Core i5 PC. However this guy must have had a screw loose when he originally made these claims!
- Boffins attempt to prove the UNIVERSE IS JUST A HOLOGRAM
- China building SUPERSONIC SUBMARINE that travels in a BUBBLE
- Review Raspberry Pi B+: PHWOAR, get a load of those pins
- Experimental hypersonic SUPERMISSILE destroyed 4 SECONDS after US launched it
- Will It Blend? Maybe. BlackBerry’s secret comeback weapon