Howdy, stranger! Ready to join the community? [log in]

Buzzkill: AMD’s fall from (good) grace

AMD has lost it. Not the performance crown or the market share. I’m talking about the people; they aren’t thinking about AMD any more.

My role at Icrontic is to keep my finger on the pulse of what the enthusiast community is interested in. Timely delivery of relevant news requires a war chest of tools that includes big RSS feeds, direct site visits, emails and Twitter. On any given day, you’ll find Icrontic carries news and articles that have been sourced from at least one of these channels.

Is this really the new face of AMD?

Is this really the new face of AMD?

Twitter is easily my favorite of the lot. The service offers a living view of the topics that scads of people are actually talking about, and I can watch that buzz in real time with Monitter. On any given day I might be watching the conversations for over 20 keywords that relate to the PC industry.

Last weekend I hovered over the Monitter dashboard and waited for a lead I could polish into an update. Instead I got a lesson in AMD’s mounting obscurity as my AMD column filled with OEM sales and typos of the word “and,” while the Intel column was overflowing with SSD, Nehalem, 32nm and Intel STS discussions

Wait for the Intel/GlobalFoundries brouhaha to clear and try it for yourself. Use Monitter to configure an AMD column and an Intel column. Watch what happens. AMD is out of sight and out of mind. That’s quite the change from 2004’s wildly collective nerdgasm for the Athlon 64.

There was a time when AMD could command premium on its processors because they were the best in the industry. The Opterons wrecked the Netburst Xeons in every metric, and the situation was no better for Intel when it came to the desktop. While Intel languished for six miserable years under the rays of their cosmically bad Pentium 4, AMD took center stage and kept it for some two years.

So just what the hell happened? Did ex-CEO Hector Ruiz milk the Athlon 64 for so long that AMD was oblivious to Intel’s driven urgency? Or did the much-delayed and ultimately disappointing Phenom force their hand? It also didn’t help that the Phenom came out swinging against Intel’s Core 2 architecture, unarguably the most robust and successful slice of silicon the firm has ever fabbed. We also know that AMD vastly overpaid for ATI which then failed to yield expected profits. And lastly, we can only surmise that AMD’s financial woes have made it incredibly hard to keep pace with a rival that can afford to burn greenbacks to keep warm in the winter.

This is not another article that foretells the demise of AMD, but hell if it doesn’t hurt to see the underdog come so far and fall so damn hard.

I suppose my recent experience with Monitter put AMD’s current situation into perspective. It’s difficult to visualize corporate struggle when expressed in losses, filings and stock prices. In contrast, it’s really quite sobering to realize that the grandpappy of x64 architecture is barely worth discussing.

AMD’s disheartening inconsequence shows little sign of abating over the next two years. Intel’s plan to release the 32nm family of Westmere processors by the end of 2009 will push AMD further into the background. In fact, AMD isn’t even competing at 32nm until 2010. By the end of 2010, Intel will already be talking about their brand new Sandy Bridge architecture. AMD’s new architecture, known as Bulldozer, isn’t expected until 2011.

The enthusiast community is a pragmatic bunch that deals in the absolutes of “first” and “best.” Intel will probably get there first. That doesn’t leave much room for buzz in the AMD camp.

But not all is downpour and brimstone. AMD has orchestrated a masterful revival of the ailing ATI brand. Not only has the Radeon HD 4000 series exploded into a potent contender, it has become incredibly popular thanks to pricing that brought reason to the entire market. Not content to rest on its laurels, AMD is already prepping a speed bump and a refresh all while NVIDIA tools around with silly name games.

AMD also has one hell of a chance to regain the word of mouth come 2011. Successfully launching a Bulldozer core that sends Sandy Bridge packing will create a groundswell of good faith. But Big Green can’t stop there, and it must take that momentum through to 22nm by year’s end. If AMD fails to die shrink on time, I could change the dates and publish this editorial all over again.

Most geeks — like me — have a soft spot for AMD in their heart. It pains us when the reality is that AMD doesn’t offer much to talk about that doesn’t begin with “well, if you’re on a budget….” These are dark, dark times for AMD. I never really noticed how bad these times were until people even stopped talking about the darkness.

Share |

53 Comments:

  1. AlexDeGruven
    I am Geek. Hear me... type?

    Hopefully, this current lull is just something like what happened with Intel for a few years. I'd like to see the Bulldozer come in and kick all kinds of butt like the Opteron did, but they need something really stellar to do so.

    I also see AMD pushing a lot of their focus on their ATi counterpart. Maybe they're working toward conditioning the market for an eventual halt in x86 space. If AMD feels they can't continue to fight Intel, positioning ATi to solidly take the crown from nVidia seems like a viable alternative.

    Who knows, though. It's really all speculation until the press releases come out.

  2. Robert,

    For full disclosure, I would describe myself as a decade long AMD/ATI fanboy. For the sake of argument here, let me limit it to the home user experience, and leave the business platform argument for someone a little better versed in that topic than myself. My love for the AMD brand started in 2000 with my purchase of a slot A Athlon based system. That moment in time, in my mind is when the real two headed microprocessor design war began between brands Intel and AMD. (Intel more or less having a home CPU monopoly preceding few years)

    AMD proved something with the slot A Athlon. They proved it took more than raw frequency to make a great performing chip. The overall design, how that chip is married to the system as a whole is just as important, and through innovation on system bus and memory architecture, along with better frequency they had manufactured a product that was in each and every way, technically superior to the Intel Pentium III at the time, and it took Intel a full manufacturing cycle, and a boat load of marketing revenue for three blue men to stay in the game.

    Why is this history lesson important? Its because its the exact point in time that real market competition spurred innovation and better pricing for consumers. Both sides playing against each other, to tap thier resource to build a better chip, faster, cooler, more performance per watt. Without AMD, I honestly think we are all still paying $1000+ a chip for something that performs like a Pentium 4. Hard to say for certain what the level of innovation would have been like if not for AMD sticking it out as the market underdog, but one thing is for sure, Intel's margins would be much higher, and computers would probably be in far fewer homes, businesses and schools.

    Since 2000 AMD has contributed much to the microprocessor market. 64 bit instructions, first true multi-core design, unlocked multipliers for enthusiast overclocking, Integrated DDR2 controllers just to name a few.

    I build systems on the side for friends and customers. Nearly every single one will ask for an Intel based system due to brand familiarity. When I educate them on the value of AMD's product offering, especially the value of platform computing with AMD's motherboard chipsets, CPU's and ATI graphics all combined in a tidy bundle that plays nice together, they always say yes to AMD, and report they are thrilled with thier systems as they use them.

    The problem may be marketing to a degree, or just a uneducated consumer, either way, AMD/ATI does have the tech, and a solid overall battle plan for bringing users chips at a value that is currently hard to believe. Its never been so inexpensive to build a powerful computer, and we have AMD to thank more than any other company in the tech landscape.

    Lets look at some AMD/ATI branded products that the market should be very excited about.

    Lets start with motherboards. Fact is, there is no better IGP for basic productivity computing or HTPC than the 780G or 790G that uses either the ATI HD3200 or HD3300 as way to deliver affordable HD graphics performance to the masses. The ATI IGP is so good you can actually run respectable 3D games on it. Sure, you still need a dedicated card to play the latest and greatest at the highest resolutions, but the 780/790G chipsets are a huge technical leap forward for IGP.

    Inexpensive enthusiast level CPU's with unlocked multipliers for overclocking. Currently you can buy the 7750 Black edition dual core for $65, The triple core Phenom II 720 BE for $145, and the crown jewel 940 BE for only $225, All of these CPU's offer enthusiast end users the kind of tweaking freedom they crave when coupled with AMD's in house developed overdrive software the experience is even better. I am personally using the triple core 720 BE in my home system, and I have to say, never in my wildest dreams would I have ever thought that real enthusiast level performance be so affordable.

    Another strong point for AMD is the ATI brand. I won't argue for or against the price of the brand, but what I will say is it positions them with who I believe is the best and most innovative producer of graphics products in the world. History lesson there, remember a little card called the 9700 pro released back in late 2002? Remember how it more or less doubled the performance of any competitive product offering from Nvidia at the time? What do you think got Nvidia spinning its wheels to produce better more innovative products? Its that moment, the 9700 pro is to the graphics market what the slot A Athlon was to CPU's. Those product releases shaped the home computing market to create the innovation and pricing we all enjoy now.

    That being said, the 48xx series of graphics products is currently the best value dollar for dollar on the market. With rebate you can buy a GDDR5 fueled 4870 for about $155 right now, that's right, thanks AMD/ATI. You have to spend nearly $100 more just to run side to side with it in a comparable Nvidia offering, and frankly, even then the GTX260 is not the clear performance winner. Not to mention, striped down, but great performing 4830's are showing up on-line for about $75 after rebate, once again, thanks AMD/ATI. To perform along side the 4830 you more or less need a 9800GT at 25% more cash on average, and then the 4830 is still going to win head to head, pixel for pixel in most contests, so better tech, less cash, its a no brainier.

    Also, soon to be released, 48xx graphics in performance notebooks!! What's not exciting about that!!

    AMD/ATI, have always taken the position as the market underdog against two brands that have a great deal of marketed consumer loyalty. Just like consumers overspend on Apple products that lack essential features due to sheer marketing might and brand recognition, people do the same when they consider Intel and Nvidia. They are just the defacto brands that have perhaps marketed a little better, and to be fair, yes, Intel gained a performance lead in the last year on the products in the higher pricing range, but if you still compare them dollar for dollar, the value winner is AMD.

    What I ask my friends and fellow tech enthusiasts, isn't it worth rewarding a hard working group of people for providing you with a competitive market that offers everyone a far better value? If the chip still more than suits your needs and costs far less than brand blue, you should spend your hard earned there. AMD/ATI has plenty up thier sleeve, the perception that they somehow make products that are vastly inferior to Intel is not a fair observation looking at the entire picture. Core2 Duo vs. the original Phenom, okay, got us, Core i7 vs. Phenom II, got us, sort of, but not dollar for dollar when you compare a new system build. A comparable Phenom II based system costs hundreds less and actually offers gamers a better platform frame for frame when coupled with Radeon discrete graphics.

    So as users, are we saying the only thing we value is the short term raw performance winner? If that were the case we would all run out, spend $1,300 on the unlocked i7, but that's obviously a realy awful value dollar for dollar when compared to something like the AMD Phenom II unlocked CPU's. You could spend $529 on a Nvidia GTX295 (if you can find one), or get a 4870X2 which is a far better value at about $120 less right now.

    What I am saying is this, value, should still have value, especially right now with the economy soft as it is. Who do we feel is winning the value race? Who is offering tech consumers more for thier hard earned dollars? I don't see how any educated person could answer anything other than AMD for motherboards, and CPU's, and ATI for graphics.

    Join me, don't settle for being one of brand blue's lemmings, support a superior value dollar for dollar. Long live AMD.

  3. First of all, thank you for commenting, Cliff. I just knew this op-ed would put a bee under your bonnet.

    I have a few quibbles with some things you pointed out:

    As I mentioned, the ATI has really done a job on the GPU market. Yep, they've brought prices down and brought the performance profile of "budget" up, but your 4870X2 vs. 295 comparison is just unfair. The GeForce GTX 285 is the X2's actual competitor, and it's a full $50 cheaper than the 4870X2 (Newegg). Not only does the 285 frequently provide better performance, its losses could be placed within the margin of error.

    You also suggest that people should spring for the Phenom II just to reward a "hard working group of people." There's no denying that AMD has toiled laboriously, but toil doesn't win my money. The best performance does. I cannot buy into AMD's message with the Phenom II because it struggles to compete against the Kentsfield, an Intel core that is two generations and two years old.

    If we're going toe to toe on enthusiast performance, an unclocked multiplier doesn't mean anything. No sane overclocker is going to pony for the $1000 i7 965 when the Core i7 920 is just $288, or only $60 more than the 940 BE. Did I mention that the chip can clock up to 4GHz without much effort? Add in the Intel X58 premium of about $60 (over comparable AMD 790 boards), and the pricetag for a Core i7 system is only $120 more*.

    What does that $120 get you? A socket with a future. Guaranteed support for a die shrink to a faster, colder chip. A more modern chipset. Superior system and memory bandwidth. Better FPU performance. ESPECIALLY better multimedia performance.

    The point I'm trying to drive home is that the word best isn't all that subjective. If $120 is nothing to you, the Core i7 is a superior CPU and platform. If you're on a budget, then the Phenom II becomes the best. It all depends on your needs.

    There's no denying that AMD is trying very, very hard to stay relevant. They're doing the best they can, but their best is not the best the market can do, and that's what the people who generate the buzz want to talk about. Few people wants to spend all their time being excited over the guy who is only in the lead when people have to pinch pennies.

    *NOTE: I did not include the price of an upgraded heatsink because both CPUs need improved cooling to achieve the clockspeeds that the Phenom's unlocked multiplier or the Core i7's bclock scaling can offer.

    //EDIT: PS, I thought the Pentium 4's price system was horse crap. Thank you, AMD, for making the market reasonable.

  4. drasnor
    124 Golden Eye Drive
    History lesson there, remember a little card called the 9700 pro released back in late 2002? Remember how it more or less doubled the performance of any competitive product offering from Nvidia at the time? What do you think got Nvidia spinning its wheels to produce better more innovative products? Its that moment, the 9700 pro is to the graphics market what the slot A Athlon was to CPU's. Those product releases shaped the home computing market to create the innovation and pricing we all enjoy now.

    I remember the 9700 Pro. I remember RMA'ing my first-party All-in-Wonder three times in as many months. I remember waiting 5 years for working drivers for it on Linux. I remember switching back to my old GeForce 3 Ti500 because at least it could manage a playable framerate in Unreal 2004. ATI would have to be *giving* away cards to get me to try their product again.

    I have owned a dual Opteron 248 workstation with the original Sledgehammer chips. At the time, no one was talking about the possibility of dual-core processors and it seemed Netburst would never go away. I built NF7-S 2.0's with Athlon XP's for friends and family. At the time, they were the best products on the market.

    Today, all of my machines are Intel boxes. When it came time to replace my dual Opteron workstation I discovered I could build an Intel Q6600 machine for a little more than the cost of a replacement motherboard for the Opterons. With the Phenom launch bugs and frequent AMD socket changes I didn't have any real desire to buy into another AMD platform.

    The choice was even easier when it came time to replace my dual Athlon MP HTPC/LAN gaming machine. None of the AMD offerings were workable with the thermal and power constraints of my HTPC case. The Core2 Duo E8400 w/ GeForce 9600GT combination I eventually settled on has an idle power footprint of 75W according to my UPS. Top that with AMD/ATI hardware.

    What I ask my friends and fellow tech enthusiasts, isn't it worth rewarding a hard working group of people for providing you with a competitive market that offers everyone a far better value? If the chip still more than suits your needs and costs far less than brand blue, you should spend your hard earned there. AMD/ATI has plenty up thier sleeve, the perception that they somehow make products that are vastly inferior to Intel is not a fair observation looking at the entire picture...

    So as users, are we saying the only thing we value is the short term raw performance winner?

    I want the best platform for my needs that my money can buy. Right now AMD isn't offering that. If I were interested in funding hardworking people providing me with a competitive market I'd still be a Mac user.

    -drasnor

  5. QCH
    Guru

    Rob, Good write-up... I really hope AMD has something up its sleeve that can make it "hip" and be the buzz again.

    Cliff, thanks for the well thought out and articulate response. I am still an AMD fanboy but it becomes harder and hard to stay that way when prices are close enough to make PC's under $600.

  6. Sledgehammer70
    Sledgehammer70

    I made a huge post and decided to not post it...

    In general I think the talk of AMD in twitter has shifted from "AMD" to other keywords. The same is true with Intel as you will most likely still get more results from intels sub words also. but the community and fanboi's are there... I know it as I converse witht he CPU side all the time but stay away from the GPU side

    NV For Life

  7. Believe me, searching the sub-words are no better:

    I have a daily column for the following keywords (as relevant to this discussion):
    AMD, ATI, Phenom, Phenom II, Agena, Bulldozer, Shanghai, Intel, Gulftown, Westmere, Nehalem, Core i7, RV790, and a long list of NVIDIA GPU code names.

    Believe me when I say that the buzz just isn't like it used to be. It's sort of dead. And it's sad.

    But this points to another thing. The "Intel" was alive and well. What's wrong with "AMD" if it's not putting up the volume? Is it a branding problem? Is it the "second place" phenomenon I have described?

  8. Good comments,

    I think we are down playing the cost of an i7 based system a bit when compared to a Phenom II. I am going to take some time, throw some prices together, see if I can draw a dollar per benchmark value comparison for us and share my findings.

    In graphics, history to say the 9700 pro did not force Nvidia to innovate would be simply wrong. Nvidia purchased 3DFX in hopes of creating a one stop shop for all advanced graphics. What would you be paying now without ATI's interference, how far would we have come since then? To be fair, ATI's customer service dept left alot to be desired back then, but the merger with AMD has improved them in every aspect, better drivers, better partnerships with board makers, better overall quality, ATI is a better brand because of AMD.

    In closing I leave you all with a simple question? Why can you build a powerful machine for far less today than you could imagine? I think more of it has more to do with competition than it does with technical advancement. The way I see it, AMD by sticking out a tough battle as an industry underdog has protected us from a big blue hardware monopoly, that would cost us all in price, performance and innovation.

    Like that i7 chip you can get for around $290? Don't forget to thank AMD in some part for that. Come on, you know Intel is not cutting its margin short because they like you.

    Look the software market, operating systems and office software. Everyone on earth knows that a revamped word processor, or an OS that is buggy as sin at each release should not cost what it does from Microsoft, but that is the cost of a monopoly. Is that what you want for the hardware market too?

    I will not argue against the innovation of a platform like the i7, but it takes a serious power user to see the benefit, and even then, I think its a market condition to just want the brand that more people associate with. Can we agree that for 90+% of the users out there, a good dual core, or reasonably priced Phenom II will be more than enough performance to satisfy them if not totally knock thier socks off? Are we are all better off with AMD in the market to protect us from a price monopoly? I think so.

  9. Sure, I completely agree that today's pricing landscape can be directly traced back to the Athlon 64. I can also completely agree that even a lowly dual core is more than sufficient for most users. We can also agree that AMD's evaporation would kill the CPU market.

    But both all of those change the direction of the discussion. Nobody is saying that AMD hasn't moved mountains in the realms of price and quality... What I am saying is that for all the earth they've moved, they're still coming up short in the very important game of public discussion.

    They can shovel inexpensive and/or bargain parts until they're blue in the face, but they're just not being talked about in the way Intel is. It's not just brand loyalty. People want to talk about Intel because they're in the lead, and everyone loves a winner. The enthusiast industry thrives on it.

    The blog-loving enthusiast cares about the game benchmarks that the i7 trounces. What's going on in Crysis? World in Conflict? PREY? They don't care about the crusty channel benchmarks that show AMD's performance/watt, transactional or platform latency advantages.

    I'm not saying they're tiptoeing back towards the utter irrelevance of their K6-III era, either. They're too big for that now. The Athlon XP, even though it was a big loser in the end, was still 6-10x less than a Pentium 4. It found a home because so many people couldn't stomach an $800 price discrepancy. But what is $120? It's a pittance to go from #2 to #1. I wish $120 would take me from my Cobalt to a McLaren.

    All I am saying is that AMD is fighting a big uphill battle when it comes to public perception. They need a solid win in the CPU market, and their GPU market isn't big enough to pick up the slack. People talking about you is a sure sign of your health, and AMD just looks sick.

  10. MiracleManS
    Mediocrity Gets You Pears

    I really like the discussion here, but part of me really believes AMD is sitting on the idea of the CPU/GPU integration with whats going on. I really feel that they can make a huge dent in that market if they do what I HOPE they're doing: Bring it all into a cost effective package that the mainstream user (not the enthusiast) can purchase on the cheap.

  11. Intel will beat them by 2 years on that, too. Westmere in '09, Fusion from AMD in '11.

  12. Robert,

    I agree that AMD has some work to do to improve the overall market perception of the brand. I actually think the Intel cross license legal threats give AMD a huge window of opportunity to gain press, and expose Intel's plot to control a monopoly share in the processor market. In the end, AMD is going to show what a scrappy competitor they are and gain some significant market share. (I don't think for a second that AMD is going to loose x86 rights per the agreement).

    The enthusiast market is but a small part of the overall computer user landscape. AMD/ATI has been attempting to market better there, and their message is right, its not not reaching. Define "enthusiast" the computer world, that's normally just another word for "gamer", or perhaps "videophile". Where do get the performance gain for gaming and video? We all know the secret sauce is in the GPU, and the Radeon 48xx line whether you pair it with an Intel or AMD platform will deliver similar frame rates in high res, high detail gaming, so how much does the CPU really matter to that user except it raw marketing perception? And ultimately, once your running the game beyond your monitors refresh rate, whats that really matter? Radeon parts get you there, why not build it on a reasonably priced AM2 platform, save some cash, run the same frame rate at a guy who spent hundreds more, buy some extra games, a bigger monitor, something that provides actual value to the "enthusiast"/ aka gamer.

    The GPU is the real power part now, the CPU is secondary in performance.

  13. Snarkasm
    The Photographer.

    I'm an enthusiast, and my primary work for that power is in photo editing. The memory bandwidth, cache size, clock speed, and number of cores matter intensely to me in my work - it cuts an HDR render from 40 minutes to 5. The GPU is great, but it's not the end-all for damn sure.

    If your argument here is to recommend AMD for people that won't know any better for their requirements, I won't mind agreeing with you. If, on the other hand, you're asking performance enthusiasts to pay insignificantly less for second-place parts because their maker made the current landscape consumer-friendly, well... to that I simply need to laugh. I'm afraid reasonable, intelligent consumers don't purchase based on past (still profit-driven) benevolence.

    I will buy Intel until AMD equals or surpasses their performance or Intel's price becomes a premium I'm entirely unwilling to pay. It's pure, hard economics.

  14. Mochan
    Geek Monster

    Sorry Cliff, I like that AMD helped bring down prices and spurred Intel into action, but I'm not going to buy an inferior product put of pity for an ailing company, or as an investment that someday we will get a fully competitive market with no monopolies. At this moment in time, you'd pretty much have to pay me to get on the Spider platform instead of riding the Intel wagon.

    I have nothing against AMD and actually I would like very much for them to get back in the game, but I'm not going to do a bailout for them.

  15. Mochan
    Geek Monster

    Also that part about the GPU being the power part is nonsense even from a gaming perspective. I'm a power game and let me tell you this: last year I had an Opteron 180 and an 8800GT. I was getting framerate dips in games like The Witcher, Jericho and even The Sims 2. I upgraded my mobo, RAM and CPU so I could use an E7200 but I kept my 8800GT. With the exact same card, the only real difference being the move from Opteron to Core 2 Duo, my hiccups went away and I am one happy gamer.

    Sorry, but the CPU is not some secondary piece in the performance puzzle, it's equally important. In fact, I'd say that most rigs today are CPU-limited because GPUs have become so powerful. That makes the CPU ever more important in the equation.

  16. Snarkasm,

    I knew as soon as I said "enthusiast" was just a code word for gamer, someone would come along and say, I'm a profesional photo editor. I was ready for this.....

    So you can thank AMD for ushering the 64 bit era so you can use all that ram, that is the key to high performance in photoshop and the gimp.

    Moving on to Mochan,

    An Opteron 180 on a socket 939 platform is not comparing current tech. Your point perhaps valid if taking me 100% litteral, but obviously I meant to cross compaire current designs. Take a Phenom II, any Phenom II for that matter (the $115 710 even), take that i7 920, run the same 4870 graphics card, run the same timedemos, make sure to turn on the bells and whistles, run in perhaps at least 1440x900 (probably less even), you are going to hit the bottleneck at the card, the framerate is going to be the same. Crossfire that 4870 now, you have a whole different ball game on either platform. Your right to a point, I can't take a three year old CPU and memory archatecture and not choke that monster card, but if we are going to compare fairly recent comperable tech, running on a similar memory platform, the GPU realy is the glory part in a good gaming benchmark. Those low res benchmarks they run against the CPU on hardware websites mean absolutely zero, zip, nothing, I hate that review sites still do that, its an empty meaningless benchmark, if your playing in 800x600 low detail just to test your frame rate, your wasting time, its not a valid benchmark. What people realy want to know is how it will game with the details on and in a HD resolution. Real world performance is what matters, and the Phenom II has plenty of it for far less cash.

    If you have seen any of AMD's marketing on the topic, the question is, what will you do with the extra cash, some new games, speakers, a bigger monitor, its up to you, but any way you slice it, the combo of a Phenom II with any of those items offers a far better value.

  17. This calls for an article where we bench a 940 BE against an i7 920 with a 4870 X2 and a 285 stock settings and at maximum achievable overclocks. We can see who wins and give a $/FPS or $/whatever metric to show the REAL value.

    ...Now if only we could get test hardware from blue. Hmm.

  18. Mochan
    Geek Monster

    Hey Cliff, you ought to register for an account here, the rate you're going you'll have posted more than some members (like me). :P

  19. Mochan
    Geek Monster
    Take a Phenom II, any Phenom II for that matter (the $115 710 even), take that i7 920, run the same 4870 graphics card, run the same timedemos, make sure to turn on the bells and whistles, run in perhaps at least 1440x900 (probably less even), you are going to hit the bottleneck at the card, the framerate is going to be the same.

    Well duh, naturally, because you hit the GPU bottleneck. Of course the framerate would be the same. When you bench a CPU, you make sure there is no GPU bottleneck!

    You have to bench a CPU at low res in order to avoid the GPU bottleneck if hi-res is what is causing the bottleneck, because otherwise you are not getting the real picture in the Phenome vs. Intel benches. The fact is, 4 years later you may still have the same CPU and decide to upgrade your GPU, that is when the "hidden potential" of the CPU comes out. Assuming their system wasn't CPU-limited (which, again, is the case with the majority of rigs in the Core 2 Duo/Phenome era) people who bought Core 2 Duos are going to have more "hidden potential" than a guy who bought a Phenome. The fact is, the Core 2 Duo is just faster. Plain and simple.

    And again, that is a moot point, because even with "just" an 8800GT any Core 2 Duo or Phenome 1st gen is typically the bottleneck. My friend has an 8800GT same as me, the difference in our rigs is our CPU. He has an E8400 and I have an E7200. Guess who's kicking ass in the benches? Again, CPU-limited. Put up my E7200 against a similarly-priced Phenome, take a guess who is more CPU limited?

    I am not too familiar with how the current Phenome 2 or i7 cores are doing paired with triple SLI or xFire configurations, but again the CPU is not to be underestimated in the gaming benchmark.

    I do not know if the Phenome 2 is the better value but even if it were, it's only the better value "now" ... how about a few years down the road when GPUs are being upgraded? I bet the Phenome 2 isn't going to look like such a great value by then.

  20. Zuntar
    Modder extraordinaire

    All I can say is that if all I have to wait is 20 seconds longer when I encode a movie(like once a month at best), or 20ms when I open Firefox, then yea, I will gladly save $150 on a build. I can't afford Intel platforms. We don't all drive high end cars for the same reasons.

    Intel has brand trust for the masses, and AMD has lost theirs.

  21. Mochan
    Geek Monster

    Not all Intel rigs are monstrosities. If you can afford a Phenome, you can afford a Core 2 Duo. And the Core 2 Duo generally will overclock more and is faster.

  22. I will heed the suggestion and get a proper account here shortly,

    Mochan, The point I want to make is this. If the general term "enthusiast" does apply mostly to PC gamers, and the benchmarks we obsess about, then what value does that low res benchmark to test the CPU limited frame rate realy honestly have? Are you going to bump down your details and resolution just so you can run a game well past your monitors refresh rate? I doubt it, your gonna play it with the eye candy you paid for, with your nice graphics card. Sure, there are limits to what a lower end CPU can feed, I would not buy a Sempron, or Celeron to pair it with a high end 3D card, but once you get into a certain arena of the better performing parts for each brand, the CPU has far less impact than the GPU in the real world.

    What will perform better? An i7 920 paired with a current, but less capable 4830, or a Phenom II 710 paired with a Radeon 4870 (which I will do the math on newegg shortly and show you, by spending $65 more on the Graphics, but nearly $250 less for the total system when you factor in the more expensive CPU, chipset, and DDR3, and you still get better results through the GPU, factor in the same GPU on each, save $300+ on the total build and just run the same frame rate, use that extra cash to buy something that actually provides you with some real world value.

    Benchmarks run at low res, low detail, may give you a warm fuzzy feeling, but they don't provide any serious gamer with any actually useable real world value.

    Now, I respect what was said earlier about the whole GTX285 vs. the 4870X2, but if you realy want to illustrate this point, pick any high end performing card, whatever brand, and keep it the same across platforms. Heck, for the meager benefits of the DDR3 currently available at its loose timings, I would take an older AM2+ paired with dual channel DDR2, and still run right next to an i7 paired with triple channel DDR3 as long as the GPU's and consistent across platforms, and we are talking about REAL performance benchmarks, you know, the benchmarks comparing the values in gameplay (no phoney timedemos, or cpu specific benchmarks), take real gameplay at a reasonable HD resolution (I might suggest something between 720, and 1080, most people should at least have monitors for 1440X900 by now), Run that puppy details cranked, with some reasonable if not all out image quality guidelines, I would say 2XAA, and 8XAF are good enough at that resolution, run that all other details cranked, run it in real gameplay, take the fraps average of a few runs per game in a few popular titles, let the i7 users swallow the bitter pill when they realize thier money might have been spent on more graphics power than on a new platform.

    And once again, to those that want to argue over performance in photoshop as being an enthusiast application, My Phenom II will generally perform as well as your i7 if I just feed it some more delicious realy inexpensive tight timed DDR2 on a 64 bit OS. Run 8 GB of DDR2, vs. your triple channel 3GB of DDR3, and you will be better off, because the real photoshop bottleneck is not raw cpu speed once you add more memory. Now, if your the kind of guy that can buy the i7, overclock it, slap 12 GB of DDR3 on a board because you swim in cash, good for you, but if your realy balancing performance against price, I can assure you I can find an angle to get at the performance you crave on whatever application for less money on a current AMD platform.

    I think the next step in the debate is to come up with agreed on real world performance benchmarks, and a way to compare them dollar for dollar. I belive in a real world well thought out benchmark, when you compare dollar for dollar, AMD is going to beat Intel. Don't just test theoretical bandwidth, or run specificly designed timedemos, or benchmark suites, I want to see some innovation on how real world performance is measured to provide users with actual meaningful results in how they will use the system. Suggestions?

  23. CyrixInstead
    http://www.cyrixinstead.com

    Wow, a great article Rob, and some interesting discussion in the tread.

    ~Cyrix

  24. CyrixInstead
    http://www.cyrixinstead.com

    "what value does that low res benchmark to test the CPU limited frame rate realy honestly have?"

    Cliff, the point is that in order to test the performance of two processors, you have to make the processor the bottleneck in the system. The only way to do that is to use low graphics settings so that the graphics card is not the bottleneck.

    True it's not real-world and you won't play your games at a low resolution, but what if you buy a new graphics card 12-18 months down the line? All of a sudden, at the settings you were playing at before the new graphics card is no longer the bottleneck, your CPU is.

    The point of these tests is to say that OK, taking the graphics card out of the equation which processor performs faster?

    12-18 months down the line with the poorer-perfoming AMD system you have to go out and buy a new CPU & system. With the Intel not so.

    For the record I've always bought AMD (after my first mistake with Cyrix) but as other people have said until AMD can offer something better, at the moment I would buy Intel if I bought a new system.

    ~Cyrix

  25. It's fairly easy. Test four system configurations:
    Phenom II X4 940 BE + 4870X2
    Phenom II X4 940 BE + GTX 285
    Core i7 920 + 4870X2
    Core i7 920 + GTX 285

    Run through a laundry-list of hand-made synthetic, productivity, gaming and Windows benchmarks. At the end of each section, you can add all the framerates or seconds (or whatever the section's test metric is) and divide it by the total cost of the platform. That will give you your value. If I have 6 game benchmarks with 500 total FPS on a $900 system (example), that's $1.8 per FPS. Maybe we're doing a productivity suite and the total time for all the tests is 1260 seconds, and again with the $900 system, that's $1.4 per second.

    The critical point to all of this is that you must avoid canned benchmarks whenever possible. That means making your own timedemos, running your own FRAPS recordings, making your own productivity scripts when possible, etc.

    And then at the end of it all, you repeat all the tests in an overclocked condition and see which one really gives you more value.

  26. Mirage
    Guest

    Rob,

    There is one review at Tom's hardware not exactly but close to what you suggest. Core i7 is unquestionably better then Phenom II in every benchmark except Crysis since that game is GPU challenged with any CPU. See the link below.

    http://www.tomshardware.com/reviews/...m-ii,2119.html

    However, I can not explain why but I am badly itching to play with a Phenom II. Maybe I am missing my retired Opteron 165. Or, I have completely transferred my constantly changing number of PCs to Intel platform lately and I need some diversity.

    Cliff, I agree with you on AMD's role in bringing computer technology forward. But I do not have any sentimental attachment to either AMD or Intel. Those are both companies working for profit (not me), and what you have described is healthy competition in the free market. My only concern is Intel's monopoly not AMD's demise.

  27. "The critical point to all of this is that you must avoid canned benchmarks whenever possible. That means making your own timedemos, running your own FRAPS recordings, making your own productivity scripts when possible, etc."

    I agree with that statement, and it has to be in things that are repatable meaningful value added tasks. Don't toss out the worst case scenerio on one realy complex job, and don't give it something realy simple to do either. If your a photo guy/gal, or a video editing guy/gal, give me a common, repetable job, something you do often enough that a time savings realy does add value to the experience, not just a benchmark.

    Same thing with gameing, I want benchmarks that reflect the way the majority will experience the games right now.

    I respect what was said about the potential upgrade path, and the CPU becoming the potential bottleneck in a gaming system 18 months from now, but I could still argue against it in terms of value presented, I want to know, how do users extract value here and now.

    I'm a gamer, and I do some light video standard def video editing, nothing too heavy, oh and music encoding too, and I convert a fair number of documents to .pdf, so I feel qualified to comment on those areas, but for our enthusiast friends who use photoshop each day, or perhaps edit HD video a couple times a week, what would be valid repeatable benchmark, something that adds real day to day task to task value? Not the easiest job you ever did, not the toughest one either, something common, repeatable, something you do on a regular basis? Thats going to be the best value added benchmark to measure with.

    Synthetics tell us nothing thats realy tangible in today's computing environment, I want to see real world results, and I want to see it in tasks that users are likely to repeat and get some real value from. Not a benchmark for a benchmark's sake, I think there is too much of that going on already, and its only purpose is to market product, I think the innovator in benchmark review sites is going to be the person that comes up with the benchmarks that refelct real users situations in the here and now, and applies a real cost to performance benchmark against it. Thats what review sites should realy be doing.

    Could be an oportunity for Icrontic, challenge the paradigm of the hardware review establishment. I think when that happens, AMD's market perception will improve.

  28. A really great way to bench Photoshop is to take a very LARGE image (think 98 megapixel), and then make a photoshop action that runs through a series of very intense transformation in sequence. Transformations like blurs and rotations require real FPU horsepower, and are a great test for a CPU's prowess. You chain these things together and then stopwatch the results. Run the action 3-5 times for an average score, and you have yourself a benchmark!

    Media encoding is pretty easy too. The x264 benchmark is a two-pass encoding of a real HD video clip that's repeated 3 times and averaged. It's not synthetic at all. You could go one step further and set up your own frameserving to encode a DVD that you've ripped yourself (this is what I'll be doing when we test Windows 7 RC1 against all current Microsoft OSes/service packs).

    @Mirage:
    Thanks for the link!

  29. neoanderthal
    New to the neighborhood

    I've been an AMD fan for a while, but not so blind that this article didn't strike a chord or three with me.

    I think AMD's done a decent job of offering their customers (which, according to AMD, are the OEMs and ISVs, not the end-user sadly) reasonably long socket life, reasonable performance per dollar, and so on. I think it's pretty cool I can honestly purchase up a Phenom II and put it into my older SB600-based mainboard and it will work (thank you Gigabyte) after a BIOS update. Not just a theoretical upgrade, but an actual supported upgrade for a three-year old motherboard no longer produced.

    I'm disappointed with the fact that many of AMD's technologies don't really have an impact on my day-to-day performance, though. With the lack of I/O contention and all of the memory bandwidth available from using HyperTransport and having an on-die memory controller, my CPU's compute power is still the major bottleneck on my home system. My SATA performance isn't particularly great, because I cannot use AHCI and gain the benefits of NCQ for my file operations.

    With a GTX 260 and 6GB of DDR2, my 3GHz dual-core is the part that keeps me from running my games with maximum settings at my modest resolution (14x9). When I picked up the GTX to replace my Radeon, I originally planned on purchasing a Phenom II 720 as well. It's supported on my mainboard, and I already have the correct BIOS installed. I just couldn't bring myself to do it. With AMD's continued issues with NCQ in their chipsets, I could see myself at some point replacing my AMD-based board with an nVidia-based unit. But, and here's the dark shame I hide from my girlfriend, if I'm going to replace the mainboard, I'm not sure I want to continue on with an AMD solution. Even though I personally despise Intel's "tick-tock" strategy (it seems to be nothing more than a marketing tool to force computing hardware into early retirement), there has been some good gained from it, and I'm beginning to worry about AMD's long-term viability as a competitive CPU manufacturer. The performance of the Nehalem platform is too good to ignore, and I'm beginning to feel a little constrained by my platform's lack of oomph. I don't doubt that the Phenom II would be a considerable upgrade over my K8. But if I need to go with a different motherboard so as to get a chipset that properly supports AHCI, do I really want to limit myself to a second-tier performer? I don't know. It's a hard time to be an AMD fan, that's for sure.

  30. Robert,

    If you got ten thousand computer users in a room, how many of them would find any value in editing a 98 megapixel raw image file? Come on, thats not a real world way to test actual useable performance.

    As far as the Tom's post goes, I can find a few other benchmark sites that proove otherwise on the gaming benchmarks, I am going to come right out and say it, those numbers are cooked.

    This is probably more accurate.

    http://forums.techpowerup.com/showthread.php?t=79635

  31. It is a real-world way, Cliff. Photo manipulation scales linearly with the size of the image. Four 24.5mp images would take the same time to alter in sequence as would a single 98mp image. Ditto eight 12mp images.

    The point is to make the image large enough that the transforms take long enough to do that it can be reliably timed by a human. The test wouldn't give one CPU an advantage over another.

  32. Buddy J
    Dept. of Propaganda

    I think Cliff is trying to say that when working with real-world file sizes and doing real world tasks, the time difference would be so small that it wouldn't make a noticeable difference. If it takes 2 seconds on a Phenom II to do a blur pass on a 5MB image and 1.8 seconds to do the same thing on a Core i7, does it matter?

  33. Not especially, but those same people are also content with their ancient Pentium 4s. I think anyone that is deliberately buying into and discussing the Phenom II vs. Core i7 debate has a very different idea about those fractions of a second.

    I know I do. Especially for $120.

  34. primesuspect
    The Icrontic Guy

    But as it was said in several earlier comments; "our" market is an infinitesimally small niche. The vast, vast bulk of buyers do no ascribe any significance to percentage points of performance.

  35. But that changes the nature of the debate again.

    Who talks about products? Enthusiasts.
    Who generates the buzz for them? Enthusiasts.
    Who writes the articles which generate the context sensitive ads? Enthusiasts.
    Who Diggs articles about these companies? Enthusiasts.
    Who is writing the predictions for the CPU markets? Enthusiasts.

    Who are they not talking about? AMD.

    Let's talk about something more important: The finance sector.

    Who does market valuation? The finance sector.
    Who can crush the value of a stock with a single prediction? The finance sector.
    Who do corporations watch to create their perception of a brand? The finance sector.

    The finance sector doesn't rate AMD well.

    Why are the enthusiasts not talking about AMD, and why are ISVs and OEMs bailing on AMD wholesale? Concerns over their future, their ability to meet demands (Spansion just filed for Chapter 11 because it couldn't meet demands).

    Why bundle a system with a Phenom II when a cheaper Core 2 Duo will offer Corporate Charles and Best Buy Bob similar performance?

    AMD is so swept up in their value proposition of "performance on a budget" that they've failed to produce a processor that can win against a cheaper Intel processor from two years ago.

    When Intel offers faster products, a lower pricepoint, better thermals and a stronger channel... Why would anyone have any need to consult or talk about AMD?

  36. Thank you Buddy J,

    That was exactly what I was saying. In reality, is there a real world, prooveable productivity gain from your i7 platform, vs. perhaps putting that money to use some other way? I think there are very rare uses where a power user might be able to argue the benefits of the platform, super heavy multi tasker, lots of HD encoding to do, perhaps, but taking an extreme task that very few users will ever experience just for the sake of the benchmark is all smoke and mirrors, its marketing, its hype, and its misleading to consumers.

    Also, for the record, the 4870X2 is besting the GTX285 according to this.... Now, once again, not making a point that its a big deal, either card is a monster, and wonderful, but it was stated earlier that the GTX285 was better, and I somehow doubted that claim.

    http://www.extremetech.com/article2/...2343324,00.asp

  37. Snarkasm
    The Photographer.

    We've circled back to the concept that people who essentially do not use their computers (those that could get by on an E5200 and be just fine) would do just fine with an AMD chip. Congratulations, you're correct. People who don't stress their machines at all will be perfectly satisfied with AMD - just like they'd be perfectly satisfied with Intel, with a Mac, or probably even with a reasonably-matched Atom or Via. Why do you think the netbook sector has taken off? A bunch of people don't need any more computing power than that.

    Meanwhile, the rest of us enthusiasts that like to push our machines and work with huge media that you immediately discount as unrealistic like to talk about good performance, and Intel has it. This article was never about whether AMD makes decent processors. We all know they do. This article was explicitly about the fact that even though they do, nobody talks about them because they've become boring and haven't offered a really competitive piece of silicon in several years.

    Is remuxing and ripping a DVD too unrealistic for you as well? What about Folding@Home? Performing a 7-layer HDR render or a 9-frame panoramic stitch? You discount these as things no normal user would ever do, but normal users are actually willing to step DOWN in performance, while the rest of us are looking for the best tools for the job.

  38. ^ Yes, my point.

    I never argued that AMD made bad processors. They make good ones! I'm saying that Intel makes better ones, and that's what captures the attention of the press, the enthusiasts, the writers, the bloggers, the benchmarkers, the tweakers, and especially the finance sector. "Say, AMD makes good things but they're 18 months behind their competitor and we're concerned about their ability to deliver volume orders." UH-OH. Instant buzzkill.

    That's all.

    AMD is not a bad company. They haven't been since they got their shit together with the Athlon XP. But they are no longer revered like they were with the Athlon 64, when everyone and their mother in this industry wanted one. Why did they want one? It was the best.

    People care about #1. Nobody remembers the loser in the superbowl.

  39. Snark,

    Video encoding and folding at home, both more GPU dependant, correct?

    Back to my arguement, the CPU and the platform its on is not the complete end all centerpiece of system performance, GPU has become as, if not more important to the "enthusiast" user.

    Snarkasm, I don't debate that there is a market for every type of user, but how many "enthusiast" users "perform a 7-layer HDR render or a 9-frame panoramic stitch" often enough that the 30 seconds they might save from one platform to the next is going to have more value than the money they can save?

    Its a big complex metric, but I am saying for the computer market at large, and gamers, and most anyone for that matter, it makes more sense to save a little on the chip set, and cpu and plug more into the graphics performance these days. Quad core CPU's are getting it done in any variety and then some for 99.9% of the users, why shoe horn the metrics to only make sense for 0.1% of the market? I think those benchmarks mislead the rest to think they are getting something that they realy are never going to experience, and it does have a trickle down effect in the market perception of a brand, part of why I think less people are talking about AMD. I think hardware review sites have been doing a diservice to the PC market as a whole for years. They look at everything flying at ten thousand feet, but its rarely a realisic way to measure performance.

  40. Robert,

    What AMD offers is a superior value. Why more people don't care about that is beyond my understanding.

    Offering a superior value certainly does not make AMD the looser.

  41. Cliff,

    I agree that AMD has amazing value. No dispute there. All I've been trying to say is that value doesn't get you talked about in the communities that will do the talking. A company's stock price reflects people's faith in your goods, and #2 in a two-person race doesn't cultivate much faith.

    Ditto our markets: Twitter, blogs, forums, etc. We love to talk about the cream of the crop, even if we may not buy into the platform (for whatever reason). We love winners! We love CPUs that trash benchmarks! We love #1! And while we all respect value, the performance per dollar, people just love to buzz about whomever does it best.

    Twitter shows that that's the case, and I'm the observer who's pointing it out.

  42. Buddy J
    Dept. of Propaganda

    I'm really enjoying this thread. Tech Report's recent article on performance/cost analysis of CPUs does a good job showing what processors offer the best value for the dollar in a wide variety of situations.

    Regarding value: We can't simply laud AMD for offering inexpensive chips with decent performance because value is a relative term. Regarding their current offerings, I don't think their value is superior, but only on-par. I'm a gearhead so I like to put things in the perspective of cars.

    If I buy a brand new Yugo for $50, I've got a great deal! As a car guy, I'd actually consider it, but for most normal people, you'd be hard pressed to convince them that it was a good purchase and not just throwing $50 away. Sure, it should get you to the grocery store and back. It might do well getting you to work. As a sound investment, it's probably not the best choice.

    You can buy a used Corolla for $4k. It should do whatever you need in your day-to-day life, gets decent mileage, and is suitable for long trips. What a steal! More people will believe you when you tell them. It makes a bit more sense.

    Or, you can buy a Bentley Turbo-R. In 1997, they cost $145,000. Now, they're down to around $15k. It's a car you'll be able to keep for a loooong time, the value has currently bottomed out and will likely rise in the future, and it offers a level of luxury few cars even today can begin to truly compete with. Is it a better value than the other two cars? It costs more. It does more.

  43. Rob, great conversation spurring article. Happy to see so many AMD supporters, we really appreciate the support.
    Cliff, have to say that you have a solid understanding of where we are and where we may go!

    There are so many commnets on this thread that I simply can't reply or answer or tackle them all, but I will state afew things:

    - AMD CPU's are leading in price/performance in almost every pricepoint we play in
    - ATI GPU's also lead in almost every price point in terms of perf/price, not to mention a true innovative feature set, not just stickers on old product
    - We are committed to gamers and enthusiasts and the communities
    - Twitter search comments are silly and unfair, AMD=AND - do i have to point out that the keyboard has "N" and "M" side by side?????

    If any of you know or read Scott's site, you know he is ruthlessly fair and honest - http://techreport.com/articles.x/16570

    Finally, I am on Twitter - IanMcNaughton - lets go chat there!
    Cheers,
    Ian

  44. primesuspect
    The Icrontic Guy

    Thanks for stopping by and commenting, Ian!

  45. Buddy J
    Dept. of Propaganda

    Thanks for chiming in Ian and welcome to Icrontic.

  46. Hi, Ian!

    No need to point out that M and N are side-by-side. The point I was trying to make with the comparison is that typos and OEM sales vastly outweighed honest AMD Tweets. Even culling the typos still puts AMD at a conversational disadvantage.

    Not meaning to imply that typos count against the firm. That would be dishonest.

    I would like to remind everyone that I am a very strong supporter of AMD. I love the Radeons and my house has two AMD systems and two Intel systems. I strongly believe in the right hardware for the right job and the right price, and AMD is a hard winner when it comes to price/performance. You'd be a super Intel fanboy or a dummy to ignore that!

    This article posed the question: Why isn't the geekiest conversational hub on earth talking about you guys?

    It bothers me, because the younger geek in me crusaded long and hard against Intel for its awful netburst architecture. I won't say the tables have turned, because the Phenom II architecture isn't bad at all (it's great at power/watt, platform bandwidth and transactionals). It DOES bother me that when I look for the top of the performance chart, I don't see much green any more.

    I want the Bulldozer to be a real ass-beater. I hope you guys do incredible things with it. Please? Ps, we want to play with one.

  47. Lincoln
    Snapperhead
    Ps, we want to play with one.

    Shameless. I approve.

    Howdy to Cliff, Mirage, neoanderthal, and Ian

  48. Snarkasm
    The Photographer.
    Snark,
    Video encoding and folding at home, both more GPU dependant, correct?
    ...
    Snarkasm, I don't debate that there is a market for every type of user, but how many "enthusiast" users "perform a 7-layer HDR render or a 9-frame panoramic stitch" often enough that the 30 seconds they might save from one platform to the next is going to have more value than the money they can save?

    First, the point isn't about how many people would save more money than time with an AMD system, the point is that the people who ARE doing this work are excited about the numbers and benchies we see from Nehalem, and are NOT excited about the numbers and benchies we see from AMD's latest batches.

    Video encoding and remuxing can fluctuate, but when I first began, I moved from an AMD 3000+, I believe, to an AMD64 4800+ the first week they came out ($1000 processor, I lolled at myself) - but I moved from AGE long rips to much improved half-hour to hour rips. I freakin' loved it. It was ENTIRELY CPU improved - I only changed mobo with the upgrade. It can really bake your processor.

    As far as Folding, while there is a GPU client that pumps out mad points, many of us also still run the SMP client, which rocks all your cores to 100% usage. Changing processors can net you significantly reduced frame times, which improves calculation speed and point totals.

    I moved from a 4800+ to a 6000+ after that, and then straight to a Q6600 and a Q9550 now. They were great processors in their day, but the Core2 and Nehalem architectures have just outclassed them in terms of raw ability. They make good chips - but so do Via. Neither of them put up the performance numbers the latest Intel architectures do. When geeks talk tech, we get excited over performance numbers for revolutionary tech. I was intrigued when Intel started talking about QPI and on-die memory controllers. I was mesmerized when I heard numbers out of an Asian tech show a year (or two?) ago that said tech-demo Nehalems were besting the BEST C2Q chips by 20% clock for clock or lower clock speeds. The boosts coming out of Nehalem actually MADE ME INTERESTED.

    I haven't heard anything about revolutionary updates to AMD architecture, and sure don't hear anything about 20% boosts over the previous gen clock-for-clock. The chips are good, but they're solid, not exciting. I'll continue to recommend them for cheap builds because, as you say, they'll never use their full potential, but when I'm building for me and I'm building for pure performance, I'm building Intel Nehalem right now.

  49. Snarkasm,

    I respect your point of view, but lets get this out of the way, comparing AMD to Via is just silly, and everyone here knows it. Via is not a major player in the CPU market. For the record, we agree that the Nehalem is an exciting piece of tech, perhaps not practical for a large segment of the market, but it has its place.

    Just to re state my point though, the only reason they offer one at under $300 is because of AMD's sticking out a tough uphill battle, against a competitor that has been called out once or twice for playing less than fair. As computer enthusiasts we have plenty of reason to want AMD to succeed, and we should give them fair credit for what they have achieved at an obvious market disadvantage.

    I still think there are plenty of things to be excited about in AMD's camp.

    - Inexpensive unlocked CPU's, The AMD 780/790 chipsets, Radeon 48xx graphics technology, Amazing pricing, Integrated computing platforms that offer great mainstream performance, potentially high end notebook graphics that wont meltdown (Nvidia's track record less than stellar here), Better Windows 7 drivers, DX 10.1 support, they also have the graphics contracts for the two winners of this console generation, Overdrive and Fusion for gaming software is awesome, I could keep going and going.

    Enthusiasts may be excited about Nehalem, understandable, but don't let that distract us from everything else that is great in the market. AMD is doing allot of great stuff, and people should be talking about it.

3 More Comments in the Forum »


Hey, be nice. Icrontic is full of good people, we promise.

New Features on Icrontic: