AMD's Project Win: A Misunderstanding

by Anand Lal Shimpi on 11/9/2011 5:49 PM EST


Back to Article

  • RussianSensation - Wednesday, November 09, 2011 - link

    Hi Anand,

    Thanks a lot for this update. Good to know that AMD doesn't intend to unveil a major change in strategic direction, at least not on a public basis until next year as you have noted. Looking forward to your 28nm HD7000 series GPU review. Can we possibly expect an update to the game test bench suite, perhaps SKYRIM, BF3?
  • Dustin Sklavos - Wednesday, November 09, 2011 - link

    Ryan, Jarred, and I are working on revamping our gaming test benches and trying to line them up between graphics, desktops, and notebooks. BF3 and Skyrim are both contenders. Reply
  • tviceman - Wednesday, November 09, 2011 - link

    Try to pick out games that scale well with CPU AND GPU's. Civ V's lategameview benchmark is one example. Far Cry 2 also scaled well, but it's pretty old now so probably not worth having in a current benchmark suite. Since Metro2033 is the current most demanding game out there, it's probably good to keep it in the mix. Reply
  • Klimax - Thursday, November 10, 2011 - link

    Would it be possible to include SupremeCommander FA? (I can supply save for SP which can overload CPU - 20/30 FPS on Core i7 920.)

    It would be good to know if those new CPUs can finally run the game with high number of units...
  • SlyNine - Friday, November 11, 2011 - link

    Here's the thing about SCFA (and I love the game BTW). The default AI caused major slowdown and only used a single core for those tasks whiched caused the slow down. So until we have 10X single core performance default AI won't run smooth on SCFA.

    I was running a Core I7 860 and I can tell you that with the modified AI (which is much better anyways) there is not very much slowdown.

    When I played Multiplayer with huge units on the lan I never had slowdown problems with 4 civs 2000 units a piece.

    Now I run a 2600k at 4.4ghz but have not tried SCFA yet ( lost my disc trying to obtain another one)

    Try these mods, FA Patch 3603 beta, Sorian mod, and DuncanE's FA AI Patch. Makes the game much much better.
  • Tormeh - Thursday, November 10, 2011 - link

    Little off-topic since it's the CPU, but I've always wanted a CIV V end-of-turn benchmark: How long do you have to wait after clicking the "end-turn" button until the enemy AIs have finished their turn and it's your turn again? It's one of very few instances in gaming where you are actually just staring at your monitor waiting for the CPU to finish.

    Pretty please? :)
  • Taft12 - Friday, November 11, 2011 - link

    I recall Ryan making a comment once upon a time that he really lamented just how lousy some of the games used in his benchmarks were.

    I agree, and many users come here looking not for a large selection of benchmarks in games they will never play (F1, Dirt, Crysis) but ones they plan to spend hundreds of hours in like Skyrim, BF3, Starcraft 2 and Diablo 3.

    Metro is a must since its still the most demanding game out there, but please, more focus on the best PC games rather than the ones that are the prettiest.
  • alphacheez - Wednesday, November 09, 2011 - link

    I think it should be called Project Winning and involve Tiger Blood and Warlocks. It would probably make more sense than Project Win. Reply
  • Roland00Address - Wednesday, November 09, 2011 - link

    sigh...I like AMD...but doing one followed by the other is just FAIL :( Reply
  • Operandi - Wednesday, November 09, 2011 - link

    Layoffs are of course not a good sign for but the marketing department indeed failed with the launch of Bulldozer. How they ever gave it the FX designation is beyond me, all they accomplished is ruin whatever good name the FX brand had. The engineering layoffs are more worry some but we really don't know the reasons behind them. Hopefully it results in better products and more timely execution and not the end of completion for Intel on the high-end.

    I also like Reads comments. Faster moving R&D, and timely product execution is exactly what AMD needs. whether he's the right man for the job or not remains to be seen but something has to be done to get things turned around.
  • Mishera - Friday, November 11, 2011 - link

    Yea, but I can't help but to think how much more streamlined can you get on the engineering side? The GPU side's doing very well so I don't know if you want to mess with that, and I thought layoffs were one of the main reasons Bulldozer disappointed.

    This is especially scary with many upper management engineers having left. I've got to think that this is going to affect AMD's long term vision, which is probably bad for the industry.
  • Beenthere - Wednesday, November 09, 2011 - link

    We'll see if Mr. Read and the board are on the same page, soon I suspect. Reply
  • frozentundra123456 - Wednesday, November 09, 2011 - link

    Anand, maybe you should run AMD. What you said in the last paragraph is much more concise and makes more sense than the marketing babble that RR was spitting out.

    Does this guy have any engineering expertise, or is he just a marketing person? Maybe he will do an excellent job, but so far I have a bad feeling about this guy. Glad I dont work for AMD.
  • Ethaniel - Wednesday, November 09, 2011 - link

    Seconded for Anand as CEO. I'll take the Catalyst department. I don't know a thing about programming but I'll whip them until the drivers are 15 MB in size. Or less. Reply
  • Alexvrb - Wednesday, November 09, 2011 - link

    Hi, it's 2011. If you can't download Catalyst packages in the blink of an eye, you need to get a friend to do it for you. Or else go leech wifi from a restaurant or coffee shop. Something. Don't cut out features for everyone else. Reply
  • Ethaniel - Thursday, November 10, 2011 - link

    I'm not talking about bandwidth access or speed. The drivers are just bloated. The hotfixes are even bigger. They could be a lot more efficient. Reply
  • Arnulf - Thursday, November 10, 2011 - link

    What features, silly ? Catalyst drivers are a huge blob of crap that should really be 10% times the size it is now, especially considering the fact that it requires yet another large blob (.Net) to even run.

    There is absolutely no reason why the driver and the tweaker application that consists of nothing more than few sliders and checkboxes should be larger combined than the minimum size for a GUI application, which on Windows I estimate to be well under 10 MB (driver taking the smaller portion of this).

    Anybody who has done low-level programming on things such as drivers knows that there is no way in heel to get them bloated up to 100+ MB without doing dumb things on purpose.
  • chiddy - Thursday, November 10, 2011 - link

    +1 Reply
  • Taft12 - Friday, November 11, 2011 - link

    Give this man a 6! (well crap, this isn't dailytech)

    The .NET requirement especially is plain idiotic. They sure don't need it on the Linux side and CCC looks about the same. Many of us like to truly minimize the OS footprint on a gaming machine Windows install.
  • Sunsmasher - Wednesday, November 09, 2011 - link

    I'm concerned about the direction of AMD these days.
    I read Rory Read's comments to the troops and it really sounded like a lot of corporate-speak.
    However, on reflection I believe that this is what he was really trying to say:

    "Last quarter, or was it fifty cents, I described a set of initiatives to streamline business and decision-making processes across our operations, (in order to speed them up to a higher speed), R&D (so that we can discover new ways to make the things we manufacture), and go-to-market functions (so that our inventories of processor component fruits and vegetables remain at competitive levels). We are in full executional deployment across each of the key work streams, getting things done at the time that they are completed, efficiently and finally. These efforts are aimed at accelerating our transformation to a world-class design company that efficiently designs World Classes better than any competitor...growing revenue to a point where it is much greater than now, lowering costs to a point far less expensive than at the present time, and reducing time-to-market (driving faster than in the past when going to shop). We expect to see material benefits from this project that will greatly benefit us in 2012, or sooner if our future projections are accurate and we are able to arrive at that year before that time."
  • nofumble62 - Thursday, November 10, 2011 - link

    At Lenovo, the engineering work is basically to take a reference design, added a little spin to make the product. At AMD, it won't work anymore. Reply
  • Taft12 - Friday, November 11, 2011 - link

    Uh, Lenovo and AMD have completely different core businesses Reply
  • mino - Sunday, November 13, 2011 - link

    That is the point ....

    (RR was at Lenovo just before ...)
  • pcfxer - Thursday, November 10, 2011 - link

    BS, if you want to improve a business you will rarely remove your engineering capacity. Your engineers are trained and have learned key knowledge specific to AMD in the industry; with such specialization you don't remove ANY of this capacity.

    It is also important that one does not reduce engineering capacity but also not to reduce engineering support capacity.

    I agree with frozentundra, Anand should run AMD. None of this political/PR crap. No wonder our old friend left AMD a couple of months ago.
  • Belard - Thursday, November 10, 2011 - link


    Yeah, I'm for positive spin for the AMD company... I'm not waiting for AMD to get their act together, again.
  • Casper42 - Friday, November 11, 2011 - link

    Why does the picture of the roadmap slide say Next Gen Bulldozer in 2012.
    Doesnt everyone pretty much already know that will be Piledriver?
  • danwat12345 - Sunday, November 13, 2011 - link

    Agrred should be Piledriver in the chart. Reply

Log in

Don't have an account? Sign up now