AMD reports small loss despite record $1.65 billion revenue

By Jos
Jul 16, 2010
Post New Reply
  1. dividebyzero

    dividebyzero trainee n00b Posts: 4,783   +639

    Let Techspot be your guide-it's representative. And you will notice why I prefaced my opinion with "real world" and not theoretical bandwidth throughput.
    If you've got any information, feel free to share with the rest of the class.
  2. PanicX

    PanicX TechSpot Ambassador Posts: 829

    Hmm.. definitely gotta be trolling.
    But I digress, like I said earlier, when I refer to Intel stagnating, I mean that they haven't had any pressure from AMD. To this you replied, "Not so dear chum! What say you of the K10?", to which I boggle, because the K10 is not an Intel product. In fact, the K10 isn't even a threat and only emboldens Intel's ego. However the article in question points at AMD in a positive light, to which I rejoice in the hope that they can again produce a product that might get Intel to push the envelope.
    What very odd statements indeed.
    You suppose that I refer to IBM's workstation production (Lenovo actually) as a premise for lack of progress from Intel? Interesting. There was a time that IBM manufactured processors that were in direct competition with Intel, however they haven't been a threat to Intel's x86 line since Apple dropped them.
    Well thats a tough one. It sure is, a regular chin grinder, this question right here.
    Hmm... well I suppose that it might be a small bit of improvement if maybe they, oh I dunno, lowered the prices to compete against the competitions offerings? Cause the old K10 is really knocking socks off.

    You see Mr. DividebyZero sir, I often post in regards to my personal experiences. Odd as that may be, I know that the 5 i7 systems I built, over the past year, all cost approximately the same and had approximately the same hardware. And my experiences have often shown that when competition is in the market, you can build a system 1 day for say $1000, then 2 months later build the same system for around $800. But here all these systems have been near identical and the price hasn't reduced. This is my qualm. The point I've been trying to make here and it seems over your head. Now if you really wanted me to, and I mean, if you really really wanted me to, I could send you a copy of a receipt for an i7 system I built in August 2009, and then another receipt for a system built in May 2010. You can say "its unsubstantiated" but my bank account sure could substantiate it.
    Or we can gander and your fine links, like a price chart without years listed and an ambiguous "now" column, or a link about "Look price drop!" from less then a month ago, (thats how long since release?) or better yet, tigerdirect.com.

    Anyhow, feel free to splice this response in quotes down to the letter to focus hard at missing the point yet again.

    Oh, and FYI the "xenophobic prejudice" was actually an insult directed at you. How embarrassing that I'd need to spell that out.
  3. dividebyzero

    dividebyzero trainee n00b Posts: 4,783   +639

    Soooooo, just a recap. Intel should lower prices to compete against AMD because AMD is providing no real competition.......gotcha..... The Oxford Union probably won't be sending out the invite anytime soon.

    So, other than lowering prices, you can't think of a thing that Intel could do that could improve the platform...and by platform I mean CPU, since the boards are the province of motherboard vendors.
    Personally, having owned an X58 since just after launch, and having built a few since, I'd say there isn't too much glaringly wrong with the platform. For those that find the extra expense a burdon, LGA1156/P55 makes a fine substitute.
    You mean this chart....the one that was circulated a month ago to system builders like myself ? And was widely circulated on the net as well as the primary subject of Intel's June press release? (hint: if you follow this link you will see July's offering)
    I see....so, a price drop that occurs before your post is only relevant if it is instituted no later than May 2010, and any retail/etail outlet that isn't newegg does not count....Interesting set of parameters to say the least
    It was apparent, obviously irony and reasoned argument aren't your strong suits.....how embarrassing I need to spell it out.
  4. Matthew

    Matthew TechSpot Staff Posts: 6,043   +84 Staff Member

    Come on guys, argue the facts and leave your cynicism out of it. It would be a shame if we had to close this thread because a couple of members couldn't exchange their thoughts without mudslinging.
  5. dividebyzero

    dividebyzero trainee n00b Posts: 4,783   +639

    I would judge this particular line of discussion at an end in any case.
    Panic attacks the argument from the emotive standpoint- that the platform should somehow guarantee lower prices as it ages, while my viewpoint is purely business based. While I can appreciate the contrary viewpoint as an abstract, it makes no business sense for Intel to lower it’s ASP’s and profit line while the product remains popular (by Panic’s own admission he has built five X58 systems) and is operating largely within a competitive vacuum. While I’m at a loss to see Panic stricken by my argument which basically summarizes what appears to be a highly effective business model, I would have thought Intel’s relative growth and the forecasts for both it, and it’s AMD over the next 2-3 quarters would speak for themselves.

    Hopefully the thread- if added to- reverts back to an AMD-centric discussion.
    My apologies for my part in the hijack…I’ll now hand control back to your regular flight crew.
  6. PanicX

    PanicX TechSpot Ambassador Posts: 829

    Putting aside all the petty bickering, this is still the point that to me is glaringly obvious and for all accounts seems you're intentionally getting wrong.

    I think we can agree that Intel will charge as much for their processors as the market will bare. In my opinion, they've been overcharging because they can. They can overcharge because there is no credible threat to their sales by another vendor.
    In an ideal world, AMD and Intel would be fiercely competing and trying to grab market share which results in better products at lower prices. As a consumer, wouldn't this be desirable?
    So the entire jest of my first post which you took so much offence to, was this. "Get in the game AMD to get competition going again."
    I've tried to illustrate this as clearly as possible in my posts, perhaps its a failure of mine to convey this opinion, or perhaps you feel that my stance on market competition is wrong. In either case, take it as you will, I just can't get any more plain than this.

    This is a non-sequitur to my original point.


    Yes, that chart doesn't make sense to me at all. Maybe this is short sighteness on my part, but it appears that that chart shows a price drop in the i7 950 happening on August '09 or will happen August 2010? When your other link clearly claims a price drop coming in June 2010 which is an odd contradiction, particularly because price checking on the 950 right now, I can't find it cheaper than $500 anywhere. And even if I assume the other prices listed there are accurate, there's no price drops listed on processors AMD can't surpass with their current offerings. Which supports my point about competition being beneficial to the consumer.
    So you could say that I'm selectively choosing to dismiss that June article because its claims would imply I'm wrong about Intel's pricing. However the claims that the article makes still don't show in the markets pricing and even if true, are set to happen over a year from the initial introduction of the chip.

    Also, its not that any retailer besides Newegg doesn't count. It's that tigerdirect is known for illicit rebates, fraud and misleading pricing.
  7. dividebyzero

    dividebyzero trainee n00b Posts: 4,783   +639

    The first version of the Intel price-cut chart for June was sent to me via email-hence the fact I put it up at Imageshack for viewing. The second link to the same chart I deliberately linked to a post (and by association VR-Zone's story link) that is dated Jun 23 2010, thus the wording under the chart reading "The following changes in Intel Price- list will occur on 29 August".
    BTW: Of course the August 2010 price drop for the 950 isn't reflected in the July pricing realignments....since it isn't dropping in price until...August. Nowhere in the Techspot link/Intel July pricing guide is the 950 mentioned for a pricing cut.

    In an ideal world, yes, most definitely desirable. But the point I'm making is that we don't live in that world. We live in a world where demand dictates pricing, the same world that allows a near one year old graphics accelerator series (HD5xxx) to retail at or above it's launch prices, where the vast majority of CPU's have locked mulipliers in order to charge exhorbitant amounts for the ones left unlocked

    No. You'll find that the phrase I took issue with was from post #5
    While in general terms the pricing is unchanged your statement is not wholly correct. The pricing is what the pricing is, b*tching about the cost never brought a price down if demand for it stays high...ergo it's a non-runner. You buy or you don't buy.
    The main issue for me was the first part of the assertion (stagnation of i7 since introduction), which from a technical standpoint is false, unless Intel is judged by a different yardstick than every other tech entity.
    Since LGA1366 was intoduced in November 2008, Intel have managed one revised stepping of 9xx CPU's, have successfully transitioned to a new process node at 32nm, brought to retail true six core (desktop) and eight core (server) CPU's (and introduced LGA 1567 for the latter Nehalem-EX CPU's), Not what I would call stagnation.
    Penultimately...
    Since the your original post contained two distinct statements seperated by a conjuction, then logically my response is not a non sequitur (there is also no hyphen between the words).
    Finally...
  8. PanicX

    PanicX TechSpot Ambassador Posts: 829

    Ok, so then it sounds like we agree that i7 pricing hasn't changed much since launch. You say this is because of demand, which I don't fault. I'm just saying demand wouldn't be so high if AMD had stayed closer on overall performance. With the Radeon pricing you mention, I'd point at exactly the same factors.

    Post #5 is my first post in this thread.

    It's not that I measure Intel by a different yardstick as you put it, I'm simply comparing developments in regards to the history of developments in the CPU industry. For example,
    In 2006, Intel releases conroe, conroe xe, kentsfield xe
    In 2007, Intel releases Allendale, Kentsfield, yorkfield xe
    In 2008, Intel releases Wolfdale, Yorkfield, Bloomfield (i7)

    In 2009, Intel releases Lynnfield.
    In 2010, Intel releases Gulftown (so far)

    When looking at the history, maybe stagnation isn't the right word, but I think it shows my point. The i7 shows up, trumps all and suddenly progress has lulled.

    Now it could be that there's a variety of factors for the reduced amount of releases that I'm not considering. I'd welcome any insight on this because as far as i see it, when Intel doesn't get heat from AMD, they stop trying.

    Also, I point at the argument as non sequitur because my defining X58 improvements doesn't demonstrate any counterpoint to my post #5. No list or lack there of would be relevant to the need for competition or note of lack of releases I pointed out.
  9. dividebyzero

    dividebyzero trainee n00b Posts: 4,783   +639

    Regarding the Intel lineage...
    Conroe and Conroe XE should be considered one family. XE's are just the unlocked multi CPU (X6800 was the only one released). Kentsfield is simply 2 Conroe cores on the same package -so 2 x 143mm² instead of the single two-core Conroe of 143mm². Kentsfield-XE is in turn the unlocked multi (eXtreme Edition QX6700/6800/6850). All have the same basic cache structure and instruction sets, as does Allendale, which is a Conroe core with a halved L2 cache (which makes the die a little smaller) and a crippled FSB (800MHz, probably to suit the budget nature of the part)...So architecturally Conroe, Kentsfield and Allendale (along with the Xeon Tigerton, Clovertown and Woodcrest) are the same chip- just that the "Quad cores" feature two of them. All are 65nm process.

    Once you die shrink the architecture to 45nm (and add SSE4.1 instruction) and a little refinement the CPU's turn into Wolfdale and Wolfdale 3M (Core 2 Duo E8xxx and E7xxx). Two Wolfdale CPU's in the same package becomes either Yorkfield 6M (Q9xxx/Q8xxx w/ 6Mb cache), Yorkfield (Q9450/9550/9650 w/12Mb cache) or Yorkfield-XE (QX9650/9770/9775). Add Xeon Harpertown and Dunnington (6 single core package) to the rebrands to complete the family.

    New architecture for Bloomfield (9x0) and Bloomfield-XE (965/975) (plus Xeon Gainestown 5500 series/Bloomfield W35xx series). Minor overhaul for Lynnfield (plus Xeon 3400).
    Die shrink to 32nm....Clarkdale (Core i3/i5 w/IGP) from Lynnfield, Gulftown (+ Xeon) from Bloomfield. In addition you have Intel's server CPU recieving a major architectural change with Beckton (Nehalem-EX) 8 core (LGA 1567) E/L/X65xx/75xx Xeon as well as the Bloomfield-Lynnfield hybrid Xeon Jasper Forest series, both of which look to supply facets for Sandy Bridge (Q4 2010) and Sandy Bridge B2 (late 2011) desktops.

    In essence, since Netburst's demise you could say the family tree consists of new architecture (C2D/C2Q at 65nm) , architecture tweak (C2D/C2Q at 45nm), new architecture (Nehalem at 45 and 32nm)

    Of course while this is all shaking out Intel still have CPU R&D investment with Atom (Silverthorne, Pineview, Diamondville) as well as the usual (ever expanding) lines of Clarksfield and Arrandale based Nehalems. Mobile and ULV in general obviously play a far larger part in Intel's plans than they did in previous generations, as might the research projects Larrabee and the 48 core SCC .
  10. PanicX

    PanicX TechSpot Ambassador Posts: 829

    Thats a great detailed explanation of Intel's desktop processor developments. I see what you mean by the XE not being a big deal on releases. So basically only architecture changes and die shrinks are what you consider to be noteworthy advancements that they should be measured against (in desktop processors).

    And in those terms, here's what I'm seeing.

    2006 - 90nm to 65nm - introduction of core arch
    2007 - 65nm to 45 nm - core 2 arch
    2008 - introduction of i7 arch
    2009 -
    2010 - 45 to 32 nm

    Now I get that the mobile market is bigger than ever and it'd make sense they're increasing efforts to develop that market. But would you say that the lack of development in 2009 would be shown in an increase in development in mobile products at that time? Is the industry leaving the desktop platform and thats the result of less development instead of the lack of competition?

    Because I'm still having a hard time seeing a fault in my original post.
  11. dividebyzero

    dividebyzero trainee n00b Posts: 4,783   +639

    No.
    No major product release for 2009 is probably co-incidental to a degree.
    What I would say, is that Intel got most¹ of it's ducks in a row in 2006-07 to allow themselves some time to regroup and have a few financial quarters without significant expenditure. The ten months between Bloomfield and Lynnfield represent Intel's return on investment for C2D/Q at 45nm.
    Bear in mind that for most people "die shrink from 65nm to 45nm" sounds like a fairly easy process. For Intel it meant building 45nm fabs to make wafers on that node. Actual cost was in the region of $US 9 billion dollars. You can see why Intel would like some time at that node to recoup capital expenditure. Bear in mind that the move to 32nm is likely more expensive per wafer run. Of course not every run at every fab is an Intel product, but the majority are.

    Personally I would view the timeline as:
    C2D @ 65nm (Mainstream) Aug 2006
    C2Q @ 65nm (Enthusiast) Nov 2006 (interval 3 months)
    C2Q @ 45nm (Enthusiast) Nov 2007 (interval 12 months...15 months between 65nm and 45nm)
    C2D @ 45nm (Mainstream) Jan 2008 (interval 2 months)
    Bloomfield 45nm (Enthusiast) Nov 2008 (interval 10 months...27 months between architectures)
    Lynnfield 45nm (Mainstream) Sept 2009 (interval 10 months)
    Clarkdale 32nm (Mainstream+IGP) Jan 2010 (interval 4 months...14 months between 45nm and 32nm)
    Gulftown 32nm (Enthusiast) Mar 2010 (interval 2 months)
    Sandy Bridge 32nm (Mainstream) c Dec 2010² (interval <9 months...<24 months between architectures)


    ¹ I say "most" because Intel deferred moving to immersion lithography (a major expense) at 45nm which is probably the limit for the dry lithography process. 32nm and smaller require immersion lithography for accuracy.
    ² Probable worst case scenario

    I think the pricing argument has been done to death. Regarding the stagnation...note the time frames between architectures and between process nodes.
     


Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...
Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.