AMD now worth less than it paid for ATI
- Thread starter Davros
- Start date
- Status
I'd say they're better in the GPU realm than the CPU realm. At least now they seem to be fairing better with the price/performance of GPUs than CPUs. Don't you think they'd have released a bugfree top of the performance crown Phenom if they could have by now? Should they go back to doing nothing but GPUs?
As of right now the ATi part of the company is the only one showing any promise at all. HD 3850 and HD 3870 are really good products in their price range, and if Nvidia doesn't have anything in the 3850 price range then it could lead to a lot of sales, we'll see. AMD's problem is that Phenom is a disaster, no getting around that.
AMDs other problem is their management, you think ATi is the better half? They disagree, they infact blame it as the main cause:
Xbit did a pretty decent summation/breakdown
http://www.xbitlabs.com/news/cpu/di...Graphics_Multimedia_Business_Performance.html
Personally i blame both. Core 2 duo crushed AMD, and they're of course still working on a reply to it 2 years after the fact, and both sections of AMD, graphics and central processors, keep blowing product release schedules which they cant afford to do in either case.
Not going to help a whole lot if you cant release your products in a timely fashion which makes OEMs reduce the size of their orders (especially with the other guy being dependable as ever) and im sure demand is in the toilet as well when the price/performance ratio is so obviously skewed in favor of Intel when it comes to computers based around their processors. The problem is everything. From what has been coming out of their R&D right down to the release and marketing of the actual product.
Last edited by a moderator:
AMD can't blame ATI, they can blame the acquisition of ATI (which is AMD's fault) but I'm going to say they were fairly certain what they were getting into with ATI. If you look at each sides most recent products though by far ATI's show the most promise. At least ATI has a real chance to sell a number of cards in the mid range market, which is where the money is at.
AMD could always go for cutting costs as much as possible. Maybe there's a market for a deluge of ultra cheap cpus with the performance of say a 2Ghz athlon?
I am really surprised how fast Intel has surpassed AMD. There was a several year period where Intel stagnated while AMD increased performance at the normal rate, and the Athlon 64 seems like a design that's about as good as x86 can get, people say Mhz and the single core are dead, and then bam, intel comes out with the core 2 duo series and they're not really letting up. It amazes me just how easily intel has surpassed AMD, but at the same time I suppose it's representative of just what a colossal failure the P4 must have been in Intel's eyes.
Not to mention that AMD's fab tech has always been behind intel's, and you'd have think that would slow as things get tougher (for everyone) to move to smaller nodes, but intel still seems to be moving along at about the same pace while everyone else is slowing down.
It's funny though, had AMD not bought ati, the market capitalization of ATI + AMD would probably be higher than it is today. People seem to ignore what ATI brings to the table, beyond the added costs to AMD.
Thank goodness for AMD though that nvidia isn't the collosal giant that intel is. Besides better dev relations and a current lead in GPU and chipset tech, nvidia competes under the same conditions that ati does, and if anything ati might have the fab advantage now if they can leverage amd.
How is AMD not already part of that market? The Sempron-LE family, which is based on the 65nm single-core Sparta, is pretty much what you describe there; it sells from $37 to $53.
However, I agree that even longer-term there's probably a market for this kind of performance due to the commoditization of the CPU industry. However, it's not like Intel doesn't have anything coming up there (Diamondville/Silverthorne microarch) and it's pretty much VIA Land right now (C7 with half that level of performance, but CN is expected to be roughly on par).
In fact, as I already said before, I'm incredibly bullish on the prospects of CN so I'm sure competition would be pretty intense there - and honestly, even with all my optimism, I don't think that market is sufficiently big to usable as a 'strategy' for a company the size of AMD.
That's not completely true; TSMC is the clear exception here. Their 45nm node (with 40nm-like design rules, there won't be a process called 40nm for the general purpose/high performance markets) is likely going to see PC products released on it before 2009.
Assuming their density is as good as they claim it to be, they'll arguably be ahead of Intel litography-wise for at least 12 months! Of course, that's not considering other important factors that define the quality of a process, but those are much less public. And it's worth pointing out TSMC's 32nm SRAM density isn't as extremely impressive as their 45nm one.
Yeah, agreed - at least in terms of not having a massive advantage over AMD, I'm not convinced CPU fabs matter whatsoever in this kind of discussion.
However, I'd still be curious to see what the GPU R&D budgets for both NVIDIA and AMD are. Neither give that kind of granularity; I saw an analyst claiming $500M vs $200M/year recently, but I can't remember where and I doubt it's very reliable...
- Status
