Slashdot Mirror
AMD's Plan To Recover From Its Perfect Storm
An anonymous reader writes "TG Daily has an interesting write-up on AMD's big Q1 loss and how the company plans to get back into the black. AMD admitted that Q1 was a meltdown and not just a miss. Looks like cost cutting, including layoffs, may be on the way. But the company says it won't change its overal direction. The CEO Hector Ruiz is quoted as saying, 'We are not going to change our strategy because of one lousy quarter.'"
Keep in mind that AMD recently greatly increased the clock frequency of their CPU's (as noted on slashdot), thereby also increasing the performace of single-thread applications and games.
This may help them get back on track.
My security clearance is so high I have to kill myself if I remember I have it...
I was at AMD during the ATI merger and I totally called layoffs in the upcoming quarter. This is what happens when MARKETING runs the company instead of the engineers. AMD makes processors, not "solutions." The moment they start to focus on the meat and potatos again and not the "whatever Intel is doing but with a green palette" the better.
Why did AMD start to eat Intels lunch? Compare the products at the time. Athlon vs. P3. Roughly equiv but the Athlon scaled, and scaled. Intel got scared and made the P4 which tanked because it was slow, drew way too much power, etc. Now that Intel has grown up a bit and caught up, AMD's answer? a 3GHz 120W core. Quad-cores in the future, etc. Where is the power savings? Where is the cheaper process? etc.
The core2 already pretty much beats the AMD64 in every measurable way. It's roughly the same in IPC, has a faster FPU, more cache, takes less power, runs cooler, etc. The only saving grace right now is HT which can help in certain applications.
Where are the lower power AMD64's for desktops/mobiles? Where are the 2MB/4MB cache parts? Where's the faster FPU? (the latter bit is coming up this year iirc)...
This isn't to say the AMD folk are bright people. The Athlon was a fairly performance driven design for the day, and the improvements in process have kept it in the running (anyone remember how hot the K7's ran?). But sadly I see AMD lagging behind Intel in both design and process for the fair length of future. Which is a shame because I've been a fanboi for a long time and would love to see AMD processors in my workstation in the future (right now it's a E6600 core2).
Tom
Someday, I'll have a real sig.
Having a monopoly isn't illegal. Abusing monopoly status is, but it's a long process to actually do anything about it. From 1969--1983, IBM was under suit for abusing their monoply status.
Do you even lift?
These aren't the 'roids you're looking for.
>> "We are not going to change our strategy because of one lousy quarter."
Without the benefit of insider knowledge, that statement wasn't hugely informative. There are so many changes afoot that it's almost impossible to forecast anything at all concerning the CPU companies at the present time.
The acquisition of ATI really complicated things, not only for share speculators but from a tech standpoint too. And while it doesn't necessarily mean that Intel will hitch up with nVidia (it seems not, given that the GMA965/X3000 competes with nVidia's lower-end offering), it does mean that both of those companies will have to respond very strongly to whatever develops from the joining of AMD and ATI. This whole area will become even more hectic than usual I think, once we start to see the fruits of the acquisition.
One of the things that will undoubtedly be on many Linux user's minds is whether the legendary disinterest of ATI in properly supporting Linux will change for the better. Once Microsoft shed nVidia in favor of ATI on going from Xbox 1 to Xbox 360, the likelihood of any such improvement plummetted drastically for obvious reasons, but the influence of AMD could of course be the exact opposite, since AMD can't afford to alienate the Linux market, one imagines.
But while we can hope that AMD will have a positive effect on ATI's attitude towards the FOSS community, what if the opposite happens, and by being tightly coupled to GPU hardware, AMD's CPUs start to lose the openness that has been traditional among CPU manufacturers until now? It's certainly a possibility, and a matter of enormous concern.
Which brings me back to the quote from TFA. It would really help AMD I think if the company removed some of the uncertainty or ambiguity in its position w.r.to FOSS as a result of the ATI thing. "No change" is a rather meaningless statement when their CPU and GPU divisions have diametrically opposite tendencies.
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
I'm not sure if I speak for anyone else, but the reason I stopped buying amd is because of the merger with ati.
/that/ bad, right?" wrong. Therefore, if AMD is going to force me to buy an ATI chipset, while still neglecting ATI support for linux, I'm going to go elsewhere.
ATI has consistently made horrendous linux drivers. They don't keep up to date, and they completely abandon "legacy" cards. Nvidia cards, however, have excellent drivers for linux, and always have. For that reason, I buy Nvidia cards over ATI ones.
With this new merger, however, it's become nigh-impossible to find a decent, small laptop which has an amd processor and an nvidia graphics chipset. I ran into this problem when buying my current laptop and thought "well, they're owned by amd now, they can't be
Intel, on the other hand, has an excellent driver for their graphics chipset, and it's even open-source. They might be the monopoly, but as far as linux is concerned, they actually seem to listen. My next laptop will be all Intel for that reason.
AMD, I've used your processors religiously for years, but if you're going to forsake your linux guys by forcing us to use ati graphics hardware with crummy drivers, don't wonder why your market share is going down. I know I'm not the only one.
Well, only since Core. Before that you had Socket 3 (486), Socket 4 (Pentium), Socket 5 (Pentium), Socket 7 (Pentium), Slot 1 (Pentium II/III), Socket 370 (Pentium III), Socket 423, 478, and 479 (Pentium 4 and M and Core), and now LGA775 (Pentium 4 and D and Core 2).
In a comparable timespan, AMD used Socket 3, 4, 5, 7 (along with Intel), Socket A, Socket 754, Socket 939, AM2, and AM3. Pretty comparable overall. So the real question is, does the recent lack of change on Intel's part show a specific intent to stick with a socket, or is it just "we're improving our internals and we don't need to play with the interface right now" ?
Then moving onto graphics, I've never really had any allegiance to nVidia vs ATI, but it's hard to ignore nVidia being the only kids on the block with DX10 cards out there, including budget ones now too... with NOTHING being shown from ATI/AMD.
It really just looks like (from this purely consumer point of view over here) that AMD is being left in the dust in terms of getting out leading edge products.
I really hope they can turn it around and bring out something to make me want an AMD core and gpu, but I see nothing that makes me want to change my mind as to my intended purchases come tax time in July!
Actually, the forced switch to AM2/DDR2 has hurt AMD badly. With their on-die memory controller Athlon 64s were very efficient (97%) at using the memory bandwidth of the original DDR400 memory, but relatively poor at using DDR2 because of it's worse CAS latency.
Pushing the change to DDR2 was a clever move on Intel's part. Not only did they make AMD change their socket design, which upset their customers, the new memory also hobbles on-die memory controller, one of the key performance advantages of the Athlon design.
"I've got more toys than Teruhisa Kitahara."
It also wouldn't be expensed. Profit is calculated by taking revenues and subtracting costs. Since the new factory goes on the books as an asset, and is counter balanced by reductions in other assets (cash on hand) and increased liabilities (loans), revenue accounts never factor into it.
On the other hand, it probably will lead to lower variable costs when up and running at full capacity. I'm not too familiar with AMD's financials, but it seems that the key question is whether it is that their fixed costs are dragging them down (in which case they need to either improve sales or cut some dead weight), or their variable costs (in which case they need better manufacturing processes or to raise unit prices).
====
Crudely Drawn Games
The stock is right at where it ended the year in 1999. A great many other tech companies from that era are no more, or are trading at pennies on the dollar. Since 2000 AMD has handed Intel their hat time and again. Ruiz is doing great work.
That said, his engineers had better pull a rabbit out of their hat. Today he's getting stomped by a very angry Chipzilla, and Chipzilla looks like the type that holds a grudge for a looong time.
Help stamp out iliturcy.
And it's been all downhill since.
Question: if you are an underdog in a hypercompetitive industry, when a little success comes your way and you are finally climbing out of debt, do you:
(a) Stop what you are doing and deeply indebt yourself in order to enter another cutthroat industry largely outside of your expertise?
(b) Freaking invest in your core competencies while you have the chance?
AMD did a lot of the former and a little of the latter. How long will it be until they spin off ATI at a multi-billion dollar loss?
To be fair, Intel got their act together in short order. However, I have to wonder if AMD could have maintained their lead if they weren't gathering wool. For at least 25 years, the market has continually payed through the nose for leading edge general purpose computing power, and AMD was finally beginning to grab a share of that high-margin turf - from a competitor an order of magnitude larger!
And they gave it all up for socket compatible GPUs, which, unlike the core2, are nowhere to be seen.
*sigh*
Time to add 0.50 SGI advantage-squandering units to AMD's tally... I hope that their accelerator gambit pays off. I hope even though I know better.
Seriously, how did you guys plan to put 512mb of multilinked DDR3 on a die + an entire video accelerator? Did you plan on doing UMA? Please tell me this isn't the unmitigated disaster it appears to be...
However, when it comes to 64bit linux, the AMD chips are arguably better performing than the core2duo. Never mind the price - AMD already wins there - Im saying that AMD64 X2's run 64bit linux better than Intel Core2Duos. People BUY these dual core AMD CPU's because they make great linux boxes.
What?
Can you please elaborate on any of these points or cite something? Are you referring to the fact that Core 2 has less of a performance delta between 32- and 64-bit than Athlon 64? Or AMD's memory architecture advantage in multi-socket boxes? Neither of those factors is Linux-specific. The ISA is identical between the two, the same binaries work, optimization support is roughly equal, there are no software incompatibilities or unsupported hardware, and Core 2 is faster, so I'm having a hard time finding a reason for why you're not talking out of your ass.
[an error occurred while processing this directive]
I'm only 30, and I've never thought the 'American Dream' was anything but 'get rich quick.' I mean, sure, maybe 200 years ago. But even in the early 1900s, all the movies show immigrants coming to America and suddenly they have nice clothes and smiles. Just from moving here.
It's been a LONG time since the 'American Dream' was portrayed as anything but 'move to America and be fat and happy'.
As for the stock market, my Dad is caught in the middle. He watches it short term (daily, ugh!) but says he wants it as a long-term investment. He curses day-traders constantly. I just stay away from it. I figure it'll eventually settle down again and be an investment, and if it doesn't, it isn't what I want anyhow.
"If you make people think they're thinking, they'll love you; But if you really make them think, they'll hate you." - DM