a critique of amd by theInq (actually worth reading)
- Thread starter epicstruggle
- Start date
That's nice and all, but they seem to be talking in theoretical terms a lot and not enough in practical terms.
Does Intel actually HJAVE anything to counter attack AMD? If they did, why haven't we heard anything about it yet?
Is clocking the A64 @ >3GHz really trouble-free? If it were, then do they really think that AMD aren't doing it "because they're ok with being just slightly faster"? Don't think so...
The CPU race is kind of stagnant at the moments from both Intel and AMD, TheInq needs to understand this. They're not god, there are reasons if certain things do or do not happen.
To me, this just sounds like a usual crappy Inquirer article.
Does Intel actually HJAVE anything to counter attack AMD? If they did, why haven't we heard anything about it yet?
Is clocking the A64 @ >3GHz really trouble-free? If it were, then do they really think that AMD aren't doing it "because they're ok with being just slightly faster"? Don't think so...
The CPU race is kind of stagnant at the moments from both Intel and AMD, TheInq needs to understand this. They're not god, there are reasons if certain things do or do not happen.
To me, this just sounds like a usual crappy Inquirer article.
I believe the point of the article was to encourage AMD to make as much distance as it can from Intel.
Even though that wouldn't leave them with much room for improvement should Intel release a "killer product".
The harder the war the more happier the customer
Even though that wouldn't leave them with much room for improvement should Intel release a "killer product".
The harder the war the more happier the customer
Ingenu said:I believe the point of the article was to encourage AMD to make as much distance as it can from Intel.
Even though that wouldn't leave them with much room for improvement should Intel release a "killer product".
The harder the war the more happier the customer
I just don't think that AMD are just "waiting because they're just fast enough for the moment" as the Inquirer says. Just like Intel is not going to magically put out a product to kill AMD in one hit anytime soon.
They're both making baby steps, relatively speaking. It's the best they can do. If it were easy, we would already have new products that are much more powerful than their predecessorts, like it happens with GPUs (every generation is at least twice as powerful as the one before, sometimes more). Instead, in the CPU market things move a bit more slowly. It's just how things are. Until we get some real breakthrough's.
I agree. There's no point in putting out a device that is 20% faster than Intel's. The ASP will be the same as a device that is 5% faster (ie. around $6-700). In fact it would devaluate the rest of AMD's offerings which make of the bulk of revenue for AMD.
Much better to bin sort and stock those fast devices, so that when Intel comes out with a faster product, AMD can immidiately respond in kind and actually meet demand.
Cheers
Gubbi
Much better to bin sort and stock those fast devices, so that when Intel comes out with a faster product, AMD can immidiately respond in kind and actually meet demand.
Cheers
Gubbi
its my understanding that the memory controlers are still a problem for amd and that is whats stopping them from the 3 gh z zone , i think a respin will show improvements .
THere is a reason why intel isn't at 4ghz and there is a reason why amd isn't at 3 ghz .
Thats because it will cost them both alot of money to get there and they are both making good money in a time when pcs aren't selling well .
THere is a reason why intel isn't at 4ghz and there is a reason why amd isn't at 3 ghz .
Thats because it will cost them both alot of money to get there and they are both making good money in a time when pcs aren't selling well .
Exactly.
So i think we have confirmed once again that TheInq should just stop talking about what they don't know.
So i think we have confirmed once again that TheInq should just stop talking about what they don't know.
jvd said:
Of course, but the fact that they think that AMD is waiting for silly reasons says it all...
hoom
Veteran
I thought the main thing holding AMD back at the mo is that it needs its new foundry running so it can pump out 90nm chips off 12 inch wafers instead dribbling out 130nm chips off 6inch (or is it 8?) wafers.
But yeah, it seems very odd to me that like the top 5 AMD processors all have the same clock speed of 2.4ghz.
They at least need to get out 2.6 & preferably 2.8ghz chips.
But yeah, it seems very odd to me that like the top 5 AMD processors all have the same clock speed of 2.4ghz.
They at least need to get out 2.6 & preferably 2.8ghz chips.
I don't think AMD is holding back out of complacency. It's being held back by hard reality.
First of all, there's pretty much no way (outside of a cascaded vapor change system or LN2) that any current AMD core is going to break 3Ghz.
That's simply not going to happen. The K8 design has not shown an affinity beyond moderate to slight clock scaling in anything but cherry picked samples. It was only recently they managed to get Athlon 64s that outclocked the older Athlons. Not that it matters if the core could. The on-die memory controller enables good performance, but tying a core to something that has to interface with a 200 Mhz bus really restricts things.
AMD can't offer 3Ghz because they can't make them. Even the 90 nm parts are showing that their process has not reached a point where there are sufficient bin splits for high clocks. None of the shipped 90nm parts outclock the 130nm parts.
Even if AMD could offer 3Ghz, they can't afford to offer them. The binsplits would be horrendous, and they would devalue their entire line of processors for the sake of a speed grade less common than a X800 PE. They are going to drag their ramp out because they need some good (as in actually making money) quarters to pay for the massive startup costs for their new fab.
Worse, AMD doesn't even have the capacity to meet much more demand than what it is supplying. This isn't complacency, AMD is fully aware that it is in a very precarious position. Right now it is somewhat in the performance lead, but it is facing competition that makes more profit than AMD is worth.
AMD can't do any better than it is doing right now, and if it does any worse, it's sunk.
First of all, there's pretty much no way (outside of a cascaded vapor change system or LN2) that any current AMD core is going to break 3Ghz.
That's simply not going to happen. The K8 design has not shown an affinity beyond moderate to slight clock scaling in anything but cherry picked samples. It was only recently they managed to get Athlon 64s that outclocked the older Athlons. Not that it matters if the core could. The on-die memory controller enables good performance, but tying a core to something that has to interface with a 200 Mhz bus really restricts things.
AMD can't offer 3Ghz because they can't make them. Even the 90 nm parts are showing that their process has not reached a point where there are sufficient bin splits for high clocks. None of the shipped 90nm parts outclock the 130nm parts.
Even if AMD could offer 3Ghz, they can't afford to offer them. The binsplits would be horrendous, and they would devalue their entire line of processors for the sake of a speed grade less common than a X800 PE. They are going to drag their ramp out because they need some good (as in actually making money) quarters to pay for the massive startup costs for their new fab.
Worse, AMD doesn't even have the capacity to meet much more demand than what it is supplying. This isn't complacency, AMD is fully aware that it is in a very precarious position. Right now it is somewhat in the performance lead, but it is facing competition that makes more profit than AMD is worth.
AMD can't do any better than it is doing right now, and if it does any worse, it's sunk.
Great read, thanks for the linkage.anaqer said:
I don't agree with this
Before the opterons this is a market that amd was never , ever in. Sure they had dual athlons but those were for enthusists , they never took off in the market .
The fact that ati has 7% of that market is a big deal and the fact that it will most likely double in a year and a half makes it even more important. This is where amd is going to move foward . This is where the techs actually know hardware and not marketing.
IN the desktop area people still say well does it have intel inside or does it have a pentium .
In the desktop market i think we will see intel slowly inch towards 4 ghz and amd slowly inch to 3ghz and then they drop dual core chips on us and suddenly they have up to 5ghz chip or 6 ghz chips (Depending on how they perform and how high they cna clock them.
Before the opterons this is a market that amd was never , ever in. Sure they had dual athlons but those were for enthusists , they never took off in the market .
The fact that ati has 7% of that market is a big deal and the fact that it will most likely double in a year and a half makes it even more important. This is where amd is going to move foward . This is where the techs actually know hardware and not marketing.
IN the desktop area people still say well does it have intel inside or does it have a pentium .
In the desktop market i think we will see intel slowly inch towards 4 ghz and amd slowly inch to 3ghz and then they drop dual core chips on us and suddenly they have up to 5ghz chip or 6 ghz chips (Depending on how they perform and how high they cna clock them.
jvd said:
Thing is, this is NOT the market AMD could live off and most definitely not the market we want to see AMD move towards. The big money is in the desktop segment, in selling relatively weak and very cheap to make processors in huge quantities. Besides, even if AMD does get by after a heavy migration to the server market - it would be of little use to us, no?
anaqer said:
No its a very big deal to us .
Huge deals with companys are very good .
You try to sell a product to a company and you go to them well we have the best performance 32/64 bit server class and 3d content class processers and based off the same tech we have the best performance 32/64 bit pc chips for busniess .
You do know that the opteron is just an athlon fx but sells for even more than the fx and in higher quanitys than the fx chips don't u ?
we want amd to expand in both areas , the more they take of one market the more money they have to develop products and move foward in other ones
Yes, I know - it's exactly what's so dangerous about this.jvd said:
Since AMD can sell the same CPU for more in servers, they are NOT inclined to press the matter of regaining/increasing desktop market share. Why do you think they would want to use the profits from the server market to spend on anything but R&D for the server CPUs? There is this idea that it's no matter which area the money comes from, it's all good as long as it comes from somewhere - but it's a flawed logic. It's very tempting for AMD to sell all those few chips that they can make to whoever pays the most for it, and it's not going to do a damn bit of good to us if we have to contend with the server market as potential buyers.
anaqer said:
Good point. Actually I was having hard times to find my CPU for weeks when I was about to build my current rig - it was out of stock almost everywhere. I was waiting for some 2 weeks... then they restocked them - for 5% higher! Shortly after I bought mine everybody ran out of stock again!
Restocked - and another $10 added everywhere!
Seriously: it's ridiculous.
No you don't seem to know
Your wrong , that is exactly how they are going to increase desktop market share , they have the enthusists or as many of them as the are going to get . They need to break into the desktop market sales that are done in the busniess side , where a company goes to dell and orders hundreds or thousands of desktops for thier busniess .
and what do you think the athlon 64 is ? its a friggen server cpu with less cache . That was the point of this product. So amd can do what intel does. Designs a xenon then sells a scaled back version (But higher mhz ) to sell to the desktop sector. That is why intel makes so much money , they can take the same design and sell it all over the market .
Sure it is , that means amd isn't going to sit around with parts of its fabs closed down because there are no buyers , with the server market paying top dollar it will keep the fabs busy cranking out chips and will allow amd to invest in more state of the art plants .
Similar threads
