Are AMD defeated ?

[Teaf]

I think AMD are trying to bite off more than they can chew at the moment, and we're left with lots of mediocre products rather than 1 or 2 great product lines. At the moment they are in the graphics card, cpu and motherboard chipset business. All 3 of those are very average and nothing specatular... kinda wonder if they just dropped one of them (chipsets maybe) and solidly focused on the other two, or even dropping 2 and focusing on 1, if we'd have much higher quality products from them.

They need to sell off ATI, (not to Nvidia or Intel) so that they can concentrate on their CPU side. That way they would have more money for R&D instead of running both companies into the ground. At the moment things are looking bad for them and i can't see it getting better any time soon.

I agree with pastymuncher, and personally I hope and wish they would sell off ATI to another company, just not a competitor as mentioned and have AMD stick to CPU's, and maybe chipsets. At the moment they have lots of products and nothing outstanding.

 

[mrthingyx]

AMD have made one socket change in... 2-odd years. Yes, that has irritated quite a few people. Yet Intel circumvent that by making their chipsets redundant, instead. And that - apparently - is fine?

X38 won't be backwards compatible with anything else; P35 onwards offers native 1333MHz FSB support, Core2Duo won't work on anything below a 945 chipset...

That - to me - is more irritating. And I haven't upgraded my computer in four years. They have successfully managed to alienate a much larger (as Intel has a larger marketshare) user population with each release, yet that goes by the by? Very interesting.

Should AMD sell off ATi? Why should they? ATi are making rather large amoutns of money at the moment and so - from AMD's point of view - that would be a very stupid thing to do indeed. Is that hindering ATi's development? Perhaps,. but remember ATi's current card offerings were probably in development a good while before AMD bought them.


Yes, they have made mistakes. I personally think their marketing is absolutely rubbish and they are only just getting rid of Hector Ruin(z). They will be around for a good, long while, however.


Will IBM get into the game? Maybe, but their products are typcially not CISC-based, which is how Windows works. It would be simpler for them to buy out AMD... which may not be their interests at this point in time. They are making huge amoutns of money off the back of their third-party support contracts to be concerned with... anything else at this point in time.

 

[something_totally_different]

can anyone tell me why IBM dont go into the consumer market? they seem to make cracking chips and have a huge number of FAB's and money and practically everything already

They did once before with the IBM cyrix XPU but the entire range was a failure. Also it should be noted that IBM purchased a HUGE amount of opterons for blue gene L so they do have a good working relationship.

AMD's biggest two mistakes were,

1. Buying ATI.
2. Getting rid of 939 and alienating a massive chunk of their users.

They need to sell off ATI, (not to Nvidia or Intel) so that they can concentrate on their CPU side. That way they would have more money for R&D instead of running both companies into the ground. At the moment things are looking bad for them and i can't see it getting better any time soon.

point 1 i agree with, many market analysts and amateurs like myself said that AMD had no need to buy ATi and it would only hamper their efforts in the short term. There is no reason why AMD could not have bought ATi at a later date when AMD were in a more suitable position to do so.

point 2 is incorrect. 939 was nearly 3 years old when its life was ended, Thats far more life span than you would get from your average Intel socket.

The 3 mistakes they have made in my book are as follows,

1) buying ATi at the wrong time.
2) Not copying intel and working on a cheap quad core via 2x opteron cores on one die whilst simultaneously working on Barcelona. Theres no shame in copying intels idea. god knows
intel have done it with AMD`s.
3) Bothering with AM2. they should have carried 939 for another 8 months then bumped out the AM3 boards and chip sets. AM2 has just been a waste of time and effort.

 

[gareth170]

but people still buying they cpu's/gpu's, so they seem to be doing not to bad..

e.g everybody knows the new 3800 series gpu is basicly a 2900 in performance wise but people still buying it , even the people with a 2900 are buying the 3870..

the same thing will be with the Phenom ..

 

[drunkenmaster]

They did once before with the IBM cyrix XPU but the entire range was a failure. Also it should be noted that IBM purchased a HUGE amount of opterons for blue gene L so they do have a good working relationship.



point 1 i agree with, many market analysts and amateurs like myself said that AMD had no need to buy ATi and it would only hamper their efforts in the short term. There is no reason why AMD could not have bought ATi at a later date when AMD were in a more suitable position to do so.

point 2 is incorrect. 939 was nearly 3 years old when its life was ended, Thats far more life span than you would get from your average Intel socket.

The 3 mistakes they have made in my book are as follows,

1) buying ATi at the wrong time.
2) Not copying intel and working on a cheap quad core via 2x opteron cores on one die whilst simultaneously working on Barcelona. Theres no shame in copying intels idea. god knows
intel have done it with AMD`s.
3) Bothering with AM2. they should have carried 939 for another 8 months then bumped out the AM3 boards and chip sets. AM2 has just been a waste of time and effort.

see the reality is they bought ATi at the BEST time, and the ONLY time they could.

ati had a great low/midrange product, they currently have by far the more sellable low/mid end, and its completely down to ATi, refusing to bring out a £400 part, aiming at £250 and below and having a far better featured low/mid end that made nvidia go completely against the norm for them and bring out a value very low priced high end part. nvidia wouldn't have done that if ati priced their cards higher.

but you have to look at the business model. AMD needs massive funds to really expand past this capacity problem and get more chips out to dell and the likes. the best way to do that is buy buying ati at a fairly damn good price. again banks care NOTHING for the short term. long term money with massive interest is where all banks money comes from. AMD on its own is at a crossroads, investing a lot and having problems increasing capacity. With ATi under their belt they keep a very good company and cut its R&D costs massively simply because they both don't have to repeat research, they share it. But it means to the bank, they have another massive share of another massive market thats still growing fast. That means their future earnings potential has increased massively, that in and of itself makes banks far more likely to lend money, countries more likely to offer good tax breaks for building fabs.

i don't know the figures, but they went from maybe a say 500million dollar turnover company forecast in the next few years, to a 800 million dollar turnover company, that difference changes everything when it comes to dealing with banks, investors. it also means they can offer further discounts to dell and the like, save $5 off a cpu, and another $5 off the gpu if you buy ati and through us. it means they can do so much, and there IS a reason dell have gone massively ATi lately, from being very nvidia heavy specs last year.

 

[pastymuncher]

X38 won't be backwards compatible with anything else

In what way? It's socket 775 so will take most processors that are compatiable. There are DDR2 and DDR3 versions. They have PCI-E 2.0 slots that are both 16x and backwards compatiable with PCI-E 1.0/1.1. They have IDE even if it is only 1. Several Sata 2 ports. They even have a Floppy socket. Why are'nt they backwards compatiable with anything else?

AMD screwed up by going from 939 to AM2 and that's partly why more people went over to Intel. The rest of it being because Intel has the fastest parts. I know that's why i swapped over to Intel. AM2 brought nothing new apart from DDR2 memory. There was no performance gain until they started bringing out the faster cpu's and that's nothing they could'nt have done with 939.

 

[cosmogenesis]

http://www.*****.net/content/item.php?item=10427&page=15

***** rated the 9600 against the Q6600 and guess what in many regards the dual core 6400+ bests it to. I know these are engineering samples but power drain and performance suffer relative to Intels equivilent offerings.

Even as a mid range offering spider offers the overclocking facilities but not a lot else really. the spider setup is not even much cheaper and no better performer even in true quad mode.

Damm they have got it wrong aint they this time around. Quad core is no performer even at 3 Ghz relative to Intels up and coming die shrinks etc.

 

[manveruppd]

Buying Ati was NOT a mistake - they knew it would bring them massive grief in the short term, but it was planned to pay off from 2009 onwards, when Fusion would come out. That chip is what'll enable AMD to break into new, expanding markets like integrated devices, mobiles and haldhelds, and set-top boxes. A lot of hardware manufacturers are frothing at the mouth at the thought of what they'll be able to do with one of those.

What I've read from various news sites is that the merger was MASSIVELY mishandled. They tried to consolidate their research facilities to reduce costs, and that mistakenly led to the laying off of dozens of key personnel (especially Ati guys). The ridiculous and protracted delays in both the R600 and Barcelona were in part due to this. Now it looks like they've re-hired a lot of the people they laid off and have (hopefully) found their stride. They've got some catching up to do.

I was as ****** off at 939 being scrapped as anyone (especially since 939 itself had been brought in after 754's introduction!), but what we don't take into consideration is that having an integrated memory controller makes the need to change sockets inevitable, as with each slight re-design of the CPU the logic that controls how it interacts with the motherboard changes. The benefits of HTT outweigh this need (as long as it's not you whose brand new mobo is EOL).

Barcelona won't find its target audience on the desktop with Phenom: the key market for AMD is servers. Its lower power consumption will be a VERY attractive feature there, especially with the advanced power management features and the ability to shut down individual cores comes about with future chipset releases. (I know the voltages the CPUs run at aren't much lower than Intel's, but that fails to take into account the fact that Barcelonas include the memory controller, so that the NB on Intel mobos tends to run far hotter and at higher voltages than on AMD chipsets.) Also, I reckon performance-wise it's only in a server environment that "true quad-core" as opposed to "2xdual core" will see any benefits.

 

[Deleted member 64478]

Theres only 2 makers of the chips Intel and AMD Theres no such thing as been defeat its a market. All the Intel and AMD fan boys hush and buy whatever is the best at the time.

 

[malfunkshun]

slightly OT, but over the last few days I have noticed one or two posts from people that were die hard AMD fans and users just a couple of years ago, all but kneeling down to kiss Craig Barretts behind. Times change, buy what you think is best but dont praise either company too much, you might look silly

Oh, someone mentioned AMD shifting $600million worth of shares? and someone else mentioned IBM coincidence? (just kidding ). ATi? genius acquisition imo, extremely brave on the price but still a shrewd move. Who else can offer a package like Spider? In the same way that Intel chipsets are considered the best for Intel processors probably because they have absolute knowledge of the CPUs they are building the chipsets for, AMD have the potential to build themselves a similar reputation. With the chipset IPs they got with ATi, add to that the intention, and arguably the ability, to make those core logics with the best integrated graphics solutions built into that core logic, and with the absolute knowledge of the CPUs that these chipsets are designed for, potentially, the office machines market would be theirs.....if the price is right, and it is increasingly likely it will be. But at what cost for AMD, so long as they get the product right (Fusion seems to be attracting a wedge of their RnD cash so they had better) and can build it and sell it at the right price point, the future looks bright. Thing is, Intel and AMD are more interested in the high volume mainstream/office computer markets, that is where the money is. The enthusiasts market brings in very little and is essentially a PR/willy waving excersize.


Edit; With Fusion afaik, AMD are intending to offload geometry processing etc to the CPU when a discrete GPU is installed, dont know if there is any truth in this but without having to do geomtry etc the 38XX series would perform a lot better? anyone know anything else about this?

 

[Final8y]

slightly OT, but over the last few days I have noticed one or two posts from people that were die hard AMD fans and users just a couple of years ago, all but kneeling down to kiss Craig Barretts behind. Times change, buy what you think is best but dont praise either company too much, you might look silly

Oh, someone mentioned AMD shifting $600million worth of shares? and someone else mentioned IBM coincidence? (just kidding ). ATi? genius acquisition imo, extremely brave on the price but still a shrewd move. Who else can offer a package like Spider? In the same way that Intel chipsets are considered the best for Intel processors probably because they have absolute knowledge of the CPUs they are building the chipsets for, AMD have the potential to build themselves a similar reputation. With the chipset IPs they got with ATi, add to that the intention, and arguably the ability, to make those core logics with the best integrated graphics solutions built into that core logic, and with the absolute knowledge of the CPUs that these chipsets are designed for, potentially, the office machines market would be theirs.....if the price is right, and it is increasingly likely it will be. But at what cost for AMD, so long as they get the product right (Fusion seems to be attracting a wedge of their RnD cash so they had better) and can build it and sell it at the right price point, the future looks bright. Thing is, Intel and AMD are more interested in the high volume mainstream/office computer markets, that is where the money is. The enthusiasts market brings in very little and is essentially a PR/willy waving excersize.


Edit; With Fusion afaik, AMD are intending to offload geometry processing etc to the CPU when a discrete GPU is installed, dont know if there is any truth in this but without having to do geomtry etc the 38XX series would perform a lot better? anyone know anything else about this?

Makes sense not to bother making 300-400 Pounds cards 600-800 dual setup when so few buy them & costs a lot to develop.
Make a low & medium range card at 100-150 that many will buy & the enthusiasts will get there 300-400 & 600-800 pound performance from multi cards & multi gpus on the card set up, saving ati/amd a bundle in costs.

 

[Grimley]

People are speculating they may one day buy out AMD...

... And rename it Cyberdyne Systems....

 

[jaykay]

Nope they aren't.
They have quad core cpus that aren't that far behind Intel. They are also taking up Intels approach and update much frequently. Instead of 3-4yrs. Its now 1-2yrs.
This is much better for us as speeds will increase much faster than normal.

Also with the quad core out they can compete against Intel where it means most. SERVERS. This is where the money is and will help AMD.

All AMD need to do is carry on the trend and they will survive.

 

[pastymuncher]

Yet more problems.

AMD likely to also postpone launch of 2.4GHz Phenom CPUs

Monica Chen, Taipei; Esther Lam, DIGITIMES [Thursday 22 November 2007]

Issues with AMD's new 65nm processors are not only dampening the launch schedule of the highest frequency version Phenom CPU, but may now also affect the 2.4GHz 9700 model as well, according to sources at motherboard makers in Taiwan.

Phenom 9500 and 9600 CPUs launched on November 19 at US$251 and US$283, respectively, and AMD was earlier reported to be planning to launch a 9700 model in mid December. However, this schedule is now likely to be pushed back to early 2008, sources at motherboard makers have revealed.

The delay of the higher-end CPUs is not the only disappointment expressed by the industry players as the thermal design power (TDP) of the upcoming 9700 is also drawing concern. The sources revealed that the 9700 will have a TDP of 125W while the the 2.6GHz 9900 model will have its TDP hit as much as 140W.

In light of the delay and the higher-than-expected TDP values, the motherboard makers expressed concerns about AMD's planned migration to 45nm production in 2008 with the possibility that the Socket AM3 CPU lineup, including Deneb FX, Deneb, Propus, Regor and Sargas, may all have their launch schedules postponed.

AMD responed in saying that AMD Phenom 9000 series processors are designed for 65nm production and are undergoing a smooth ramp at Fab36. AMD's strategy consists of bringing the right technology to the right markets at the right time to meet the needs of our clients and consumers.

Taken from Digitimes today.

 

[FatRakoon]

I feel that people who are slagging off AMDs over the 939 are seeing the wrogn side of this coin.

the 939 chips have outlasted Intels life span of any of their CPUs.

What AMD should have done, is simply not bothered with AM2 at all, and simply held the support for 939 a while longer, and then jumped to the AM3 when it was ready, and not bothered at all with AM2

What we will have now, is simply that the AM2 users will all of a sudden find that their brand new AM2 system is no longer supported, and yet they moved from 939 specifically to stick with AMD ( Hmmm ) and they are just going to be even more miffed than ever.

AMD only brought out AM2 to try to keep hold of some of their current fanboy base I recon.

As for getting rid of ATI, I think this would at this time be a big mistake. Sure, they should not have bought it at this time, but they have, and ATI are bringing in the money right now, so perhaps AMD had no choice? - I dont know, but they should not get rid.

 

[Foxor`]

at the end of the day no one really knows whats going to happen, i say let time and your wallet decide

 

[pastymuncher]

I feel that people who are slagging off AMDs over the 939 are seeing the wrogn side of this coin.

the 939 chips have outlasted Intels life span of any of their CPUs.

What AMD should have done, is simply not bothered with AM2 at all, and simply held the support for 939 a while longer, and then jumped to the AM3 when it was ready, and not bothered at all with AM2

What we will have now, is simply that the AM2 users will all of a sudden find that their brand new AM2 system is no longer supported, and yet they moved from 939 specifically to stick with AMD ( Hmmm ) and they are just going to be even more miffed than ever.

AMD only brought out AM2 to try to keep hold of some of their current fanboy base I recon.

As for getting rid of ATI, I think this would at this time be a big mistake. Sure, they should not have bought it at this time, but they have, and ATI are bringing in the money right now, so perhaps AMD had no choice? - I dont know, but they should not get rid.

That is exactly what they should have done. AM2 brought nothing new with it apart from ddr2 and brassed off a lot of people. They did the same with 754 as well. So that was twice i was left without an upgrade path with AMD. Personally i go with whatever is fastest at the time and don't care about whether it's AMD or Intel. Exactly the same with graphics cards as well.

 

[something_totally_different]

I feel that people who are slagging off AMDs over the 939 are seeing the wrogn side of this coin.

the 939 chips have outlasted Intels life span of any of their CPUs.

What AMD should have done, is simply not bothered with AM2 at all, and simply held the support for 939 a while longer, and then jumped to the AM3 when it was ready, and not bothered at all with AM2

What we will have now, is simply that the AM2 users will all of a sudden find that their brand new AM2 system is no longer supported, and yet they moved from 939 specifically to stick with AMD ( Hmmm ) and they are just going to be even more miffed than ever.

AMD only brought out AM2 to try to keep hold of some of their current fanboy base I recon.

As for getting rid of ATI, I think this would at this time be a big mistake. Sure, they should not have bought it at this time, but they have, and ATI are bringing in the money right now, so perhaps AMD had no choice? - I dont know, but they should not get rid.

i posted that in post 24 on the first page so id say we agree

939 was nearly 3 years old when its life was ended, Thats far more life span than you would get from your average Intel socket.

The 3 mistakes they have made in my book are as follows,

1) buying ATi at the wrong time.
2) Not copying intel and working on a cheap quad core via 2x opteron cores on one die whilst simultaneously working on Barcelona. Theres no shame in copying intels idea. god knows
intel have done it with AMD`s.
3) Bothering with AM2. they should have carried 939 for another 8 months then bumped out the AM3 boards and chip sets. AM2 has just been a waste of time and effort.

 

[FatRakoon]

i posted that in post 24 on the first page so id say we agree

So you did mate, so you did.

I agree on the copy Intel idea.

 

[something_totally_different]

So you did mate, so you did.

I agree on the copy Intel idea.

indeed, im still unsure to this day why they didnt and why they bothered keeping AM2 alive with AM2+ a very bad move that could be seen from a mile off.