AMD on 32nm

[Harrisson]

http://www.theinquirer.net/default.aspx?article=38063

For lazy ones

Mass production will begin in 2010

This technology is likely to end up being used to manufacture AMD microprocessors.

If AMD could mass deliver 32 nm cpu's in 2010, that would be a great news.

 

[Labs23]

Whoah, from that site, everything could be a hearsay... Well, we'll just hope its gonna be a fruitful year for the green's.. Right now, they're even struggling to release the Barcelonas'.. :roll:

 

[m25]

The only problem is that we're in early 2007 now and AMD is still to deliver it's 65nm K10s.

 

[Harrisson]

Barcelonas are on track, as far as I know. In the same way you could tell "Intel is strugling to release Penryn". But for some reason we only hear negative interpretations about AMD, usualy unfounded

 

[NMDante]

The only problem is that we're in early 2007 now and AMD is still to deliver it's 65nm K10s.

True, but Dell is gonna suck all those CPUs too (as per another member's adamant reasonings).

The thing about 32nm is that AMD still hasn't been able to fully test immersion lithography with 45nm, to see if it's a viable solution, let alone 32nm. Sure, they could ride IBM's R&D's coattails, but again, it might require retooling of a fab for those specific layers/processes.

Until 65nm or even 45nm is running stable for AMD, talking about 32nm is just PR.

I'm with m25 on this. Until AMD actually does start producing in quantity some newer stuff, both to OEM and channel, it's all talk.

 

[Harrisson]

NMDante, AMD isnt talking about 32nm as of yet (that I know of), its IBM plans and IMO AMD could jump on this train. 3-4 years is enough time frame for it, again IMO.

As you work for Intel, share some news when Intel is planning to switch to 32 nm (skipping possible PR )

 

[NMDante]

NMDante, AMD isnt talking about 32nm as of yet (that I know of), its IBM plans and IMO AMD could jump on this train. 3-4 years is enough time frame for it, again IMO.

As you work for Intel, share some news when Intel is planning to switch to 32 nm (skipping possible PR )

I personally wouldn't know when Intel is planning to go to 32nm. That's not what I do at work. I am just a grunt in the fab, not one of the decision makers. Sorry. I am guessing, 32nm transition is still on a roadmap somewhere.

As for AMD not talking about 32nm. Yes, I know the article is about IBM. But the OP titled it "AMD on 32nm", so I am sticking with that, along with the last sentence in the Inq. article, which is quoted in OP's first post.

3-4 years is viable for a transition, but it won't be cheap, and without really stablizing 45nm, it could be a diaster. Again, that's my opinion. If IBM has a stable 32nm process, and retooling is not a major cost factor, then AMD can, in theory, almost skip 45nm and go to 32nm. Again, that's just a theory.

 

[gloppppp]

http://www.theinquirer.net/default.aspx?article=38063

For lazy ones
Mass production will begin in 2010

This technology is likely to end up being used to manufacture AMD microprocessors.

If AMD could mass deliver 32 nm cpu's in 2010, that would be a great news. :lol: Yeah!Great for Intel!

 

[LordPope]

The only problem is that we're in early 2007 now and AMD is still to deliver it's 65nm K10s.

True, but Dell is gonna suck all those CPUs too (as per another member's adamant reasonings).

The thing about 32nm is that AMD still hasn't been able to fully test immersion lithography with 45nm, to see if it's a viable solution, let alone 32nm. Sure, they could ride IBM's R&D's coattails, but again, it might require retooling of a fab for those specific layers/processes.

Until 65nm or even 45nm is running stable for AMD, talking about 32nm is just PR.

I'm with m25 on this. Until AMD actually does start producing in quantity some newer stuff, both to OEM and channel, it's all talk.

If IBM does all the legwork and AMD gets to coatail it .... then AMD will get 32um out alot earlier..... but they still have to tool a fab for it...

 

[LordPope]

Barcelonas are on track, as far as I know. In the same way you could tell "Intel is strugling to release Penryn". But for some reason we only hear negative interpretations about AMD, usualy unfounded

people tend to buy what INTEL is selling alot easier because INTEL has all the dollars.... AMD is smaller so people tend to think they cant get it done...or take a SHOW ME stance...

 

[Periander]

Uh..where is the news here? IBM is working on 32nm? Duh. It's on the roadmap for 2010? Once again duh, but remember that roadmaps don't always translate to reality and there is a hell of a lot more to any process than just photomasks.

Right now IBM/AMD needs to get .65 yields and performance up to par. Intel has always been ahead in process tech, but I can't recall a time that they have been as far ahead as they seem to be now.

 

[LordPope]

Whoah, from that site, everything could be a hearsay... Well, we'll just hope its gonna be a fruitful year for the green's.. Right now, they're even struggling to release the Barcelonas'.. :roll:

What makes you think they are struggling to put out Barcelona..... is there a roadmap i havent seen that AMD put out that shows the K10 to be out earlier?

 

[MarcusL]

3 years to make 2 process steps when 65nm is not even baked is a pretty bold announcement. Trying to skip 45nm could be a nightmare since the lessons learned at 45nm are crucial to a smooth 32nm.

Technical issues aside, full production on 32nm in 2010 means that AMD needs to drop $3-4 billion right now to build the fab... I don't see that happening unless they managed to lease some fab capacity from Intel.

 

[LordPope]

the NY FAB?

they getting that pretty much free....they could build that for 32um

 

[bixplus]

the NY FAB?

they getting that pretty much free....they could build that for 32um

Not quite free...a $1 billion dollar subsidy isn't enough to build a fab these days. It'll take another $2-$3 billion more.

 

[sailer]

Barcelonas are on track, as far as I know. In the same way you could tell "Intel is strugling to release Penryn". But for some reason we only hear negative interpretations about AMD, usualy unfounded

Actually, speaking as a former stockholder of AMD, I think the bad news about AMD is founded. Ever since last fall, starting with the FX70-74 disaster, AMD hasn't come out with much of anything. Oh yes, a new motherboard and that's all well and good, but that motherboard was in the works before AMD bought ATI. The R600 was supposed to be out last fall, then got delayed, and delayed again, and delayed even again. The promised new cpu's get delayed over and over. The company execs are selling out their own personal shares of AMD. No exec has bought any AMD stock this year, only sold.

Having mainly bought AMD processors and ATI video cards for the past few years, I'd like to see some new stuff come out, but for now, we get delays and excuses. There comes a point when even the faithful start opening their eyes and seeing reality. At least some do.

 

[Major_Spittle]

Barcelonas are on track, as far as I know. In the same way you could tell "Intel is strugling to release Penryn". But for some reason we only hear negative interpretations about AMD, usualy unfounded

When did I start hearing about Barcelona by AMD and how it would be a C2D killer...... Oh yeah, a long time ago. When did Intel start touting Penryn as being the end all CPU..... Oh yeah, just recently.

How long has Intel had 45nm processors released, but no Penryn?
How long has AMD had 65nm processors out but no Barcelona?

Me thinks Barcelona has issues.

 

[crazypyro]

The only problem is that we're in early 2007 now and AMD is still to deliver it's 65nm K10s.

True, but Dell is gonna suck all those CPUs too (as per another member's adamant reasonings).

The thing about 32nm is that AMD still hasn't been able to fully test immersion lithography with 45nm, to see if it's a viable solution, let alone 32nm. Sure, they could ride IBM's R&D's coattails, but again, it might require retooling of a fab for those specific layers/processes.

Until 65nm or even 45nm is running stable for AMD, talking about 32nm is just PR.

I'm with m25 on this. Until AMD actually does start producing in quantity some newer stuff, both to OEM and channel, it's all talk.

the trouble with AMDs 65nm chips now, is they're scaling a 90nm to 65nm. ATi had the same problem just scaling there 90nm chips to 80nm, nVidia reported problems too thats why they didn't transition/haven't so quickly. If AMD can design a 32nm part from scratch then they won't need 45nm, they can skip right into 32, providing IBM and partner can make a stable 32nm process.

 

[LordPope]

Barcelonas are on track, as far as I know. In the same way you could tell "Intel is strugling to release Penryn". But for some reason we only hear negative interpretations about AMD, usualy unfounded

Actually, speaking as a former stockholder of AMD, I think the bad news about AMD is founded. Ever since last fall, starting with the FX70-74 disaster, AMD hasn't come out with much of anything. Oh yes, a new motherboard and that's all well and good, but that motherboard was in the works before AMD bought ATI. The R600 was supposed to be out last fall, then got delayed, and delayed again, and delayed even again. The promised new cpu's get delayed over and over. The company execs are selling out their own personal shares of AMD. No exec has bought any AMD stock this year, only sold.

Having mainly bought AMD processors and ATI video cards for the past few years, I'd like to see some new stuff come out, but for now, we get delays and excuses. There comes a point when even the faithful start opening their eyes and seeing reality. At least some do.

in other words... AMD is done ?

 

[Lacostiade]

sure great news, but this is still too far away.

 

[gOJDO]

WOW...another amazing BS from the_INQ. :roll:

 

[Major_Spittle]

http://www.theinquirer.net/default.aspx?article=38063

For lazy ones
Mass production will begin in 2010

This technology is likely to end up being used to manufacture AMD microprocessors.

If AMD could mass deliver 32 nm cpu's in 2010, that would be a great news.

1. AMD may be a totally different company by 2010.

2. The article is talking about making 32nm Masks, not working logic chips. Nobody even knows IF 32nm is possible for Mass production at this point. (well, that has been released)

3. IBM would have to be transferring a working process to an AMD that is already tooled for 32nm and ready to produce it by fall 2009 to Mass deliver chips in 2010. IBM would never do this, they would release it themselves.

4. Not sure if Barcelona will be out by 2010. LOL

5. The only way I could see this working is IF IBM aquired AMD.

 

[sailer]

Barcelonas are on track, as far as I know. In the same way you could tell "Intel is strugling to release Penryn". But for some reason we only hear negative interpretations about AMD, usualy unfounded

Actually, speaking as a former stockholder of AMD, I think the bad news about AMD is founded. Ever since last fall, starting with the FX70-74 disaster, AMD hasn't come out with much of anything. Oh yes, a new motherboard and that's all well and good, but that motherboard was in the works before AMD bought ATI. The R600 was supposed to be out last fall, then got delayed, and delayed again, and delayed even again. The promised new cpu's get delayed over and over. The company execs are selling out their own personal shares of AMD. No exec has bought any AMD stock this year, only sold.

Having mainly bought AMD processors and ATI video cards for the past few years, I'd like to see some new stuff come out, but for now, we get delays and excuses. There comes a point when even the faithful start opening their eyes and seeing reality. At least some do.

in other words... AMD is done ?

In my opinion, for what its worth, the writing is on the wall for AMD. It might survive, but it will be a long time before anything good happens. I suppose it depends on how much they can stretch things and how many people they can get to buy stock. They are hoping to get shareholder approval at the next stockholder's meeting to issue and sell about 1.5 billion shares of new stock. I don't think they will get the permission, because that would dilute the value of the stock that is at large and I can't see any stockholder willingly cutting the value of his stock in half.

Again, in my opinion, the best hope for AMD is a buyout by IBM, or at least a large infussion of money from IBM in exchange for a lot of shares/options. That might happen, but I don't know if IBM wants to take on the debt and troubles that AMD has.

The best thing going for AMD is the OEM contracts that it has, such as with Dell. If it can provide enough low end chips to them and they keep buying, then they might survive, but they would be back to being a very second rate company again. If Dell, etc, back away from AMD, then I think AMD might well be done.

Personally, I hope that AMD does get out a cpu that will put it back on top again in the performance arena. Maybe their new cpu(s) will do that, but they have to get them on the market and not keep up the excuses and delays.

 

[gloppppp]

http://www.theinquirer.net/default.aspx?article=38063

For lazy ones
Mass production will begin in 2010

This technology is likely to end up being used to manufacture AMD microprocessors.

If AMD could mass deliver 32 nm cpu's in 2010, that would be a great news.

1. AMD may be a totally different company by 2010.

2. The article is talking about making 32nm Masks, not working logic chips. Nobody even knows IF 32nm is possible for Mass production at this point. (well, that has been released)

3. IBM would have to be transferring a working process to an AMD that is already tooled for 32nm and ready to produce it by fall 2009 to Mass deliver chips in 2010. IBM would never do this, they would release it themselves.

4. Not sure if Barcelona will be out by 2010. LOL

5. The only way I could see this working is IF IBM aquired AMD. :idea: Definite possibility(AMD would be a nice new toy for IBM).

 

[rodney_ws]

All of this discussion over an article from The Inquirer?!? LOL. Going to that site is just a waste of bandwidth and talking about it is just a waste of key strokes if you ask me.