When is the new Phenom coming out?

I don't think anyone can answer when a 3.0GHz Phenom will retail, but the B3 stepping with fixed errata should arrive in this quarter.
 
Probably after 45nm parts hit the street. My opinion is that we will probably see 65nm parts top out at 2.6 or 2.8GHz, depending how late they finally show up. However I could be wrong and B3 could turn out to be pretty good and we will see 3-3.2GHz on the Agena B3's. God knows AMD needs a break.
 
I agree with the above; we won't see 3ghz stock Phenom processors for a while, at least not with a reasonable TDP. It may not take until 45nm (which is currently slated for 2h 08, which in all honesty probably means meager availability just as Christmas hits, with more arriving in 2009 IF they don't delay it further), but like Apollo mentioned nobody can give you an honest answer because nobody really knows. Hell, with all the changes in the roadmap, does AMD even know?
 
I just want to see the B3 steppings. I'd love to pick one up, but I use my desktop in a manner that might expose the TLB (aka, running VMware in Windows, and several different VMs in linux)...:(
 
kllrnohj said:
I just want to see the B3 steppings. I'd love to pick one up, but I use my desktop in a manner that might expose the TLB (aka, running VMware in Windows, and several different VMs in linux)...:(
Click to expand...
Last I heard, it will be around March time frame and it doesn't look they'll be delayed any further. So, it shouldn't be that much longer, probably 2 months tops.
 
Last I heard/paid attention to the B3 news, I thought AMD had stated that the TLB still existed in the B3's. Or was that just some FUD?
 
Version_3 said:
Last I heard/paid attention to the B3 news, I thought AMD had stated that the TLB still existed in the B3's. Or was that just some FUD?
Click to expand...
It was FUD and discredited a few days later.
 
quadnad said:
I agree with the above; we won't see 3ghz stock Phenom processors for a while, at least not with a reasonable TDP. It may not take until 45nm (which is currently slated for 2h 08, which in all honesty probably means meager availability just as Christmas hits, with more arriving in 2009 IF they don't delay it further), but like Apollo mentioned nobody can give you an honest answer because nobody really knows. Hell, with all the changes in the roadmap, does AMD even know?
Click to expand...

I doubt it will be 2009 before we see Shanghai and Montreal (45nm quad and octo cores, the latter being an MCM) because if AMD pussyfoots it much longer how are they going to keep above water? They can't afford it to be late!

But it seems that they've finally got the bugs sqashed, and hopefully 45nm parts will have some IPC tweaks as a cherry on top to help make it more competitive. Not to mention that these parts are rumored to have doubled L2 (up to 1MB) so that will help with throughput as well. But, again I'm not sure and what I've heard could always be wrong. If the transition of the old K8 cores to 65nm is any indication, the 45nm shrink should go relatively well. Early reports from AMD (grain of salt time) say that all is well and they already have Shanghai cores up and running.

I have high hopes for the shrink, but at the same time how they really, really fucked up on Phenom, AMD has a lot to prove, and they don't have much more time to do it.
 
Mr. Bluntman said:
I doubt it will be 2009 before we see Shanghai and Montreal (45nm quad and octo cores, the latter being an MCM) because if AMD pussyfoots it much longer how are they going to keep above water? They can't afford it to be late!

But it seems that they've finally got the bugs sqashed, and hopefully 45nm parts will have some IPC tweaks as a cherry on top to help make it more competitive. Not to mention that these parts are rumored to have doubled L2 (up to 1MB) so that will help with throughput as well. But, again I'm not sure and what I've heard could always be wrong. If the transition of the old K8 cores to 65nm is any indication, the 45nm shrink should go relatively well. Early reports from AMD (grain of salt time) say that all is well and they already have Shanghai cores up and running.

I have high hopes for the shrink, but at the same time how they really, really fucked up on Phenom, AMD has a lot to prove, and they don't have much more time to do it.
Click to expand...

To be honest, I thought that they couldn't do any worse when they continually delayed Phenom after the launch of C2D. They constantly promised more and more, and they delivered very little.

Also, what 65nm transition were you looking at? Their 65nm process never clocked as well as their 90nm process (and still don't). Sure there weren't any catastrophic problems, but their 65nm process didn't compete with their 90nm process in terms of clock speeds.....that isn't right.

To be honest, (and I've said this so many times I'm sure it annoys people) I don't see how they can compete until a new architecture arrives. They're significantly down on IPC, TDP, and process. All the implications that has for them, as well as their track record for "on-time" arrivals of products, combined with their current financial situation means that their roadmap is looking even more uncertain.
 
quadnad said:
Also, what 65nm transition were you looking at? Their 65nm process never clocked as well as their 90nm process (and still don't). Sure there weren't any catastrophic problems, but their 65nm process didn't compete with their 90nm process in terms of clock speeds.....that isn't right.
Click to expand...

Brisbanes were what I was reffering to. Right now they've got 2.8GHz cores (5400+), they're getting to release a 2.9GHz 5600+, and the G2 cores for the most part can be pushed just as far on air as the Windsor F3 cores. At least they are able to match each other through OC'ing.

Why wasn't 65nm competitive to 90nm? Because AMD didn't have as much experience and time working on 65nm compared to their 90nm process. Plus SOI is harder to scale up as well in comparison to bulk silicon, IIRC.

Given their history, it shows that AMD is generally slower to scale up clocks compared to Intel, with the very noteworthy exception of the 1GHz+ Thunderbirds. Then again, they have 11 more fabs and much more manufacturing resources so of course you will see faster parts come out from them sooner, because there are a lot more wafers to get those cherry picked dies in sellable volumes. AMD just has Dresden. (Yeah, I know, there's Chartered and IBM too, but that really doesn't count so much IMO...)
 
Yep supposed to be March, the B3 stepping and the TLB fix, maybe even the new 700 chipset.
 
Mr. Bluntman said:
Brisbanes were what I was reffering to. Right now they've got 2.8GHz cores (5400+), they're getting to release a 2.9GHz 5600+, and the G2 cores for the most part can be pushed just as far on air as the Windsor F3 cores. At least they are able to match each other through OC'ing.

Why wasn't 65nm competitive to 90nm? Because AMD didn't have as much experience and time working on 65nm compared to their 90nm process. Plus SOI is harder to scale up as well in comparison to bulk silicon, IIRC.
Click to expand...

The 65nm process took ages to become competitive (clock-speed wise) with their 90nm process. Thats what I'm talking about.
 
quadnad said:
The 65nm process took ages to become competitive (clock-speed wise) with their 90nm process. Thats what I'm talking about.
Click to expand...

IIRC, 90nm didn't exactly scale through the roof off the bat either...

But I do see your point, it did take a little longer than what it should.
 
You must log in or register to reply here.
Back
Top