Sign in with
Sign up | Sign in
Your question

AMD K8L ships Q1 2007

Tags:
Last response: in CPUs
Share
July 11, 2006 12:00:35 PM

http://www.theinquirer.net/default.aspx?article=32948

The INQ refutes yesterdays digitimes report

the INQ says its sources inside AMD say the K8L is coming even sooner than they thought.

Digitimes got its report from taiwanese MB makers...these the same guys who had no clue about AMD's 4x4 till they announced it publicly...so basically they know about as much as we know...

More about : amd k8l ships 2007

July 11, 2006 12:03:51 PM

Nice.

Despite it being from the Inq.

Will this thing slot in my AM2 socket?
July 11, 2006 12:04:34 PM

Odd, its not on their desktop roadmap.
Related resources
July 11, 2006 12:08:52 PM

well its one of 3 things

1 AMD isnt being slick...and trying to spring the K8L on Intel

2 There is not K8L till 2008 and the INQ is on drugs again

3. AMD will limit release the K8L for the server side in early 07
July 11, 2006 12:10:37 PM

1. No.

2. Wouldn't be the first time.

3. Possible but unlikely.
July 11, 2006 12:13:28 PM

Still shows the theinquirer dont know what right or wrong. 80% it dead wrong. and the 20% is a lucky guess.
July 11, 2006 12:32:17 PM

Maybe this will be the first displayed sample? With launch still in 08? I could see that as a possibility.
July 11, 2006 12:39:04 PM

Quote:
http://www.theinquirer.net/default.aspx?article=32948

The INQ refutes yesterdays digitimes report

the INQ says its sources inside AMD say the K8L is coming even sooner than they thought.

Digitimes got its report from taiwanese MB makers...these the same guys who had no clue about AMD's 4x4 till they announced it publicly...so basically they know about as much as we know...


Not the first time they're wrong and most probably they still are. But if the 3600+ ships within 2 months, it means they're accelerating times and K8L early next year will not look that far.
July 11, 2006 12:39:37 PM

It does make sense for AMD to keep quiet. Who wants a K8 processor when K8L is just around the corner? Not me!

I hope this news is true. I would love to see Intel and AMD be like ATi and nVidia; continually besting each other.
July 11, 2006 1:59:46 PM

<<<It does make sense for AMD to keep quiet. Who wants a K8 processor when K8L is just around the corner? Not me!>>>

what you just posted makes no sense....

if you are selling sprockets.... u have a warehouse full of them

your new better sprockets is abotu ready to go...

why would u blab about the new sprocket too soon when u still trying to sell all those old sprokets u have in the warehouse...
July 11, 2006 2:24:48 PM

Yes we will see K8L next year, not 2008 like some other person posted. Yes it will slot in AM2.
That roadmap you guys refer to was a joke, it was obvious to me.
July 11, 2006 2:25:52 PM

So AMD released roadmaps are jokes... :lol: 
July 11, 2006 2:36:57 PM

Quote:
Odd, its not on their desktop roadmap.


K8L is a server chip. In my opinion they may forego K8L for the first X2 desktops and just use K8. FX @ 65nm will be quad ~definitely.

Since clock speed can decrease with quad, the "X2" may be back down to 1.8 - 2.4 while the FX may be 2.6 and up.

I guess not playing the clock speed game has a method to it.

Wow, I think I just figured out why the FX will stay at 90nm. Because every 939 chip that was 1MB will be tuned to be FX64. That way they don't have to special runs for it. Just tune the current runs for power. I mean how many people can actually spike an FX with general use?
July 11, 2006 2:59:05 PM

Quote:
<<<It does make sense for AMD to keep quiet. Who wants a K8 processor when K8L is just around the corner? Not me!>>>

what you just posted makes no sense....

if you are selling sprockets.... u have a warehouse full of them

your new better sprockets is abotu ready to go...

why would u blab about the new sprocket too soon when u still trying to sell all those old sprokets u have in the warehouse...


Exactly. AMD doesn't have to blab about a new processor coming out like intel did because they've got the performance crown right now. They're just moving chips. Intel needs the press to help hang on to market share.

I do think you'll get a new quad core server chip on socket F well before you'll see it at the desktop, so if the release date is true, that's a server chip in Q1 and a desktop chip probably in Q3ish.
July 11, 2006 3:04:14 PM

Do you work for AMD that you know for a fact it will be out next year?
July 11, 2006 3:04:55 PM

Quote:
http://www.theinquirer.net/default.aspx?article=32948

The INQ refutes yesterdays digitimes report

the INQ says its sources inside AMD say the K8L is coming even sooner than they thought.

Digitimes got its report from taiwanese MB makers...these the same guys who had no clue about AMD's 4x4 till they announced it publicly...so basically they know about as much as we know...


I assume 0-inch got banned and you took over. Isn't that true?
July 11, 2006 3:20:32 PM

Nice, but ... does anybody really need a quad-core? I mean, for the most common things we use our computers for.
Look, I like AMD. My rig's based on an X2 3800+. When I bought that chip in February I weighed my options and almost bought a single core Opteron instead (for overclocking potential) but I decided to go the dual core route instead (and bump the stock 2ghz up to 2.2).
Five months later, I'm generally happy with my purchase. But I can't honestly say that I've seen huge gains or advantages from having a dual core.
I don't encode much. I cruise the net and I play a few video games. Oblivion is the newest thing I've got, and it's not optimized for dual cores.
Sometimes I think I'd be happier with a much faster single-core chip.
So excuse me if I can't get too excited about a quad-core in 1st quarter 2007.
July 11, 2006 3:32:24 PM

yeah.. read my REALISTIC PREDICTIONS FOR AMD 2006-2010

to see i how i am taking over for 9-inch.

anything else cute to say?
July 11, 2006 3:38:07 PM

Quote:
It does make sense for AMD to keep quiet. Who wants a K8 processor when K8L is just around the corner? Not me!

I hope this news is true. I would love to see Intel and AMD be like ATi and nVidia; continually besting each other.



Some how I don't think this is going to be K8L because KL is for servers, not desktops. AMD had already planned for K8 quad, so it will probably be just a 4 core K8. Then K8L can be on the server while X2L can be enthusiast level.
July 11, 2006 5:07:26 PM

Quote:
Yes we will see K8L next year, not 2008 like some other person posted. Yes it will slot in AM2.
That roadmap you guys refer to was a joke, it was obvious to me.


Of course. And in 2009, we will see quantum computing processors from AMD. And it will slot in AM3.
a c 102 à CPUs
July 11, 2006 5:37:39 PM

They don't want to have happen to them what happened to Intel with the HUGE run-up to Conroe. The "something MUCH better is upcoming!!" line we've been hearing from Intel did make people not buy AMD units...but it also rendered their current (large) stocks of NetBurst chips about as desirable as space heaters in the Sahara and as a result we've seen some very inexpensive prices on those chips. Intel can afford to lose a lot of money that way and keep afloat, but AMD cannot, so AMD has reason to sandbag their future roadmap so that people buy K8s confidently today and pay full price for them.
a c 102 à CPUs
July 11, 2006 5:43:21 PM

I was very seriously debating whether I would sink the big bucks to go get a dual-DC Opteron workstation as I can and do regularly peg both cores of my X2 4200+ and can do that to more cores if I had them. However, budgetary constraints pretty much ensured that I could not get one as this was my first desktop build and I had to get things like a monitor and those aren't cheap. Sure, *games* don't usually need more than 1 or 2 cores, but real work can usually use them.
July 11, 2006 5:53:56 PM

Has TheInquirer.net EVER been right about a story?

Did a 7900 GTX have 32 pipes? Nope.
July 11, 2006 5:58:45 PM

Quote:
2 There is not K8L till 2008 and the INQ is on drugs again


This is the most likely scenario
July 11, 2006 6:26:56 PM

Quote:
2 There is not K8L till 2008 and the INQ is on drugs again


This is the most likely scenario

Is it not out of the question that AMD knows that we, as in the enthusiasts, know how much they're gonna get smashed on in the coming months and they WANT people to speculate, including the Inquirer? It seems like either AMD, the motherboard companies, or the media sources are all messing with our heads; is it 2008 or 2007 for K8L?? This is so ridiculou; if no one's got engineering samples for K8L right now, chances are, it's not going to come out in Q1 2007.
July 11, 2006 6:41:39 PM

The thing I don't like about this article is that it implies that a K8L based quad core will be available in Q1 2007. At the same time, what the text actually saids is that "AMD's quad core Athlon 64 X2 successor is on schedule", although it's unclear whether the one they are talking about is actually K8L based. Someone really aught to go out an confirm using the most current information whether or not Deerhound is K8 based or K8L based since that would really clear things up.

It's actually ironic that just yesterday Fuad Abazovic explicitly said:

http://www.theinquirer.net/default.aspx?article=32885
Quote:
Its response, K8L won't be ready before Q2 2007

Now, he's pulling his estimates in a quarter. What's even more interesting, is that Mr. Abazovic is the one who saids RHT doesn't exist, and he's bias has been questioned. Now, if people are to believe his Q1 2007 K8L estimate, they are also going to have to believe his RHT denial, because the reports come from the same person. I guess AMD traded RHT for K8L, which in the end is probably the right move anyways.

Now, specifically toward your K8L server, K8 desktop differentiation. Just by observation, the timing just doesn't work. Now we know that Rev G (probably a K8 Rev F die shrink) is launching on desktop in December. At the same time, we're speculating that Server is moving directly to K8L (Rev H) quad cores for Q1. The question is can AMD do both at once?

From a design perspective yes, Rev G seems to be just a die shrink anyways, and AMD has 2+ design teams right now. From a production perspective though no. Fab 36 won't be at 50% production until Q1 2007. I can't see AMD being able to sustain enough production to manufacture 2 completely different designs on less than half a fab. Take into account the yields of Rev G is low because they are just starting up although it'll ramp quickly, but Rev H yields would be even lower because its a completely new microarchitecture and it'll ramp slowly. What's worse is that quad cores will eat up all the savings from going 65nm. The only saving grace is the reduction in cache for desktop processors.

I just can't see AMD being able to manufacture sufficient numbers Rev G dual core desktops and Rev H quad core servers on less than 50% production capacity. It makes even less sense, when we take into account optimistic reports that Bulldozer appears in H1 2007 too. 3 different microarchitectures, 2 of them brand new, on a still ramping Fab 36 doesn't make much sense.

Conversely, your K8L server, K8 desktop split does seem to reflect other facts to date though. Namely, the suspicious lack of 65nm desktop processors on the roadmap and the fact that they are only mid-range and below. That could point to Fab 36 capacity going toward those quad cores, with desktop getting what little remains, which would support your split. I really can't see that being a smart idea though, since in terms of volumes the desktop market is critical, and it'd be poor form to antagonize people with older architecture 65nm processors with limited availability. If the desktop future were that bad, you'd think that Dell would have continued to keep their AMD plans limited to the server market.
July 11, 2006 9:21:09 PM

Quote:
K8L is a server chip.


WRONG! Stop posting this BS when you've proven wrong about this before.
July 13, 2006 5:37:07 AM

Look at this Inquirer article about Intel processors. Inquirer CPU Article Now try and take their info seriously.
July 13, 2006 6:08:06 PM

Quote:
Do you work for AMD that you know for a fact it will be out next year?


There is confirmation on the AMD website that K8L is due to ship in 2007.
Let's see how long it takes you to find it!!! :lol: 
July 13, 2006 6:27:10 PM

The server varient maybe... at the end of 2007... making it over a year away :roll:
July 13, 2006 7:46:31 PM

Quote:
The server varient maybe... at the end of 2007... making it over a year away :roll:



K8L will be Opteron. End of statement.
July 13, 2006 7:49:16 PM

So AMD is not going to compete with Intel on the desktop? Ever?
July 13, 2006 8:11:50 PM

Opteron is always first. The first K8 was Opteron. the first X2 was opteron. The first K8L will be Opteron.
July 13, 2006 8:48:27 PM

It's supposed to fit AM2, we'll see.
July 13, 2006 9:42:29 PM

Quote:
So AMD is not going to compete with Intel on the desktop? Ever?


They always release the Opteron models first. With Core 2 they may change that but I doubt it. Sun is waiting for it as is Dell.
July 13, 2006 9:45:18 PM

Quote:
<<<It does make sense for AMD to keep quiet. Who wants a K8 processor when K8L is just around the corner? Not me!>>>

what you just posted makes no sense....

if you are selling sprockets.... u have a warehouse full of them

your new better sprockets is abotu ready to go...

why would u blab about the new sprocket too soon when u still trying to sell all those old sprokets u have in the warehouse...



What your saying makes very good buisness sense.

It has one bad side effect. It screws any AMD fans who bought the
old sprocket. And since you are an AMD fanboy (I'm not saying that's
bad). How can this be true. I for one don't like surprises. I want to
know what's coming, so I can make an informed decision. If AMD
is holding secrets, it screws people personally. If you don't think so,
if you bought the old sprocket because you love AMD so much, and
found out the secret weapon sprocket is being released. How would
you feel? Betrayed?

The so-called evil Intel with its so-called slimey buisness practices,
lost a lot of money with P4 chips due to conroe.
July 13, 2006 9:58:58 PM

you cant feel betrayed...not with COMPUTERS....because no matter what ...there is always something better coming out that trumps what you just bought.


But usually, if you keep your ear to the ground , nothing with totally shock you .
July 13, 2006 10:18:46 PM

Quote:
BM says: K8L is a server chip.


Action_Man answer: WRONG! Stop posting this BS when you've proven wrong about this before.



a fanboy was owned part-1



K8L is not a server chip only.... it is the next cpu architecture of AMD.
July 13, 2006 10:20:00 PM

Quote:
The server varient maybe... at the end of 2007... making it over a year away :roll:



K8L will be Opteron. End of statement.




a fanboy was owned part-2



K8L is not an Opteron chip only.... it is the next cpu architecture of AMD.
July 13, 2006 10:58:46 PM

i think what BM means is that iinitially AMD will intorduce the K8L core on the high end server side

which makes sense because iniital yields of a new core are generally low.... But Server chips sales are low as wll albiet highly profitable...

as the fabs yields raise...then AMD will start to replace its upper to mid desktop line with K8Ls and pushing K8 cores to the sempron line

this is the most liekly scenario.... when all this happens is the big mystery
July 13, 2006 11:47:53 PM

Quote:
The thing I don't like about this article is that it implies that a K8L based quad core will be available in Q1 2007. At the same time, what the text actually saids is that "AMD's quad core Athlon 64 X2 successor is on schedule", although it's unclear whether the one they are talking about is actually K8L based. Someone really aught to go out an confirm using the most current information whether or not Deerhound is K8 based or K8L based since that would really clear things up.

It's actually ironic that just yesterday Fuad Abazovic explicitly said:

http://www.theinquirer.net/default.aspx?article=32885
Its response, K8L won't be ready before Q2 2007

Now, he's pulling his estimates in a quarter. What's even more interesting, is that Mr. Abazovic is the one who saids RHT doesn't exist, and he's bias has been questioned. Now, if people are to believe his Q1 2007 K8L estimate, they are also going to have to believe his RHT denial, because the reports come from the same person. I guess AMD traded RHT for K8L, which in the end is probably the right move anyways.

Now, specifically toward your K8L server, K8 desktop differentiation. Just by observation, the timing just doesn't work. Now we know that Rev G (probably a K8 Rev F die shrink) is launching on desktop in December. At the same time, we're speculating that Server is moving directly to K8L (Rev H) quad cores for Q1. The question is can AMD do both at once?

From a design perspective yes, Rev G seems to be just a die shrink anyways, and AMD has 2+ design teams right now. From a production perspective though no. Fab 36 won't be at 50% production until Q1 2007. I can't see AMD being able to sustain enough production to manufacture 2 completely different designs on less than half a fab. Take into account the yields of Rev G is low because they are just starting up although it'll ramp quickly, but Rev H yields would be even lower because its a completely new microarchitecture and it'll ramp slowly. What's worse is that quad cores will eat up all the savings from going 65nm. The only saving grace is the reduction in cache for desktop processors.

I just can't see AMD being able to manufacture sufficient numbers Rev G dual core desktops and Rev H quad core servers on less than 50% production capacity. It makes even less sense, when we take into account optimistic reports that Bulldozer appears in H1 2007 too. 3 different microarchitectures, 2 of them brand new, on a still ramping Fab 36 doesn't make much sense.

Conversely, your K8L server, K8 desktop split does seem to reflect other facts to date though. Namely, the suspicious lack of 65nm desktop processors on the roadmap and the fact that they are only mid-range and below. That could point to Fab 36 capacity going toward those quad cores, with desktop getting what little remains, which would support your split. I really can't see that being a smart idea though, since in terms of volumes the desktop market is critical, and it'd be poor form to antagonize people with older architecture 65nm processors with limited availability. If the desktop future were that bad, you'd think that Dell would have continued to keep their AMD plans limited to the server market.


If you loo at AMDs previous strategy the better stuff goes to server. By usingQuad K8 for desktop there is a line of differentiation between those and the K8L Opterons.

Adding K8L to mobile dual will fill the slots at the three main target markets. It's even possible that the plan maybe to make X2, 1xx and 2xx quad while FX and Opteron get K8L quad. Though even FX can get along with K8 quad. AMD is just so tightlipped thatit's hard to say what they'll do from what they've said.


I think we can count on quad K8 6 months after 65nm and Opteron K8L around the same time. K8 is so flexible that any or all of these things could happen next year.

I ws thiniking abotu it today and they could really just modularize units and add on to them. So that in 2008, K8 could get 8 decoders, 4 fp units and 8 cores at 45nm. Just thinking. No flames please.

Thogu they say they are doing a whole new architecture, that is a relative term because it will still need executions units connected to decoders, etc.


I am not of the opinion that it will ship in Q1 - maybe sample, but they maybe able to push it up if they really had it a few months ago when it was mentioned at Computex. People hae to remember that AMD is not in the position to have tons of ES' floating around as Intel is.
July 14, 2006 12:44:44 AM

Quote:
The thing I don't like about this article is that it implies that a K8L based quad core will be available in Q1 2007. At the same time, what the text actually saids is that "AMD's quad core Athlon 64 X2 successor is on schedule", although it's unclear whether the one they are talking about is actually K8L based. Someone really aught to go out an confirm using the most current information whether or not Deerhound is K8 based or K8L based since that would really clear things up.

It's actually ironic that just yesterday Fuad Abazovic explicitly said:

http://www.theinquirer.net/default.aspx?article=32885
Its response, K8L won't be ready before Q2 2007

Now, he's pulling his estimates in a quarter. What's even more interesting, is that Mr. Abazovic is the one who saids RHT doesn't exist, and he's bias has been questioned. Now, if people are to believe his Q1 2007 K8L estimate, they are also going to have to believe his RHT denial, because the reports come from the same person. I guess AMD traded RHT for K8L, which in the end is probably the right move anyways.

Now, specifically toward your K8L server, K8 desktop differentiation. Just by observation, the timing just doesn't work. Now we know that Rev G (probably a K8 Rev F die shrink) is launching on desktop in December. At the same time, we're speculating that Server is moving directly to K8L (Rev H) quad cores for Q1. The question is can AMD do both at once?

From a design perspective yes, Rev G seems to be just a die shrink anyways, and AMD has 2+ design teams right now. From a production perspective though no. Fab 36 won't be at 50% production until Q1 2007. I can't see AMD being able to sustain enough production to manufacture 2 completely different designs on less than half a fab. Take into account the yields of Rev G is low because they are just starting up although it'll ramp quickly, but Rev H yields would be even lower because its a completely new microarchitecture and it'll ramp slowly. What's worse is that quad cores will eat up all the savings from going 65nm. The only saving grace is the reduction in cache for desktop processors.

I just can't see AMD being able to manufacture sufficient numbers Rev G dual core desktops and Rev H quad core servers on less than 50% production capacity. It makes even less sense, when we take into account optimistic reports that Bulldozer appears in H1 2007 too. 3 different microarchitectures, 2 of them brand new, on a still ramping Fab 36 doesn't make much sense.

Conversely, your K8L server, K8 desktop split does seem to reflect other facts to date though. Namely, the suspicious lack of 65nm desktop processors on the roadmap and the fact that they are only mid-range and below. That could point to Fab 36 capacity going toward those quad cores, with desktop getting what little remains, which would support your split. I really can't see that being a smart idea though, since in terms of volumes the desktop market is critical, and it'd be poor form to antagonize people with older architecture 65nm processors with limited availability. If the desktop future were that bad, you'd think that Dell would have continued to keep their AMD plans limited to the server market.


If you loo at AMDs previous strategy the better stuff goes to server. By usingQuad K8 for desktop there is a line of differentiation between those and the K8L Opterons.

Adding K8L to mobile dual will fill the slots at the three main target markets. It's even possible that the plan maybe to make X2, 1xx and 2xx quad while FX and Opteron get K8L quad. Though even FX can get along with K8 quad. AMD is just so tightlipped thatit's hard to say what they'll do from what they've said.


I think we can count on quad K8 6 months after 65nm and Opteron K8L around the same time. K8 is so flexible that any or all of these things could happen next year.

I ws thiniking abotu it today and they could really just modularize units and add on to them. So that in 2008, K8 could get 8 decoders, 4 fp units and 8 cores at 45nm. Just thinking. No flames please.

Thogu they say they are doing a whole new architecture, that is a relative term because it will still need executions units connected to decoders, etc.


I am not of the opinion that it will ship in Q1 - maybe sample, but they maybe able to push it up if they really had it a few months ago when it was mentioned at Computex. People hae to remember that AMD is not in the position to have tons of ES' floating around as Intel is.

Poser.
Stupid Is As Stupid Does!
July 14, 2006 12:55:41 AM

Quote:
well its one of 3 things

1 AMD isnt being slick...and trying to spring the K8L on Intel

2 There is not K8L till 2008 and the INQ is on drugs again

3. AMD will limit release the K8L for the server side in early 07


Good follow-up, I think the votes for number two are running ahead so I will need to go there -- for a few reasons:

- First AMD definitely had RHT, then they didn't.
- Intel switched to google, then they didn't.
- April 4th Prague bonaza would show working 65 nm parts, then they didn't.

Then a decent commentary here:
http://www.overclockers.com/tips00993/

Unless they have taped out and at least have first silicon in debug, it is highly unlikely they will make a Q1 goal --- of course, AMD may not be forthcoming with the tape out schedule for K8L.


AMD NEVER SAID THEY HAD DMT(RHT). the Inq and some Russian website said it. AMD reps SAID they have K8L taped out.

Leave "the Inq is on drugs again" and number two may work out. AMD will not precipitously do anything because Intel has a majority of the world market. Hype can destory a good product. People hyped AM2 while AMD said 0-10%.

They probably dont want too much attention to what any "K8L" might be capable of.
July 14, 2006 12:56:05 AM

The specs look good though.
July 14, 2006 12:58:31 AM

Quote:
The specs look good though.


Poser.
Why Do You Talk To Yourself You Crazy!
July 14, 2006 1:15:28 AM

Quote:
I would love to see Intel and AMD be like ATi and nVidia; continually besting each other.


Yes, like Ali-Frasier III - the fighters are bloody and lumpy and we sit back and cheer.
July 14, 2006 1:46:12 AM

Quote:
People hyped AM2 while AMD said 0-10%.


So how come you caught up in the hype?
July 14, 2006 2:02:17 AM

Quote:
AMD NEVER SAID THEY HAD DMT(RHT).

But some AMD fanboys were BS-ing around. For example you.
July 14, 2006 2:44:40 AM

Quote:
I don't understand your response, I did not say AMD said they had it, I was commenting on the track record of the Incubator



I had to work last Sat and Sun so I'm jet lagged.
July 14, 2006 3:45:08 AM

Quote:
I would love to see Intel and AMD be like ATi and nVidia; continually besting each other.


Yes, like Ali-Frasier III - the fighters are bloody and lumpy and we sit back and cheer.

Good analogy - Ali called one of his tactics' "rope a dope" -
Where he hung around the rope looking like he was beaten. This encouraged Frasier to pursue him, giving Ali a flurry of punches. Ali meanwhile, covered himself up during the onslaught, and Fraiser worn himself out. Fight over. Now, AMD is now on the rope - will it release
K8L? Stay tune.

I'm buying conroe, but I like a good fight.
!