Sign in with
Sign up | Sign in
Your question

Conroe killer ?

Last response: in CPUs
Share
August 23, 2006 2:11:12 PM

Wider execution units and overall improved architecture speak for themselves. Although highly optimized and better performer, Conroe is still older architecture than K8/K8L, it has a lot of drawbecks especially ppreventing future upgrades like plain FSB, lack of an IMC and less efficient cache architecture.
Overall performance I think will be higher than conroe.
August 23, 2006 2:22:04 PM

Quote:
Wider execution units and overall improved architecture speak for themselves. Although highly optimized and better performer, Conroe is still older architecture than K8/K8L, it has a lot of drawbecks especially ppreventing future upgrades like plain FSB, lack of an IMC and less efficient cache architecture.
Overall performance I think will be higher than conroe.


WTF? 8O
Related resources
August 23, 2006 2:28:18 PM

I know,... it will hurt conroe fanboys a lot.
August 23, 2006 2:41:37 PM

Well AMD have to do something back as the fastest AMD CPUs looks like a joke compared to conroe right now.

I think AMD and Intel both still have plenty more stuff in the lab. They like to hold back the fastest stuff from the market so that they have something to pull out quickly if/when they need to.
August 23, 2006 2:50:28 PM

By the time that's out Conroe's replacement will be here or on it's way.

As to what m25 said, really Intel will need to switch off the Northbridge BUS, but Intel already has plans to do so with CSI.
August 23, 2006 3:08:37 PM

Also it depends will it scale higher then K8 ? if not then it might not be enough.
August 23, 2006 3:31:00 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.
August 23, 2006 3:35:38 PM

Quote:
Alas, developing and perfecting a new core is a tremendously laborious and time-consuming task. Even if a new architecture is under development at AMD, it will hardly become market-ready in the next couple of years.


I wonder what they meant by that? :?

I wonder by how much K8L will beat Conroe? 5, 10, 15, 20%? Guess we will have some time to find out.
a c 473 à CPUs
August 23, 2006 3:56:25 PM

Quote:
Alas, developing and perfecting a new core is a tremendously laborious and time-consuming task. Even if a new architecture is under development at AMD, it will hardly become market-ready in the next couple of years.


I wonder what they meant by that? :?

I wonder by how much K8L will beat Conroe? 5, 10, 15, 20%? Guess we will have some time to find out.

As Xbitlabs stated, K8L is only an evolutionary product of the K8 architecture. May be calling it K8 revision L makes it more clear.

AMD needs a new core, the K10 supposedly is that core, but it's still under development. I am assuming that the K10 will be AMD's first CPU that will be able to execute 4 instructions per clock cycle.

However, this is what Xbitlabs regarding the 4 instructions per clock cycle:

Quote:

The inability to decode and retire 4 commands per clock cycle in some cases may result into tangible performance gaps in integer applications. However, it may not be of that much importance in most cases, because the typical instruction execution pace in real integer applications does not exceed 2-2.5 instructions per cycle because of the data dependence.


No one will know the K8L's potential until they can release some engineering samples for testing.
August 23, 2006 3:57:20 PM

Quote:
Wider execution units and overall improved architecture speak for themselves. Although highly optimized and better performer, Conroe is still older architecture than K8/K8L, it has a lot of drawbecks especially ppreventing future upgrades like plain FSB, lack of an IMC and less efficient cache architecture.
Overall performance I think will be higher than conroe.

How is it older? Its based on older stuff but its been reworked. Also around the same time KL8 is around Intel will have an IMC.
August 23, 2006 3:59:35 PM

Quote:
Alas, developing and perfecting a new core is a tremendously laborious and time-consuming task. Even if a new architecture is under development at AMD, it will hardly become market-ready in the next couple of years.


I wonder what they meant by that? :?

I wonder by how much K8L will beat Conroe? 5, 10, 15, 20%? Guess we will have some time to find out.

As Xbitlabs stated, K8L is only an evolutionary product of the K8 architecture. May be calling it K8 revision L makes it more clear.

AMD needs a new core, the K10 supposedly is that core, but it's still under development. I am assuming that the K10 will be AMD's first CPU that will be able to execute 4 instructions per clock cycle.

However, this is what Xbitlabs regarding the 4 instructions per clock cycle:

Quote:

The inability to decode and retire 4 commands per clock cycle in some cases may result into tangible performance gaps in integer applications. However, it may not be of that much importance in most cases, because the typical instruction execution pace in real integer applications does not exceed 2-2.5 instructions per cycle because of the data dependence.


No one will know the K8L's potential until they can release some engineering samples for testing.
I agree, we should all wait for benchmarks before we make a more solid conclusion. But, I must say those specs do look nice.
August 23, 2006 4:10:50 PM

Quote:
I know,... it will hurt conroe fanboys a lot.

Funny that you said that in reply to somebody with an AMD CPU in their sig.

Anyway, it's inevitable that AMD will release something to combat the Conroe. It's just a question of how many customers they will lose in the meantime, and what Intel does in the meantime (since K8L is still a long way out). Since Intel seems to currently be doing something right, it's not impossible to think that they have something in the works that will be even better than conroe that they will release around the same time AMD puts out their next-gen chips.

I used to buy only AMD CPUs (haven't bought intel since I got my first K6-II), but if I build a new PC I'm going C2D all the way. They just kill in price/performance right now, and nobody can deny it (well, nobody can deny it that has a logical argument at least). When AMD is back on top, I'll switch back to them. Right now they are losing, and I'm not going to pay 2x the money for a processor because of brand loyalty.
August 23, 2006 4:16:02 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.
August 23, 2006 4:27:06 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.
Kentsfield is supposed to be out in Q4 of this year according to X-Bit.
a b à CPUs
August 23, 2006 5:29:50 PM

Quote:
Wider execution units and overall improved architecture speak for themselves. Although highly optimized and better performer, Conroe is still older architecture than K8/K8L, it has a lot of drawbecks especially ppreventing future upgrades like plain FSB, lack of an IMC and less efficient cache architecture.
Overall performance I think will be higher than conroe.


Ahhh an opinion. To use the term "I think" involves an opinion to follow shortly.

Core 2 Duo (conroe) still has improvements that have yet to be implemented. Such as an Intergrated Memory Controller and a two-way high speed bus (like Hypertransport). The Core 2 Duo architecture also has the ability (with a shrink) to further add to it's FPU performance.

You're looking at it from a blinded fanboy perspective. You're seeing AMD, as a knight in shinning armor... sending out it's venerable warrior K8L to fend off evil Conroe.

Reality, yes reality has the situation pegged differently. K8 was AMD's card, Conroe was it's response, K8L is AMD's last response to Conroe before K10... Intel still has cards to play with Conroe.. still features it can add.

Reality has it that Intel is in a much better position... biased preferences aside.
August 23, 2006 5:38:06 PM

Quote:
The fate of K8L depends on the potential of AM2 65nm. i.e. scalibility
That is if AMD can master 65nm+SOI right.

And if AMD can bring AM2 65nm to market before November then they'll have a bigger chance with K8L, if not then the chance maybe slim and further delaying K8L.

Something has to keep Core2Duo's price low and that's all I care about.


AMD getting 65nm to us by November sounds highly unlikely to me. From what i have read/heard 65nm AMD chips won't be available till Q1 '07.
August 23, 2006 5:38:44 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
August 23, 2006 5:46:30 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
First, Penryn is a 45nm shrink of Merom. Its competitior is (AFAIK) the AMD mobile chip codenamed "Bulldozer." Now, I have no idea when Intel will release a true quad core - it took this long for them to introduce a true dual core.

K8L will probably compete against Wolfdale (The Conroe shrink) and whatever quad core intel has on the maket at that time. Conroe has shown that it can scale very high at 65nm. What happens when they take it to 45nm? Then, one wonders how K8L wil compete against Nehalem - that'll be interesting...
August 23, 2006 5:53:21 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
First, Penryn is a 45nm shrink of Merom. Its competitior is (AFAIK) the AMD mobile chip codenamed "Bulldozer." Now, I have no idea when Intel will release a true quad core - it took this long for them to introduce a true dual core.

K8L will probably compete against Wolfdale (The Conroe shrink) and whatever quad core intel has on the maket at that time. Conroe has shown that it can scale very high at 65nm. What happens when they take it to 45nm? Then, one wonders how K8L wil compete against Nehalem - that'll be interesting...

To put it all simply, the future doesn't look bright for AMD.
a b à CPUs
August 23, 2006 5:59:47 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
First, Penryn is a 45nm shrink of Merom. Its competitior is (AFAIK) the AMD mobile chip codenamed "Bulldozer." Now, I have no idea when Intel will release a true quad core - it took this long for them to introduce a true dual core.

K8L will probably compete against Wolfdale (The Conroe shrink) and whatever quad core intel has on the maket at that time. Conroe has shown that it can scale very high at 65nm. What happens when they take it to 45nm? Then, one wonders how K8L wil compete against Nehalem - that'll be interesting...

To put it all simply, the future doesn't look bright for AMD.

Oh it looks bright.. just not as clear and certain as AMD fanboys would have you believe.
August 23, 2006 6:04:13 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
First, Penryn is a 45nm shrink of Merom. Its competitior is (AFAIK) the AMD mobile chip codenamed "Bulldozer." Now, I have no idea when Intel will release a true quad core - it took this long for them to introduce a true dual core.

K8L will probably compete against Wolfdale (The Conroe shrink) and whatever quad core intel has on the maket at that time. Conroe has shown that it can scale very high at 65nm. What happens when they take it to 45nm? Then, one wonders how K8L wil compete against Nehalem - that'll be interesting...

To put it all simply, the future doesn't look bright for AMD.

Oh it looks bright.. just not as clear and certain as AMD fanboys would have you believe.

I know this is off topic, but i was wondering could you post me some benchies from your Core 2 extreme? I'm thinking of upgrading to one of those but i wana see the speed difference between my D930 @ 4.2 and an x6800.
August 23, 2006 6:18:46 PM

Don't forget that his Core 2 Duo is clocked at 3.8GHz so it'll really put your Pentium D to shame. :wink:
August 23, 2006 6:24:05 PM

Yeah i know it'll put it to shame, but i want to know if the upgrade is worth the £800 or so.
August 23, 2006 7:38:19 PM

From rumors, I've heard that quad-core will be for enthusiast only for a while, is that true?
a b à CPUs
August 23, 2006 7:51:34 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
First, Penryn is a 45nm shrink of Merom. Its competitior is (AFAIK) the AMD mobile chip codenamed "Bulldozer." Now, I have no idea when Intel will release a true quad core - it took this long for them to introduce a true dual core.

K8L will probably compete against Wolfdale (The Conroe shrink) and whatever quad core intel has on the maket at that time. Conroe has shown that it can scale very high at 65nm. What happens when they take it to 45nm? Then, one wonders how K8L wil compete against Nehalem - that'll be interesting...

To put it all simply, the future doesn't look bright for AMD.

Oh it looks bright.. just not as clear and certain as AMD fanboys would have you believe.

I know this is off topic, but i was wondering could you post me some benchies from your Core 2 extreme? I'm thinking of upgrading to one of those but i wana see the speed difference between my D930 @ 4.2 and an x6800.

What aps would you like me to test?
August 23, 2006 8:40:53 PM

Quote:
AMD needs a new core, the K10 supposedly is that core, but it's still under development. I am assuming that the K10 will be AMD's first CPU that will be able to execute 4 instructions per clock cycle.


Why do you say that, revisions of old cores can do just fine, look at Conroe, the basic design is close to 15 years old and it's doing just fine...
August 23, 2006 8:49:16 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.
Kentsfield is supposed to be out in Q4 of this year according to X-Bit.

Oh, I know. I was just goofing off about the whole Conroe fanboi thing.

In fact, I am still hoping to get an ES of Kentsfield, but the engineers I know are messing with me. The last ES sample they gave me was a PD 805, I think. They like to tease me with stuff...they told me it was an ES of X6800. Bastages...
August 23, 2006 8:59:37 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

AMD has no worries. K8L is for the server. AMDsays it maybe the next FX class chip also, but it will be a server first, so Clovertown is it's competitor not Kentsfield.

But by the time quad comes around it will be greeted by 4x4.
All I want for XMas is my 4x4.
August 23, 2006 9:04:40 PM

Quote:
Also it depends will it scale higher then K8 ? if not then it might not be enough.



Theyre not that far behind Core 2. Since it's 65nm I figure it'l get a 20-30% higher clock, which starting from 3.0GHz gives 3.5-4.0GHz. It should get close to 4 IPC, since it's just a matter of superscaling the execution and fetch to fire twicw in one clock. Wait they are doing two 128 bit loads per cycle.

K8L will be a killer. If they really push it down to FX, WOW!
August 23, 2006 9:12:25 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

AMD has no worries. K8L is for the server. AMDsays it maybe the next FX class chip also, but it will be a server first, so Clovertown is it's competitor not Kentsfield.

But by the time quad comes around it will be greeted by 4x4.
All I want for XMas is my 4x4.

Ah...I see. I didn't know that K8L was a server only chip. I thought it was for desktops.

Either way, until AMD shows a working model of 4X4, and not just say that it's "going smoothly in production", it's vaporware, as far as I'm concerned. And showing a static demo of the 65nm, without showing the actual machine or what parts were used, is again, vaporware. Hell, it would've been a boost if the 65nm chip was running on existing AM2 boards with no issues. Imagine the sales of AM2 socket boards with that fuel to feed the 65nm fire for AMD.

I would like to see both the 4X4 and AMD's 65nm at work and running benchies, even the same old benchies. Once that happens, then I can truly give an opinion of the situation at hand. Until then, it's all speculation and rumor.

It's nice to see graphs of what it can do, but it's better to actually see it in action.
August 23, 2006 9:22:49 PM

Quote:
Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.


Kentsfield is a Two-Conroes-Glued-Togeather CPU and if as predicted, AMD's quad cores are K8L they're also true quad cores and get more advantages.

Yeah, Kentsfield is 2 Conroes in one package.

But by the time K8L is out, or even paper-launched, its competitor will most probably be Penryn, which is a "true" quad core. So, it will be whole different proposition.

However, I agree with your basic assumption, I think K8L will be better than the Core architecture. I would be very disappointed if it does not outperform the Core CPUs, because K8L will be about a year after Core.

But then, Intel is set to replace Core by 2008; which would be a year after K8L.
First, Penryn is a 45nm shrink of Merom. Its competitior is (AFAIK) the AMD mobile chip codenamed "Bulldozer." Now, I have no idea when Intel will release a true quad core - it took this long for them to introduce a true dual core.

K8L will probably compete against Wolfdale (The Conroe shrink) and whatever quad core intel has on the maket at that time. Conroe has shown that it can scale very high at 65nm. What happens when they take it to 45nm? Then, one wonders how K8L wil compete against Nehalem - that'll be interesting...

To put it all simply, the future doesn't look bright for AMD.

Oh it looks bright.. just not as clear and certain as AMD fanboys would have you believe.

I know this is off topic, but i was wondering could you post me some benchies from your Core 2 extreme? I'm thinking of upgrading to one of those but i wana see the speed difference between my D930 @ 4.2 and an x6800.

What aps would you like me to test?

Hows about a 3dmark05 CPU score and a SuperPi 1m score?

Thanks in advance.
August 23, 2006 9:22:57 PM

Quote:
Yeah i know it'll put it to shame, but i want to know if the upgrade is worth the £800 or so.
What's running too slow for you? Upgrading simply for the sake of doing so is foolish.
August 23, 2006 9:26:23 PM

Quote:
Wider execution units and overall improved architecture speak for themselves. Although highly optimized and better performer, Conroe is still older architecture than K8/K8L, it has a lot of drawbecks especially ppreventing future upgrades like plain FSB, lack of an IMC and less efficient cache architecture.
Overall performance I think will be higher than conroe.


You're looking at it from a blinded fanboy perspective. You're seeing AMD, as a knight in shinning armor... sending out it's venerable warrior K8L to fend off evil Conroe.

Reality, yes reality has the situation pegged differently. K8 was AMD's card, Conroe was it's response, K8L is AMD's last response to Conroe before K10... Intel still has cards to play with Conroe.. still features it can add.
Reality has it that Intel is in a much better position... biased preferences aside.

I am just seeing AMD how it's got to be seen; it will have an answer for conroe but Intel cannot take conro and simply ADD an IMC, replace FSB etc all at once. It took them ages to get away from netburst and putting on the market something totally new 1 year after conroe... I really can't believe that.
August 23, 2006 9:26:25 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

AMD has no worries. K8L is for the server. AMDsays it maybe the next FX class chip also, but it will be a server first, so Clovertown is it's competitor not Kentsfield.

But by the time quad comes around it will be greeted by 4x4.
All I want for XMas is my 4x4.

So for Christmas you want a ridiculously overprived piece of server hardware that will run REALLY hot AND that will run slower than Kentsfield. While you could have a proper quad core that uses a lot less heat, most probably for a much lower price (2 x fx62 will costs $1600, i can't see Intel charging that for a single CPU.).

The logic of some people escapes me....
August 23, 2006 9:28:06 PM

Quote:
Yeah i know it'll put it to shame, but i want to know if the upgrade is worth the £800 or so.
What's running too slow for you? Upgrading simply for the sake of doing so is foolish.

I know it's foolish, i admit i'm a fool. But I really do like having shiny new pieces of hardware in my machine, if for no other reason than bragging rights.
August 23, 2006 9:36:16 PM

Quote:
So for Christmas you want a ridiculously overprived piece of server hardware that will run REALLY hot AND that will run slower than Kentsfield. While you could have a proper quad core that uses a lot less heat, most probably for a much lower price (2 x fx62 will costs $1600, i can't see Intel charging that for a single CPU.).

The logic of some people escapes me....
A quad core chip on a 64-bit 1066MHz BUS is going to be SLOOOOOOOOOW. :lol: 
August 23, 2006 9:38:00 PM

Quote:
So for Christmas you want a ridiculously overprived piece of server hardware that will run REALLY hot AND that will run slower than Kentsfield. While you could have a proper quad core that uses a lot less heat, most probably for a much lower price (2 x fx62 will costs $1600, i can't see Intel charging that for a single CPU.).

The logic of some people escapes me....
A quad core chip on a 64-bit 1066MHz BUS is going to be SLOOOOOOOOOW. :lol: 

That's what people said with Conroe too...Did their predictions come true?

From what i have heard they are upping the FSB to 1333mhz, possibly even higher. There should be a very little bottleneck, if there is one at all.

Only time will tell....
a c 473 à CPUs
August 23, 2006 9:44:44 PM

Quote:
AMD needs a new core, the K10 supposedly is that core, but it's still under development. I am assuming that the K10 will be AMD's first CPU that will be able to execute 4 instructions per clock cycle.


Why do you say that, revisions of old cores can do just fine, look at Conroe, the basic design is close to 15 years old and it's doing just fine...

Then what you are saying is that Conroe is an Intel 486DX CPU from circa 1991.

While I believe the fundamental building blocks and concepts are the same, there has been a lot of innovations since then that vastly improves the architecture design and performance. For example the ability of Conroe to execute 4 instructions (complex and simple) per cycle can be veiwed a revolutionary compared to the 1 instruction per cycle of yesteryear.

Intel was able to only push their Pentium 4 design so far. Netburst was originally thought to herald the era of the multi-gigahertz CPU, 5GHz+. And the Pentium 4 eventually became a very hot item indeed; literally. The transistion to 90nm did not reduce temperatures. Eventually Intel abondoned Netburst and the Pentium 4 architecture as a whole for something different. The Pentium M was more or less a PIII/4 hybrid and became the dark horse for Intel next's core architecture, Conroe.

K8L is "merely" a dressed up K8 core, fundamentally speaking. But you can only tweak a design so much before you need to abandon on old design in favor of a new one for improved performance.

I'm sure K8L will perform well since AMD has been taking their sweet time with it, but I prefer to see what the K10 is capable of.
August 23, 2006 9:49:08 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

Again, you think Intel is done with the C2D. Sit on their donkey and wait for AMD to come out with something better.
You shouldn't of smoked whatever that was.
August 23, 2006 9:51:49 PM

Denial is in their dictionary.
Defeat is not.
August 23, 2006 9:54:13 PM

Prepare for the 9-Inch/BM barrage of "i told ya so", and other such oral diarhea. :wink:


First there are two r's in diarrhea and second, I never say I told you so, especially when in this case I don't know what you're talking about.Okay...maybe not exactly...i told ya so, but you do like to pat yourself on the back.

Quote:
Maybe it's because I said that a long time ago in a post about 5000+ being the highest demand chip of the year. Also I remember saying that Core 2 prices would be jacked up too.


http://forumz.tomshardware.com/hardware/AM2-X2-5000-pap...

Quote:
I said more than a month ago that every system builder and their mother would want these for systems. They canmake more selling the system than just the chip.


Quote:
1Tanker doubted me. I knew they would disappear. Just like the 3600+. It is around but only for systems - at least they are earmarked for it.


Quote:
AMD is looking to sell 9 million chips in H2. Maybe you should try intellectual independence.I told everyone in so many words that AMD would end up in Dimension right bewteen Optiplex and Precision.

I aslo said that 5000+ would show up in retail around now and that it may be expensive due to demand.


http://forumz.tomshardware.com/hardware/AMD-X2-5000ftop...

Just a few examples of your never saying "i told ya so". Same message. Namely........."I'm Baron Matrix.....mark my words, i'm always right!!"
August 23, 2006 9:56:19 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

AMD has no worries. K8L is for the server. AMDsays it maybe the next FX class chip also, but it will be a server first, so Clovertown is it's competitor not Kentsfield.

But by the time quad comes around it will be greeted by 4x4.
All I want for XMas is my 4x4.

So for Christmas you want a ridiculously overprived piece of server hardware that will run REALLY hot AND that will run slower than Kentsfield. While you could have a proper quad core that uses a lot less heat, most probably for a much lower price (2 x fx62 will costs $1600, i can't see Intel charging that for a single CPU.).

The logic of some people escapes me....


Obviously you don't know much about it. It will have several diferent grades and at least one will be well under $1000 for two chips and the mobo, according to the last demo with VooDooPC. Even if it is 120W+, look at how much QUAD SLI uses. It's a good trade off. I want the two chips and just a good GPU.

Since I will be using it for development it is more than worth it. i will be able to run a DC, Exch Server, and three clients in VMs. I just hope they quickly raise the RAM limit to 8GB.
August 23, 2006 9:58:42 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

AMD has no worries. K8L is for the server. AMDsays it maybe the next FX class chip also, but it will be a server first, so Clovertown is it's competitor not Kentsfield.

But by the time quad comes around it will be greeted by 4x4.
All I want for XMas is my 4x4.

So for Christmas you want a ridiculously overprived piece of server hardware that will run REALLY hot AND that will run slower than Kentsfield. While you could have a proper quad core that uses a lot less heat, most probably for a much lower price (2 x fx62 will costs $1600, i can't see Intel charging that for a single CPU.).

The logic of some people escapes me....


Obviously you don't know much about it. It will have several diferent grades and at least one will be well under $1000 for two chips and the mobo, according to the last demo with VooDooPC. Even if it is 120W+, look at how much QUAD SLI uses. It's a good trade off. I want the two chips and just a good GPU.

Since I will be using it for development it is more than worth it. i will be able to run a DC, Exch Server, and three clients in VMs. I just hope they quickly raise the RAM limit to 8GB.

Surely if you just want 2 CPUs 2 x Opteron 265 or similar would be more than sufficient? It's already available.
August 23, 2006 10:10:59 PM

Quote:
That's what people said with Conroe too...Did their predictions come true?

From what i have heard they are upping the FSB to 1333mhz, possibly even higher. There should be a very little bottleneck, if there is one at all.

Only time will tell....

Actually the only people who said that were AMD "fanboys"; there is a big difference between dual core and quad core. Even at a 1333MHz BUS, Kentsfield/Clovertown may suffer from bottlenecking between the cores. Once Intel gets CSI rolling they'll really have what it takes to compete with AMD in the multicore segment.
August 23, 2006 11:35:25 PM

Quote:
I'm sure K8L will perform well since AMD has been taking their sweet time with it, but I prefer to see what the K10 is capable of.


I would bet that it will still fundamentally be a K8 but will have HW threading and 8 cores @ 45nm. It may even move to dual 256bit pipes. I think it would be a mistake for Intel to create a "new" core every two years.
August 23, 2006 11:50:00 PM

Quote:
I know,... it will hurt conroe fanboys a lot.


Yeah, the Conroe fans will be disappointed, but I wonder if the Kentsfield fans will be too? Since it might be available by K8L's release. I think K8L should be more concerned about Kentsfield than Conroe.

AMD has no worries. K8L is for the server. AMDsays it maybe the next FX class chip also, but it will be a server first, so Clovertown is it's competitor not Kentsfield.

But by the time quad comes around it will be greeted by 4x4.
All I want for XMas is my 4x4.

So for Christmas you want a ridiculously overprived piece of server hardware that will run REALLY hot AND that will run slower than Kentsfield. While you could have a proper quad core that uses a lot less heat, most probably for a much lower price (2 x fx62 will costs $1600, i can't see Intel charging that for a single CPU.).

The logic of some people escapes me....


Obviously you don't know much about it. It will have several diferent grades and at least one will be well under $1000 for two chips and the mobo, according to the last demo with VooDooPC. Even if it is 120W+, look at how much QUAD SLI uses. It's a good trade off. I want the two chips and just a good GPU.

Since I will be using it for development it is more than worth it. i will be able to run a DC, Exch Server, and three clients in VMs. I just hope they quickly raise the RAM limit to 8GB.

Surely if you just want 2 CPUs 2 x Opteron 265 or similar would be more than sufficient? It's already available.

The whole idea of 4x4 is to remove the PCI-X, SCSI/SAS, management chips, not to mention ECC. As a dev I don't need ECC. Nor do I as a gamer. Not putting those things on the mobo gives more space for PCIe x16 slots and AzaliaHD. With HW virtualization, my VMs will fly. The cheapest 2 socket board is the ASUS K8N-DL @ $212 with SLI. Because AM2 is 940 pins, the form factor stays the same and it's possible to drop the price. nForce 5xx will do the job just fine and keep AMD in nVidia's good graces.

I imagine ASUS is a partner in 4x4 and will have mobos around the end of next month. I can see DFI definitely getting a mobo out quick to keep their enthusiast-level status.
August 24, 2006 1:00:14 AM

Quote:
I am just seeing AMD how it's got to be seen; it will have an answer for conroe but Intel cannot take conro and simply ADD an IMC, replace FSB etc all at once. It took them ages to get away from netburst and putting on the market something totally new 1 year after conroe... I really can't believe that.


a. Seems that AMD will be placing a lot of additional features in K8... just to add an 'L' to it;

b. Those 'NetBurst ages' were AMD's golden times: In a way, AMD should thank Intel for that;

c. The old, 'unsuccessful' NetBurst microarchitecture still co-exists with Intel's new Core's; despite the NetBurst flop, Intel managed to turn everything upside down, in a bit more that a year;

d. Where's K8L?

Bottom line: You see... AMD is a much, much better company than what you try to make of it, with unsubstantiated, erroneous, pretentious ("claiming or demanding a position of distinction or merit, especially when unjustified", just suits perfectly) and whimsical statements about a company you "admire" so much. I'm positive that AMD would reprove it, as you're just staining their reputation.

What about 'back into reality', where & when Intel has an unbeatable product, right here, right now?! Humm?!
!