Next gen AMD mobile part - Griffen

[dmens]

Originally posted by: Zstream
"That said, HT uses far less power than a FSB, and it's cheaper to design around and manufacture."

It is on the AMD website. That is why Intel is copying it and putting it on the new cpu's.

well... says so on their own website, so it must be true.

so why is AMD is spinning a new mobile part with an heavily tweaked HT controller? because the original was so low power to begin with?

 

[Zstream]

Originally posted by: dmens

Originally posted by: Zstream
"That said, HT uses far less power than a FSB, and it's cheaper to design around and manufacture."

It is on the AMD website. That is why Intel is copying it and putting it on the new cpu's.

well... says so on their own website, so it must be true.

so why is AMD is spinning a new mobile part with an heavily tweaked HT controller? because the original was so low power to begin with?

No I did not say because it is on the website it must be true, it was a reference. Why would Intel want to copy a similar form of HT?

 

[dmens]

Originally posted by: Zstream
No I did not say because it is on the website it must be true, it was a reference. Why would Intel want to copy a similar form of HT?

first off, CSI is not a copy of HT, that's just a made up fantasy. secondly, mobile power isn't exactly one of the planned benefits for integrating the controller onto the cpu... in fact, it'd be a liability imo.

 

[Zstream]

Originally posted by: dmens

Originally posted by: Zstream
No I did not say because it is on the website it must be true, it was a reference. Why would Intel want to copy a similar form of HT?

first off, CSI is not a copy of HT, that's just a made up fantasy. secondly, mobile power isn't exactly one of the planned benefits for integrating the controller onto the cpu... in fact, it'd be a liability imo.

CSI is very similar to HT, second the reason why it uses less voltage is because it does not need to use the NB thus lowering voltage.

Straight from wiki, I know you hate to believe me.

HyperTransport facilitates power management as it is ACPI compliant. This means that changes in processor sleep states (C states) can signal changes in device states (D states), e.g. powering off disks when the CPU goes to sleep.

 

[dmens]

Originally posted by: Zstream
CSI is very similar to HT, second the reason why it uses less voltage is because it does not need to use the NB thus lowering voltage.

Straight from wiki, I know you hate to believe me.

HyperTransport facilitates power management as it is ACPI compliant. This means that changes in processor sleep states (C states) can signal changes in device states (D states), e.g. powering off disks when the CPU goes to sleep.

no hate or anything, i just don't believe you or wikipedia. i assume you meant it will use less current (not voltage), because "it does not need to use the NB", whatever that means. well all that logic is still there on the cpu, and now that it is coupled with the cpu logic, which makes it more difficult to downclock and/or undersupply.

also in regards to the power management stuff, HT doesn't facilitate anything, it can be done with any bus protocol. i have no idea where you got all jazz from.

and finally, about CSI... is it copied or not? you just retracted back from "copy" to "very similar"... so what is it?

 

[Zstream]

Originally posted by: dmens

Originally posted by: Zstream
CSI is very similar to HT, second the reason why it uses less voltage is because it does not need to use the NB thus lowering voltage.

Straight from wiki, I know you hate to believe me.

HyperTransport facilitates power management as it is ACPI compliant. This means that changes in processor sleep states (C states) can signal changes in device states (D states), e.g. powering off disks when the CPU goes to sleep.

no hate or anything, i just don't believe you or wikipedia. i assume you meant it will use less current (not voltage), because "it does not need to use the NB", whatever that means. well all that logic is still there on the cpu, and now that it is coupled with the cpu logic, which makes it more difficult to downclock and/or undersupply.

also in regards to the power management stuff, HT doesn't facilitate anything, it can be done with any bus protocol. i have no idea where you got all jazz from.

and finally, about CSI... is it copied or not? you just retracted back from "copy" to "very similar"... so what is it?

Quad core for Intel requires the use of the FSB which uses the NB thus requiring more voltage. And yes I CSI is a copy of HT, you know they have patents right? Intel can not use the same criteria for HT. It is very very similar, so similar that I am willing to call it a copy.

 

[dmens]

Originally posted by: Zstream
Quad core for Intel requires the use of the FSB which uses the NB thus requiring more voltage.

- more voltage? lolz. i hope you know the difference between current and voltage.
- so when amd integrated the controller, did it vanish into thin air? or maybe the relocated logic takes zero power or something?!? maybe you can explain that phenomenon.

Originally posted by: Zstream
And yes I CSI is a copy of HT, you know they have patents right? Intel can not use the same criteria for HT. It is very very similar, so similar that I am willing to call it a copy.

wtf same criteria? patents are meaningless; CSI is a proprietary protocol. you just assume it is a copy because CSI/HT functionality roughly overlap, keyword roughly.

amd was not the first to do this kind of interconnect, nor was it the first to have an imc, and i certainly wouldn't accuse them of copying anything. here's a clue, don't use the word "copy" so lightly, it makes you sound retarded.

 

[Viditor]

Originally posted by: Zstream

Originally posted by: dmens

Originally posted by: Zstream
No I did not say because it is on the website it must be true, it was a reference. Why would Intel want to copy a similar form of HT?

first off, CSI is not a copy of HT, that's just a made up fantasy. secondly, mobile power isn't exactly one of the planned benefits for integrating the controller onto the cpu... in fact, it'd be a liability imo.

CSI is very similar to HT, second the reason why it uses less voltage is because it does not need to use the NB thus lowering voltage.

Straight from wiki, I know you hate to believe me.

HyperTransport facilitates power management as it is ACPI compliant. This means that changes in processor sleep states (C states) can signal changes in device states (D states), e.g. powering off disks when the CPU goes to sleep.

I need to correct you here Z...CSI performs like HT, but it is entirely different.
If you understand network models, think of the difference between P2P and a Token Ring network...at least that was the case the last time I saw anything on the CSI model. Things change so this may no longer be true, but the model makes sense to me...

 

[Nemesis 1]

First things first guys . Nehalem will come in 2 packages . One with CSI and one with FSB. So 09 for the notebook Nehalem is a complete line of BS.

Second thing CSI is not the same as HT. So intel is not coping AMD . You guys act like Amd invented HT . Which is a lie . AMD copied Dec . But HT is differant than what Dec had. So did Amd copy Dec . No they didn't . Amd improved on what Dec offered .

The same as Intel . Intel isn't coping AMD or Dec they are only doing what AMD did. They are improving an existing tech. Thats It End of the lies.

Amd bringing a K8 processor based laptap in late 08 isn't going to cut it.

Besides we can argue that point much later . LETS see some K10 results.

In Jan. Of this year the AMD fanboys were running around the forums saying K10 in april/ may. Were is K10?? All were saying they could buy one in July of 07 . Were are they???
If we see A K10 desktop part in retail befor 08 I will be excited. But won't hold my breath . AMD is long on promises and short on delivery. Also very short on performance.

If it wasn't for intel we would still be paying $300 for a weak ass X3800. Amd kept prices high for along time thank God Intel brought some sanity into the market.

When we see the Intel penryn parts. prices are going to fall even further. Buy this time next year we will see sub $200 Qcores that outperform anything thats on the market today by 20%+ and on the top end penryn will outperform todays topend parts by 60% in its highest stock clocked part. AMD can't compete with that. Even if K10 is faster clock for clock.

2 weeks ago the story line was AMD is dancing in the streets over its BO parts. 2 weeks later the truth comes out . That the AMD people weren't dancing in the sreets about the BO parts. They were running into the streets because the BO parts smelled like a K9(dog) They were stinking up the place. Kill the dog.

Than it would seem you guys have missed what Intel is doing with the Penryn mobile parts. One core will shut down and the other will self O/C when there isn't enough work to keep 2 or more cores busy . I like this and its should work great. 1 penryn core should be able to roughly the same amount of work as 2 K8 based cores in the same amount of time.

Viditor your very good at pushing back Intel release time frames and moving AMD's forward.
The problem is Viditor and we all know it . In the REAL WORLD the trueth is just the other way around Intel has been bringing its products to market ahead of their time frame and AMD is way behind its.

 

[Viditor]

Originally posted by: Nemesis 1
First things first guys . Nehalem will come in 2 packages . One with CSI and one with FSB. So 09 for the notebook Nehalem is a complete line of BS.

So what (in your view) will be the specifications of a non-CSI Nehalem?
I guess you have a source for Intel's wanting to:
1. spend the extra billions to develop 2 radically different chips (FSB and CSI)
2. lose an entire cycle of Penryn by releasing Nehalem early and just eating the loss
3. actually call 2 chips with radically different architecture the same thing (the architecture required for CSI is radically different than it is for FSB)

Second thing CSI is not the same as HT. So intel is not coping AMD . You guys act like Amd invented HT . Which is a lie . AMD copied Dec . But HT is differant than what Dec had. So did Amd copy Dec . No they didn't . Amd improved on what Dec offered .

The same as Intel . Intel isn't coping AMD or Dec they are only doing what AMD did. They are improving an existing tech. Thats It End of the lies.

I don't think you get it...CSI is very different than HT. In fact it had very different goals as well. The original purpose of CSI was to create a unified platform for both Itanium and x86...in addition, it's purpose was to create a directly connected type of architecture though it uses a Token Ring design instead of a P2P design.

Amd bringing a K8 processor based laptap in late 08 isn't going to cut it.

Besides we can argue that point much later . LETS see some K10 results.

In Jan. Of this year the AMD fanboys were running around the forums saying K10 in april/ may. Were is K10?? All were saying they could buy one in July of 07 . Were are they???

Ummm...this is May. July comes after June which is next month.

If we see A K10 desktop part in retail befor 08 I will be excited. But won't hold my breath . AMD is long on promises and short on delivery. Also very short on performance.

If it wasn't for intel we would still be paying $300 for a weak ass X3800. Amd kept prices high for along time thank God Intel brought some sanity into the market.

When we see the Intel penryn parts. prices are going to fall even further. Buy this time next year we will see sub $200 Qcores that outperform anything thats on the market today by 20%+ and on the top end penryn will outperform todays topend parts by 60% in its highest stock clocked part. AMD can't compete with that. Even if K10 is faster clock for clock.

2 weeks ago the story line was AMD is dancing in the streets over its BO parts. 2 weeks later the truth comes out . That the AMD people weren't dancing in the sreets about the BO parts. They were running into the streets because the BO parts smelled like a K9(dog) They were stinking up the place. Kill the dog.

Sigh...just...sigh...I can't even read that anymore...

 

[Nemesis 1]

Originally posted by: Viditor

Originally posted by: Nemesis 1
First things first guys . Nehalem will come in 2 packages . One with CSI and one with FSB. So 09 for the notebook Nehalem is a complete line of BS.

So what (in your view) will be the specifications of a non-CSI Nehalem?
I guess you have a source for Intel's wanting to:
1. spend the extra billions to develop 2 radically different chips (FSB and CSI)
2. lose an entire cycle of Penryn by releasing Nehalem early and just eating the loss
3. actually call 2 chips with radically different architecture the same thing (the architecture required for CSI is radically different than it is for FSB)

Second thing CSI is not the same as HT. So intel is not coping AMD . You guys act like Amd invented HT . Which is a lie . AMD copied Dec . But HT is differant than what Dec had. So did Amd copy Dec . No they didn't . Amd improved on what Dec offered .

The same as Intel . Intel isn't coping AMD or Dec they are only doing what AMD did. They are improving an existing tech. Thats It End of the lies.

I don't think you get it...CSI is very different than HT. In fact it had very different goals as well. The original purpose of CSI was to create a unified platform for both Itanium and x86...in addition, it's purpose was to create a directly connected type of architecture though it uses a Token Ring design instead of a P2P design.

Amd bringing a K8 processor based laptap in late 08 isn't going to cut it.

Besides we can argue that point much later . LETS see some K10 results.

In Jan. Of this year the AMD fanboys were running around the forums saying K10 in april/ may. Were is K10?? All were saying they could buy one in July of 07 . Were are they???

Ummm...this is May. July comes after June which is next month.

If we see A K10 desktop part in retail befor 08 I will be excited. But won't hold my breath . AMD is long on promises and short on delivery. Also very short on performance.

If it wasn't for intel we would still be paying $300 for a weak ass X3800. Amd kept prices high for along time thank God Intel brought some sanity into the market.

When we see the Intel penryn parts. prices are going to fall even further. Buy this time next year we will see sub $200 Qcores that outperform anything thats on the market today by 20%+ and on the top end penryn will outperform todays topend parts by 60% in its highest stock clocked part. AMD can't compete with that. Even if K10 is faster clock for clock.

2 weeks ago the story line was AMD is dancing in the streets over its BO parts. 2 weeks later the truth comes out . That the AMD people weren't dancing in the sreets about the BO parts. They were running into the streets because the BO parts smelled like a K9(dog) They were stinking up the place. Kill the dog.

Sigh...just...sigh...I can't even read that anymore...

First off Viditor we all know nehalem is coming in 2 differant packages One CSI and one FSB unless Intel is lying to us.

Nehalem is a processor CSI is the interconnect. Thats all . You might want to look at some of the cross licensing agreements between Intel and ATI . Think ATI memoery controller. As for What Nehalem is bringing to the table . Look for It to be a five issue core probabaly 18 wide. Think Northwood C

Your 2 is really stupid . Penryn Isn't about a new design. Conroe has done its Job. Beat AMD down like a dog. Penryn is about HighK and metal gates that will carry over into Nehalem after Penryn proves its Value . Same applies to Nehalem C Nehalem C will see the introduction of 3D gates than it will move on to Geshner. Intels Plan is brillant and its working. I here a lot of people laughing about Larabee I can't wait for this GPU it should be great. Some of ATI's top people left ATI just befor the AMD take over. One of which was the lead engineer for the ATI memory controller.

The last thing anyone needs Viditor is a lesson from you on Intel Teck. You have seldom said anything factual About Intel .