ltcommander_data
Distinguished
- Dec 16, 2004
- 997
- 0
- 18,980
Not really interested in getting to involved in the thread, but I'll just say that that's not necessarily true. Even if Intel simply put 2 current B2 E6700s together in Kentsfield the TDP isn't going to be 2x65W or 130W. One of the reasons is that Kentsfield is has 1 1067MHz FSB not 2 1067MHz so the interface itself will be saving power. Also, Intel's TDP numbers are for 75% load which is in their favour since the chances of finding 4 threads that fully load all components of all 4 cores is unlikely. This means that lose sleeping transistors and work to shut down unused portions of the cores to good effect. In terms of a 100% load using all components of all 4 cores, my feeling that it'll be difficult to implement since X-Bit Labs already had to go to lengths to try to fully load (extra load really because of the 4 issue versus a regular full load geared toward 3 issue) Conroe so doing it to Kentsfield will be even harder. In a max load situation, Kentsfield would be limited by it's 1067MHz which may reduce performance potential (although the absolute performance is probably still very good), but will again mean that portions of the core will remain idle and save power.
Now all that is assuming that you are using 2 standard E6700s. However, for one thing, Intel's 65nm process is constantly improving at the same time that they are gaining more experience with Conroe production. Between the 2, the power levels of the average E6700 will drop in the next quarter before Kentsfield launches. It's almost certain right now that the E6700 doesn't use all of it's 65W TDP anyways. For instance, the 2.67GHz 5150 Woodcrest also has a 65W TDP and it has a 1333MHz FSB vs. Conroe's 1067MHz FSB. The current E6700's TDP is probably closer to 60W, the extra 5W just means they can don't have to bin as tightly. Now by Q4 2007, the actual TDP will probably be even lower, maybe say 57-58W. Now the TDP estimates for Kentsfield are either 110W or 120W so this would make sense.
Now, that is assuming that Kentsfield is a regular high-volume consumer part, but of course it's not and it's an Extreme Edition. That means, that they can easily bin tighter than the E6700. I can imagine that they can easily bin for 55W, which again points to a 110W TDP with room to spare. Something also to note in terms of binning is that the 2.33GHz 5148 Woodcrest with a 1333MHz FSB has a TDP of 40W. If this were to be the basis for Kentsfield, at 2.67GHz with a 1067MHz FSB a 50W TDP should be achieveable meaning that even a 110W TDP is conservative.
The likely reason why we aren't going to see very low TDP numbers is because Intel's is planning on releasing higher clocked models so it's best to define the standards higher now so they don't have to keep revising them. This works in Intel's favour since they don't need to bin as tightly. The other reason is that Cloverton is higher volume than Kentsfield so can't bin as tightly and it wouldn't look good to have Cloverton's TDP at 120W and Kentsfield's at 95W. Still, reports are that the B1 stepping is noticeably cooler than the B0 stepping so things are looking up. Hopefully, Intel decides on a 110W TDP rather than the 120W TDP that they seem to be more recently leaning toward.
Anyways, I seem to have splurged, which I was not meaning to do, but I haven't posted here recently so I guess I'm making up for it.
gOJDO
Distinguished
- Mar 16, 2006
- 2,309
- 1
- 19,780
Interesting, very intelligent.
Jack
That's the best BaronBS can. You know, his respect....
Jack, why are you trying to polite with a grown up stupid child?
The relationships between people are bilateral process. It is very hard to build them and very easy to destroy them. All I can see is your effort trying to talk with BBS like talking with a man. On the other side is the stupid little child just arguing to waste his time(which probably has no dimension for him in his world). Does he deserves your treatment?
JonathanDeane
Distinguished
- Mar 28, 2006
- 1,469
- 0
- 19,310
Not really interested in getting to involved in the thread, but I'll just say that that's not necessarily true. Even if Intel simply put 2 current B2 E6700s together in Kentsfield the TDP isn't going to be 2x65W or 130W. One of the reasons is that Kentsfield is has 1 1067MHz FSB not 2 1067MHz so the interface itself will be saving power. Also, Intel's TDP numbers are for 75% load which is in their favour since the chances of finding 4 threads that fully load all components of all 4 cores is unlikely. This means that lose sleeping transistors and work to shut down unused portions of the cores to good effect. In terms of a 100% load using all components of all 4 cores, my feeling that it'll be difficult to implement since X-Bit Labs already had to go to lengths to try to fully load (extra load really because of the 4 issue versus a regular full load geared toward 3 issue) Conroe so doing it to Kentsfield will be even harder. In a max load situation, Kentsfield would be limited by it's 1067MHz which may reduce performance potential (although the absolute performance is probably still very good), but will again mean that portions of the core will remain idle and save power.
Now all that is assuming that you are using 2 standard E6700s. However, for one thing, Intel's 65nm process is constantly improving at the same time that they are gaining more experience with Conroe production. Between the 2, the power levels of the average E6700 will drop in the next quarter before Kentsfield launches. It's almost certain right now that the E6700 doesn't use all of it's 65W TDP anyways. For instance, the 2.67GHz 5150 Woodcrest also has a 65W TDP and it has a 1333MHz FSB vs. Conroe's 1067MHz FSB. The current E6700's TDP is probably closer to 60W, the extra 5W just means they can don't have to bin as tightly. Now by Q4 2007, the actual TDP will probably be even lower, maybe say 57-58W. Now the TDP estimates for Kentsfield are either 110W or 120W so this would make sense.
Now, that is assuming that Kentsfield is a regular high-volume consumer part, but of course it's not and it's an Extreme Edition. That means, that they can easily bin tighter than the E6700. I can imagine that they can easily bin for 55W, which again points to a 110W TDP with room to spare. Something also to note in terms of binning is that the 2.33GHz 5148 Woodcrest with a 1333MHz FSB has a TDP of 40W. If this were to be the basis for Kentsfield, at 2.67GHz with a 1067MHz FSB a 50W TDP should be achieveable meaning that even a 110W TDP is conservative.
The likely reason why we aren't going to see very low TDP numbers is because Intel's is planning on releasing higher clocked models so it's best to define the standards higher now so they don't have to keep revising them. This works in Intel's favour since they don't need to bin as tightly. The other reason is that Cloverton is higher volume than Kentsfield so can't bin as tightly and it wouldn't look good to have Cloverton's TDP at 120W and Kentsfield's at 95W. Still, reports are that the B1 stepping is noticeably cooler than the B0 stepping so things are looking up. Hopefully, Intel decides on a 110W TDP rather than the 120W TDP that they seem to be more recently leaning toward.
Anyways, I seem to have splurged, which I was not meaning to do, but I haven't posted here recently so I guess I'm making up for it.
Thank you ! I enjoyed that one very informative and what not
gOJDO
Distinguished
- Mar 16, 2006
- 2,309
- 1
- 19,780
I know that what ever he gets it is BS. What I don't like is him trolling around. BBS is capable to destroy every intellegent discussion, and not everybody can do that. At the begining I tried to discuss with BBS, than I tried to ignore him. But both methods gived no results and BBS was trolling hardcore BS further. The only way is destroy BBS or he will destroy everything valuable around his fat stupid ass.
thematrixhazuneo
Distinguished
- Jun 17, 2005
- 84
- 0
- 18,630
AMD is working on some nifty technology for their quad core solution, in which case the cores will dynamically and independently adjust clock and voltage to keep power under control.
This, however, is different from the 4x4. The 4x4 is a dual socket MB with 2 dual core CPUs in each socket to yield 4 execution cores.
AMD will not have a quad core chip (4 core on the same die or in the same package) on a new architecture until mid 2007. The rumor is that AMD has quad core rev G designed and will be released earlier --- the rumor is unsubstantiated and it started with the Inquirer -- some forum members who claim to work at Best Buy confirmed that AMD has told them expect quad core in stores by x-mas or something like that. AMD is very ambiguous what it is that taped out and what the timings are in general.
It is hard to tell. What is currently confirmed and what we do know is that Intel will release Kentsfield in Q4 2006, AMD will also launch 4x4 by the end of the year. Kentsfield looks to be about 110 watt 4 core part at 2.67 GHz, AMD has not said what CPUs will fit into the 4x4. This will be tough for AMD in general as it will take a full 2.8 GHz part on all 4 cores to come even close to matching any kind of performance metric, these are 125 watt parts so a 4 core Intel system will run you 110 watts and 4 core AMD system (on the 4x4 platform) will run you 250 watts as it stand. This could all change of course.
Jack
8) Thanks for clearing that up for me.
After what AMD pulled in ( I forget which country) and said k8's where better then C2D and messed with a bunch of benchmarks scores. I don't trust anything that comes out of there mouth anymore. They lost all credit in my books with that stunt. Intel has did the same, but we have seen the C2D has proformed just as much as they said it would.
Like most are saying we'll have to wait and see. Nothing is for 100% yet and even tho we have seen a bit from Kentsfield. It was a first stage CPU. In lines of the 4x4 we haven't seen anything yet. But we can look at the K8's right now and get a good idea of what can happen.
clairvoyant129
Distinguished
- May 27, 2006
- 164
- 0
- 18,680
turpit
Splendid
- Feb 12, 2006
- 6,373
- 0
- 25,780
Interesting, very intelligent.
Jack
That's the best BaronBS can. You know, his respect....
Jack, why are you trying to polite with a grown up stupid child?
The relationships between people are bilateral process. It is very hard to build them and very easy to destroy them. All I can see is your effort trying to talk with BBS like talking with a man. On the other side is the stupid little child just arguing to waste his time(which probably has no dimension for him in his world). Does he deserves your treatment?
Baron is obviously playing dumb. If he can get it wrong almost every single time, he can get it right. Your the dumb one wasting your time trying to correct him. What a hypocrite, a child responding to a child with an added tantrum to it. Although one is just playing stupid, and the other one lacks any capacity to realize that, not sure which one is worse. By this time you should be ignoring him already. Nice one...
Im sorry, but you are mistaken. You've been around long enough to know BM is not playing dumb. He truly believes what he spews.
The problem is he believes himself so strongly, that he creates fictional notions to support himself. Most of the time, he cannot provide data to support the himself as it does not exist. When data is available which counters his claims, he will say it is biased, or change his theory and manipulate the data to match, then claim "I was right, thats what Ive been saying all along". In short, he is incapable of admiting when he is wrong.
If you watched the who Core 2 debate starting back in Jan/Feb, then you know Baron makes stuff up as he goes along to support his own agenda. The consistancy of his contrary nature disproves the idea that he is just playing dumb.
I know I really dont care that BM lives in a fantasy world. What I and a few others care about is that BM spreads his BS to people who lack the knowledge/information to know better.
In BMs defense, he does have some good things to say, buts he's like the boy who cried wolf. He's spewed so much BS, that often a person will miss the good simply because they're so used to the BS that they wont even bother reading what he has to say.
GiDDY_SOUL
Distinguished
- Apr 11, 2006
- 80
- 0
- 18,630
The 4x4 will most probably perform better than a theoretical Quad core K8 arc.
My guess... The 4x4 will not perform better than a Kentsfield. Though I think 4x4 will b able to perform the same as Kentsfield in most applications.
AMDs K8L will b the killer!
My guess... The 4x4 will not perform better than a Kentsfield. Though I think 4x4 will b able to perform the same as Kentsfield in most applications.
AMDs K8L will b the killer!
- Jul 11, 2006
- 21
- 0
- 18,510
jus when i was beginnin to have doubts of what my query actually was...😱
the above is a relief....seems that the topic of discussion had changed to BM...hey bro ur quiet famous around here.... :lol:
gOJDO
Distinguished
- Mar 16, 2006
- 2,309
- 1
- 19,780
I think that you are wrong. All 4 cores on the quadcore will be connected much more efficient using the crossbar, rather than the HTT link. When reworking the crossbar, they will add something they allready have. For example shared L3. The bandwidth of the 128bit IMC and the 800MHz DDR2 RAM will be more than enough to feed the quad K8. 4x4 will waste more silicon than the quad K8.
turpit
Splendid
- Feb 12, 2006
- 6,373
- 0
- 25,780
Nope, I’m not saying he's crazy, not at all.
JJ suspects its metacognative recognition, that BM lacks the intellectual capacity to recognize when he is mistaken.
I suspect it’s a little deeper than that. That it is an emotional personal perspective issue.
Ultimately, the point is moot as the causal factors do not change the end result. So let’s just take the middle ground and say he's stubborn. Nothing crazy about that.
Regardless of what the truth is, the end result is that he provides incorrect information. This is the problem in a nut shell. The false information he is providing may be misleading to individuals who come here seeking knowledge. Those people simply lack the knowledge to recognize false information.
Now, if as you theorize, BM is doing this deliberately, then I would say he was crazy, as there is no logical or socially acceptable reason to be deliberately deceptive.
But, we should seriously consider getting back on topic before we get this thread locked. Since I helped perpetuate this little "side note", I will now cease.
Peace
GiDDY_SOUL
Distinguished
- Apr 11, 2006
- 80
- 0
- 18,630
Wont the 2 IMC in the 2 different Dual cores have any effect on performance?
This is the reason why i said->
This is the reason why i said->
GiDDY_SOUL
Distinguished
- Apr 11, 2006
- 80
- 0
- 18,630
Thanks 4 ur reply.
I thought the 2 IMC would have some other advantages too, but I was wrong.
The 4x4 may have some tweaks, but I dont think thats enough to beat Intels Kentsfield!
I thought the 2 IMC would have some other advantages too, but I was wrong.
The 4x4 may have some tweaks, but I dont think thats enough to beat Intels Kentsfield!
darkhorror
Distinguished
- Sep 5, 2006
- 24
- 0
- 18,510
We shouldn't worry about the FSB on kentsfield too much, I mean we have core2's doing 400-500 mhz fsb(1600-2000 total fsb). Now if they are running stronger chipsets when kentsfield comes out we may go even higher.
My E6300 had no problem at all hitting 400 fsb from 266. not pushing it anymore untill I get permenent active cooling on northbridge.( brought it back down to 320 untill then).
Seems that when kentsfield comes out we should be able to push it farther.
My E6300 had no problem at all hitting 400 fsb from 266. not pushing it anymore untill I get permenent active cooling on northbridge.( brought it back down to 320 untill then).
Seems that when kentsfield comes out we should be able to push it farther.
JonathanDeane
Distinguished
- Mar 28, 2006
- 1,469
- 0
- 19,310
I agree with the faster FSB thing I just got done reading an anandtech article they had the FSB of a 6300 pushed up to 532mhz !!! Crazy I tell you
http://www.anandtech.com/mb/showdoc.aspx?i=2822&p=3
BaronMatrix
Splendid
- Dec 14, 2005
- 6,655
- 0
- 25,790
AMAZING!!!!!!!!!!!!!!!!!!!!!!!
The secret of life the universe and everything revealed!!!!!
More cores process data faster!!!!!!
ASTOUNDING!!!!!!!!!!!!
Alert The Joint Chiefs immediatly!
So now links don't prove anything?
Yeah, right!!!
:twisted:
BaronMatrix
Splendid
- Dec 14, 2005
- 6,655
- 0
- 25,790
BaronMatrix
Splendid
- Dec 14, 2005
- 6,655
- 0
- 25,790
BaronMatrix
Splendid
- Dec 14, 2005
- 6,655
- 0
- 25,790
That was a most pertinent "splurge", indeed!
Cheers!
"almost electrically impossible"
Unless of course they do some kind of VRM trick that doesn't run 100% of the chip when idle. I would be interested to see what load power is. I would bet that 4x4 will do as well or better than 2xx or 2xxx.
I hope Kentsfield doesn't suck, but I think that two 2.0GHz+ AM2 chips will FINALLY unlock the bandwidth potential of K8 the way Opteron does.
turpit
Splendid
- Feb 12, 2006
- 6,373
- 0
- 25,780
AMAZING!!!!!!!!!!!!!!!!!!!!!!!
The secret of life the universe and everything revealed!!!!!
More cores process data faster!!!!!!
ASTOUNDING!!!!!!!!!!!!
Alert The Joint Chiefs immediatly!
So now links don't prove anything?
Yeah, right!!!
:twisted:
Not what Im saying the BM. Not even remotely. By any stretch of the imagination. Normal, warped or otherwise.
Let me explain it for ya.
Pointing out that more cores process data faster places you squarely in the catagory:
TRENDING THREADS
-
-
-
-
-
Question Asrock B450 Steel Legend motherboard is not loading the OS installer from pendrive ?
- Started by surya_sapui
- Replies: 9
-
-
Question RAID 10 disk failure and reallocated sectors count ?
- Started by rcfant89
- Replies: 5
Facebook
Twitter
Instagram