Conroe, AM2, and Memory controllers.

[OC4life]

First off, im not a fan boy, Ive switched between Amd and Intel so many times its countless, I LIKE CHEAP FAST PRECESSORS. I bought my 820 (sig) for $121 w/ $12shipping on ebay. This is what a OC'ing juckie wants to know and his thoughts on it.

Why didnt conroe impiment HTT (or some form of it)?

Everyone should be asking that but isnt. Its definite by now the the P-3/P-M/Israel architecture is vastly superior to anything AMD or Intel currently has to offer.A company that has 85%+- marketshare should have a vastly superior R&D driven product to a company that has 15%+- marketshare. Why amd ever took such a performance lead just baffles me. I loved AMD as an underdog but doubt them as a leader. EVERY intel board member should be fired.

A hexux report is showing a AM2 system hitting 10K in sandra memory bechies with ddr2 800 . We all know that most of this is because of AMD has a ondie memory controller that provides lower latencies as well as faster busses. WHY cant intel do this with conroe??? I know its in the works with the nehemiah chip but why not now?

X2's beat P-D's because of the memory controller. It gets 1&0's in and out of the processing unit a lot faster than a FSB.


For those who dont know a AMD64 processor memory controller is on the chip whereas the intel memory controller is on the northbridge ( a seperate chip) so the processor has to go through the FSB to get memory instuctions from the northbridge.

If a small company ( compared to intel) can successfully implement a on die memory controller for the last 3 years WHY CANT INTEL RIGHT NOW?

If they did putt htt on their chips Intel would have the fastest chip on the market today by far. AMD would become a cheapo alternate chip that they used to be. Intel needs to strike while the irons hot. And before K8l its hot right now.

 

[Gautam]

Inch by inch I guess. Can't wait for it to happen myself though.

 

[Shelnutt2]

Conroe was suppose to have an IMC and CSI (HTT). Conroe was to be the Indian (Based in India) design team's move to the big times. Up until then all processors had been designed in Israel. There were many delays and Conroe was going over budget so Intel investigated. Turns out that the Indian team had embezzled $750K USD. So what had been completed of Conroe was shipped to Israel and Israel finished up the job, but in order to get it out on time they had to axe the CSI and IMC.

Correct me if I am wrong but I belive the next architecture by Intel, not Kentsfield but the one being released in 2008, is suppose to have IMC, CSI and more.

 

[NODES]

Conroe was suppose to have an IMC and CSI (HTT). Conroe was to be the Indian (Based in India) design team's move to the big times. Up until then all processors had been designed in Israel. There were many delays and Conroe was going over budget so Intel investigated. Turns out that the Indian team had embezzled $750K USD. So what had been completed of Conroe was shipped to Israel and Israel finished up the job, but in order to get it out on time they had to axe the CSI and IMC.

Correct me if I am wrong but I belive the next architecture by Intel, not Kentsfield but the one being released in 2008, is suppose to have IMC, CSI and more.

wow such big company and we are behind AMD by what now 3yrs in the HTT game?

 

[MikeyLikesItSI]

wow such big company and we are behind AMD by what now 3yrs in the HTT game?

Pretty sick that intel has a non IMC cpu comming out that beats AMD's top IMC'd chips clock for clock...... don't you think?

 

[OC4life]

750K$!!! (I lost more than that in the .com bust). Thats nothing for intel. That cant be the answer. Very interesting info though. Where did you learn that?

So the next gen. intel chips will have a htt counterpart??

 

[OC4life]

Pretty sick that intel has a non IMC cpu comming out that beats AMD's top IMC'd chips clock for clock...... don't you think?

Kill the cat while you can!!!!!!!!! Put IMC in conroe now and dont let AMD catch up.
The FX-62 isnt far behind in gaming compared to the conroe e6700. and the cheap X2's will have a much higher % OC because of no FSB/HTT limitations. Cant see FSB going beyond 450mhz any time soon.

Who can get over 400 on air/water now?

 

[OC4life]

Pretty sick that intel has a non IMC cpu comming out that beats AMD's top IMC'd chips clock for clock...... don't you think?

Double post

 

[xFlankerx]

And the answer isn't losing $750,000; Its the fact the because of that, Conroe was moved from India to Israel. That caused the Israel team to chop off the IMC to make the due date.

The FX-62 isnt far behind in gaming compared to the conroe e6700.

Are you kidding me? I've got an AMD system, and even I know that the FX-62 is only comparable to the E6400!

The ratio of clock efficiency in [Conroe : AMD] is [1 : 1.33]. That puts a 2.0Ghz Conroe at the performance of a 2.66Ghz Athlon 64. That also puts the 2.13Ghz Conroe at the performance of the 2.8Ghz Athlon 64 FX-62.

 

[OC4life]

Are you kidding me? I've got an AMD system, and even I know that the FX-62 is only comparable to the E6400!

The ratio of clock efficiency in [Conroe : AMD] is [1 : 1.33]. That puts a 2.0Ghz Conroe at the performance of a 2.66Ghz Athlon 64. That also puts the 2.13Ghz Conroe at the performance of the 2.8Ghz Athlon 64 FX-62.

BS. The fx-62 is only a fps of less than a 6700 ( unless you calculate fear into it). BF2 and Doom are almost the same

 

[Shelnutt2]

750K$!!! (I lost more than that in the .com bust). Thats nothing for intel. That cant be the answer. Very interesting info though. Where did you learn that?

So the next gen. intel chips will have a htt counterpart??

And the answer isn't losing $750,000; Its the fact the because of that, Conroe was moved from India to Israel. That caused the Israel team to chop off the IMC to make the due date.

Thanks xFlankerx.

and yes Intel's next gen will have CSI.

 

[greenmaji]

450FSB is possible with air as long as the processor doesnt have problems with high FSB, wich conroe obviously doesn't.
so far the processors core has hit 4.5Ghz and that is pretty much the end of the line for the 10X multiplier (only $360).. SOOO... Can AMD clock that high at that price point? And we dont know what 11X or 12X will do both are cheaper then the FX-62.. I dont see how in the world current AMD's and AM2 are going to clock that high.

 

[Ross]

So the next gen. intel chips will have a htt counterpart??

Depends what you mean by "generation". Next up will be quad-cores, followed by 45nm versions of everything, followed by *server* versions with CSI (Xeon - Tukwila core). I think desktop procs with CSI probably won't be seen until '08.

FSB is going to bite them in the butt sooner or later (probably sooner with the quad cores coming quickly). The shared/dynamic L2 is helping a lot, but there's only so much of it you can fit (and it's expensive and generates a lot of heat). I'd imagine we'd see something like dual FSBs implemented on desktops long before CSI. Combined with a still decent size of shared/dynamic L2, that might be something to hold Intel over until CSI is ready to go.

450FSB is possible with air as long as the processor doesnt have problems with high FSB, wich conroe obviously doesn't.
so far the processors core has hit 4.5Ghz and that is pretty much the end of the line for the 10X multiplier (only $360).. SOOO... Can AMD clock that high at that price point? And we dont know what 11X or 12X will do both are cheaper then the FX-62.. I dont see how in the world current AMD's and AM2 are going to clock that high.

FSB will be limited by the board. I don't know what kind of cooling these guys are using on their NBs, but it ain't stock PLLs get sloppy the closer they get to max which all clocks unstable. Digital clock gen based boards will reach max much more easily/stable. That's not to say 400FSB+ wouldn't be possible with active aircooling, but running the FSB at max with any sense of stability won't happen on air or probably even water cooling for the NB/PLL chip. And the 2.40GHz is $360, which is 9x multi. 2.66 = 10x for $520 I've already heard at least one source say the 11x (2.93GHz which will supposedly be dubbed an XE now) does 4GHz on air

 

[greenmaji]

Hmmm Ross.. I didnt see anything fancy on colanders NB.. Got to go check out the pic's again..

 

[Ross]

I'll believe 450FSB on stock cooling when I see it. You don't see pics of the boards with LN2 either, but that's what they are using I've done over 400FSB with active aircooling on the NB, so I know that's easily possible, but max FSB is a totally different story.

 

[greenmaji]

Yes, but you have admited that 400FSB was the limit of the processor
And I said air, that doesnt discount heatpiped aftermarket air coolers

 

[Ross]

Hell yeah I'll admit that, but that doesn't mean a stressed PLL/NB wasn't helping the situation They also seem to be going in steps...400->425->444->450FSB. I don't know if that's the proc or if they were making some changes in areas, like NB cooling. I guess someone shoud ask them or we'll find out soon enough for ourselves.

 

[ochungry]

well because there was'nt enough room. If you read the reviews on the core architecture you will see why there isn't enough room.
Also could be their pride, not wanting to copy the little guy.
Instead of trying to break AMD's back by selling chips at loss that eventually drove dell out, they could have spent time and money on this (IMC).

 

[NODES]

Pretty sick that intel has a non IMC cpu comming out that beats AMD's top IMC'd chips clock for clock...... don't you think?

oh yeah agree with u 100%, but for maximun perforamnce they should also take the crown in diff depts. aswell

 

[greenmaji]

He just hit 460FSB.. someone with a XS membership please ask him what the heck he is cooling that north bridge with