why why why??? Why are the Phenoms Priced SO HIGH????

[rwayne]

what is "SOI"?

 

[randomizer]

I'm probably talking about something completely different lol. But I'm not going to explain something I barely understand myself, so I'll let somebody else do that. Google it, you'll find more info than I can tell you 😉

 

[yomamafor1]

I heard they are going SOI in 45nm, not sure after that.

It certainly looks like it...

Don't know what AMD is thinking here...

EDIT:
Here you go kittah 😀, courtesy of JJ.
http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/9BF3EC138A82223B8725708B004C6471/$file/ASIC%20Brochure_June0407_final.pdf

 

[randomizer]

Ah yes, that's what I was referring to when I said "I heard" so I actually meant "I read"

 

[turpit]

I'll take this one:



AMD has a flip flop strategy.

AM2 was a flip.

K10 was a flop.



Sorry, I just had to!

Thats just mean.....funny,..... but mean.

Besides, AM2s are great little processors. They didnt live up to the hype the fanboys laid on them, but they did what AMD said they would. And K10 isnt a flop....yes it does not even remotely approach AMDs (or the fantards) claims, but it is an improvement over K8, which was not a 'flop' either and once they get the bugs ironed out and it settles into its appropriate price range, it will sell. If it sells enough, at high enough margins, well.......that will remain to be seen.


Isnt it sad that so much dissapointment is generated by a few people who cluelessly raise expectations, then get frustrated when the actual performance is revealed. One would think they would get a clue and keep silent.

 

[jimmysmitty]

I have to address the "What is SOI" question. Basically it is the process for making the transistors that run on a chip for the past 40 years. SOI stands for Silicon on Insulator. It is what stands between the transistor and the insulator in order to make sure the transistor does what it needs to do.

The reason Intel went to Hafnium is because Silicon was starting to let the electrons roam around(leak) and cause more heat at smaller than 65nm. Hafnium allows the Intel to shrink the transistors even smaller but keps the electrons in place allowing for less energy used and less heat produced along with better performance. Think of SOI as mayonase(that thins out with holes) and Hafnium as peanut butter(which is thick and blocks).

Basically AMD doing SOI at 45nm would be amazing especially since Intel couldn't and they are the leaders in shrinking transistors. I doubt AMD will be able to do it as SOI will cause too much leakage and they will have a Prescott core x2 in terms of heat and power usage. for more on the High-K dielectric check here: http://en.wikipedia.org/wiki/High-k_dielectric (but doesn't entail Hafnium for it just the basics)

As for the guys original question I would go for either a P35 mobo or a X38 mobo as those will support Yorkfeild(Penryn Quad). I have a Asus P5K-E WiFi/AP and it as 6 Sata ports and 2 eSATA ports. There are plenty with 8 SATAII ports but most are set for DDR3 as they are the high end mobos($200+).

As for the bugs getting ironed out don't forget that they haven't even compared the Phenom to the mid end Yorkfeilds yet since they have yet to be out. Plus with revisions Intel could even lower the power usage an increase the OC ability of the chips like they did with the G0 stepping of Conroe/Kentsfeild.

*Little off topic*
As for the only other twoteams with 5 superbowl championships they are the 49ers and the Cowboys. I doubt Steelers will break to 6 first. Probably 49ers or Cowboys.

 

[turpit]

I think it's sad that some just come on here to bash AMD

One reaps what they sew. AMD could have avoided this 'bashing' had they done one thing.....;kept quiet. It wasnt the people 'bashing' AMD that said '40%', it was AMD. AMD made a claim, then failed to meet that claim. They lied. No once forced them to, it was a choice...free will. Even that by itself was not bad enough to trigger this response. The reaction you are seeing has been excaserbated by AMDs 'holier than thou', superior attitude...hitting Intel with claims of falsifying benchmarks, making 'glued' quad cores, playing dirty, etc. You cant shake a finger at someone for having their hand in the cookie jar, then try and stick your own hand in the cookie jar when you think no one is watching.....pure hypocrasy....and hypocrasy doesnt sell too well. This reaction is one of the consequences of AMDs own marketing/public relations actions.

 

[tamalero]

I heard they are going SOI in 45nm, not sure after that.

considering IBM already sayd they had "METAL GATES" too.. using hafnium.
I suppose AMD will suck that tech, who knows..

 

[yomamafor1]

The reason Intel went to Hafnium is because Silicon was starting to let the electrons roam around(leak) and cause more heat at smaller than 65nm. Hafnium allows the Intel to shrink the transistors even smaller but keps the electrons in place allowing for less energy used and less heat produced along with better performance. Think of SOI as mayonase(that thins out with holes) and Hafnium as peanut butter(which is thick and blocks).

You also watched that video on Youtube? :lol: Man I love that scientist... always cracks me up....

As for HK, the main purpose of switching to Hafnium based material is because it displays similar electrical property as SiO2, while being a lot thicker than SiO2. As a result, electrons won't "roam" around, and leak through the holes of the dielectric.

 

[jimmysmitty]

You also watched that video on Youtube? :lol: Man I love that scientist... always cracks me up....

As for HK, the main purpose of switching to Hafnium based material is because it displays similar electrical property as SiO2, while being a lot thicker than SiO2. As a result, electrons won't "roam" around, and leak through the holes of the dielectric.

Yea I did watch it(and he was funny to me too) but I knew a lot of it from college. I am going for a IT Admin degree at a 2 year college but they require a lot of basic-advanced electronic classes in order to complete it(thanks to the hard ass guy in charge).

I go to college at Pima Community College in Tucson and they have a lot of funding from Intel. In fact my friend graduated(with a 4.0+ none the less) and now works for Intel in Chandler. They have a direct program and will pay for all of your college and then give you a part time job at the FAB then if you are good you get a full time job. I would have but I prefer taking the technology and putting it together rather than making it.

Now what I find amazing about Intels new process is they state that a Penryn has 800million transistors does that mean Yorkfeild will run around 1.6billion? Or is that 800million for Yorkfeild. I can imagine that Nehalem should have about 1billion+ for a quad and then double for their Octos.

 

[erocker]

I think AMD needs Daniel $ynder more than anyone else.

But then again he has invested $100 million dollars before and still did not get a winning product.

Yeah, AMD would roll out a new "flashy" processor out every couple weeks and they would constantly fail.

 

[erocker]

I'll take this one:



AMD has a flip flop strategy.

AM2 was a flip.

K10 was a flop.



Sorry, I just had to!

939 was a flip
AM2 was a bridge
K10 was a flop off the bridge.

 

[wolverinero79]

What was 4x4 in that scenario? The diving board?

 

[K_M82]

Well some site say that AMD still have problems with their Phenom yield. Wait a few weeks if you want to buy phenom. But you got to admit that for people with some AM2 main board the phenom have great value because they can just upgrade the processor and nothing else (well maybe BIOS).

 

[exit2dos]

Now what I find amazing about Intels new process is they state that a Penryn has 800million transistors does that mean Yorkfeild will run around 1.6billion? Or is that 800million for Yorkfeild. I can imagine that Nehalem should have about 1billion+ for a quad and then double for their Octos.

Yorkfield has 820M transistors, while the upccoming Nehalem quads will "only" have 731M due to smaller cache.
http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3102&p=2

I figure that will about double with their octos, but it will depend on how they work the cache. Intel has already shown 1.9B at 32nm (granted, it's SRAM - but still impressive):
http://www.intel.com/technology/architecture-silicon/32nm/index.htm?iid=tech_arch_45nm+rhc_32nm

 

[rwayne]

One reaps what they sew. AMD could have avoided this 'bashing' had they done one thing.....;kept quiet. It wasnt the people 'bashing' AMD that said '40%', it was AMD. AMD made a claim, then failed to meet that claim. They lied. No once forced them to, it was a choice...free will. Even that by itself was not bad enough to trigger this response. The reaction you are seeing has been excaserbated by AMDs 'holier than thou', superior attitude...hitting Intel with claims of falsifying benchmarks, making 'glued' quad cores, playing dirty, etc. You cant shake a finger at someone for having their hand in the cookie jar, then try and stick your own hand in the cookie jar when you think no one is watching.....pure hypocrasy....and hypocrasy doesnt sell too well. This reaction is one of the consequences of AMDs own marketing/public relations actions.

I agree with that but I think that Henri was the one spewing most of the propaganda and he is gone now. Just the way AMD names their products sounds like they are talking tough. "Phenom", "Bulldozer", "Bobcat". Talking tough is one thing but doing it when knowing in advance you are not going to be able to walk the walk is totally another. I have really lost respect for AMD. I wish I did not buy my 64x2 back in March. If I could do it all over again I would go with Intel.