What's the next big Intel Processor to come out?

[[BB] Rick James]

What speeds will we be seeing out of it?

 

[Dillusion]

Tejas was cancelled. intel is sticking with prescott until their enginers finish the dual core design chips, the 'pentium 5' ... scheduled for Q1-Q2 2005, until then, prescott is said to reach 4GHZ

 

[Lord of Shadows]

Ive heard rumors of a dual core prescott. (can we say hot)

 

[siegecraft4]

Ive heard rumors of a dual core prescott. (can we say hot)

If thats the case AMD had better start mass producing those Prescott Survival Kits.

 

[RancidWAnnaRIot]

Ive heard rumors of a dual core prescott. (can we say hot)

As far as i know.. they aren't rumors it is absolutly true.. and i remember hearing that AMD is gonna go dual core as well..

 

[AMDXP]

Well I know the new processors will be running at about 4-6 ghz not sure on 6 but I know of 5ghz

 

[nst6563]

the 'pentium 5' ...

funny how Intel originally named their chip "Pentium" because it was a derivative of the word Penta meaning 5 and 5 was the number of transistors in the millions that the new chips contained. Now if they keep the name Pentium and put a 5 behind it then they've just named it 5 twice

At least I *think* that's where they got the Pentium name. I could be totally wrong and full of shit too and we'll all never know


anyway...I'd like to see a dual core chip. it'd be very interesting to see what kind of cooling it required if it were dual Prescott cores!

 

[Lord of Shadows]

I think that hot chip will be a cooling dream, intel will have to underclock them to keep them on air. (And the overclockers rejoice)

 

[Bodega]

i heard that one of the reasons intel scrapped the tejas was heat issues... i dont think 2 prescotts in 1 is gonna be any cooler

 

[kronchev]

Ive heard rumors of a dual core prescott. (can we say hot)

the point of dual core is to reduce heat. i dont feel like finding out HOW but its not two prescott cores on one chip

 

[Pongi]

funny how Intel originally named their chip "Pentium" because it was a derivative of the word Penta meaning 5 and 5 was the number of transistors in the millions that the new chips contained. Now if they keep the name Pentium and put a 5 behind it then they've just named it 5 twice

At least I *think* that's where they got the Pentium name. I could be totally wrong and full of shit too and we'll all never know


anyway...I'd like to see a dual core chip. it'd be very interesting to see what kind of cooling it required if it were dual Prescott cores!

I thought they came up with "Pentium" because they couldn't trademark the number 586. So being the 5th generation design they called it Pentium. That's the impression I was under.

 

[RancidWAnnaRIot]

funny how Intel originally named their chip "Pentium" because it was a derivative of the word Penta meaning 5 and 5 was the number of transistors in the millions that the new chips contained. Now if they keep the name Pentium and put a 5 behind it then they've just named it 5 twice
:

???? LOL

 

[Dan_D]

Intel AMD and Cyrix had a court battle over the 486 name. If you recall everyone had a 486. It was that time that Intel started their Intel Inside marketting campaign. But that wasn't enough. So chipzilla took AMD and Cyrix to court claiming copyright infringment on the 486 name.

The court ruled that numbers themselves can not be copyrighted. So Intel had to differentiate tier chips from Cyrix and AMD's. So they called their 5th generation 586 chip Pentium. Denoting that it was infact a 5th generation architecture. Which the architecture is reffered to as P5.

They've never thought about the name after that. It was so well recognized and almost everyone that's ever even looked at a PC knows what a Pentium is to a degree. But what's funny is the Pentium Pro is a completely different design than the Pentium and is known as the P6 microarchitecture. And is in fact thier 6th generation chip design. However it co-existed with the P5 due to the fact that at the time the P6 was too expensive to produce. Mostly due to cash design. Enter the Pentium II. Intel's push to cheapen the P6 and introduce it to the masses.

The rest is histroy. But what was funny about that is the Pentium II was slower than the Pentium Pro at the same speeds. Due to the cache memory being on a card running at half the speed of the processor rather than being on die like the Pentium Pro's was.

This is similar to the Northwood vs. Prescott scenario of today. Prescott is generally slower than the Northwood that was before it. Oh well, enough ranting.

One more thing though, the Netburst architecture is thier 7th generation core but does not retain the P7 style architectual designation of the P6 and P5's. However the Pentium M is actually a P6 based design with a few inhancements.

Good old P6's whipping that ass since 1995!!!

 

[nst6563]

lol....I knew it had something to do with "5"

but I still think it would be funny if they named it "Pentium 5"....thus naming it 5 twice

 

[V0ltage]

judging by this article: http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2117
it seems single-core dothan chips for LGA-775 will replace prescott, followed by dual-core dothans.
both will be much, MUCH cooler than prescott.
I cant wait to get one of those LGA dothans at then end of the year

 

[emorphien]

the point of dual core is to reduce heat. i dont feel like finding out HOW but its not two prescott cores on one chip

Yeah, seems counter intuitive to me too.

 

[topdragon147]

maybe it is two cores just shared instruction parts

 

[widefault]

1st, dual cores increase the core size, meaning less wattage per square mm to remove from the package.
2nd, Prescott's big issue is current leakage at higher speeds. Two half speed Prescott cores will use less power than one full speed core and should run significantly cooler. Only problem will be in how well they can get the chips to work together. A dual core 2x2GHz Prescott will likely be behind a full 4GHz single core in computing power.

Add those together and you can see why Intel will at least try a dual core Prescott.

 

[Duke3d87]

1st, dual cores increase the core size, meaning less wattage per square mm to remove from the package.
2nd, Prescott's big issue is current leakage at higher speeds. Two half speed Prescott cores will use less power than one full speed core and should run significantly cooler. Only problem will be in how well they can get the chips to work together. A dual core 2x2GHz Prescott will likely be behind a full 4GHz single core in computing power.

Add those together and you can see why Intel will at least try a dual core Prescott.

not necessarly. The die might be bgigger, but wouldnt the heat out/squre mm be similiar? I think the current actually leaks at any speed w/prescott. Transistor gate leakage has been a problem for years, but only when companies started transitioning to the 90 nm process did it become a big problem. its estimated that 20-40% of Prescotts power consumption is b/c of transistor gate leakage. Intel needs to sample their tri-gate architecture really really soon. I like the idea of 2x2 GHz Prescotts, but then the idea of the NetBurst architecture is gone. Intels logic was 20% less efficient 20% faster. If they shortened the pipeline to 14-16 stages (just enough for HT to work well) it would be really good as the prescott core itself isnt that bad.

 

[oldpablo]

I heard rumor that they are going to be calling it "Pentathlon 64". Something about being able to convince more enthusiasts to buy...

 

[Synful Serenity]

I have to say, I liked the model designations like the 80486 better than the "name" Pentium. Not only did Intel want something it could trademark, but being at the dawn of the mainstream Internet, Intel was leaning in that direction anyway so CPUs could be dumbed down for the masses like a lot of things in computing during this time (I can't remember how many times I've asked someone to look in a directory and they responded with something like "Ohhh, you mean the folder!" *groan*). It figured the average person would be more comfortable and have better brand recognition with a word rather than a bunch of numbers. It's going both a step forward and back with the new model numbers, for one thing it adds more confusing numbers like the old ones, but it also de-emphasizes GHz (I had thought was a good selling point since it runs higher than AMDs) which many marginally computer-literate customers apparently did not understand. I bet Intel intentionally wanted people to think of BMW with the numbering..."Hey does that Pentium 750iL come with 12 cylinders?"

Recognition is probably why they kept Pentium name even after the 5th generation...the big joke was always that the 686 would be called Sexium (or the voodooesque Hexium). We'd all be happily using Octiums right now if Intel didn't stick with Pentium!

Hard to believe it's been 11 years...I remember the first Pentium comps selling for $6000-$9000.

Just a quick note: the first Pentium processor, which ran at 66MHz, had 3.1 million transistors.

 

[RancidWAnnaRIot]

A processor that will do all your work without you asking.

 

[Josh_B]

"But what was funny about that is the Pentium II was slower than the Pentium Pro at the same speeds. Due to the cache memory being on a card running at half the speed of the processor rather than being on die like the Pentium Pro's was."

Not all of the time. Pentium Pro's 16-bit performance was poor, so Intel added more L1 cache (not CASH) to the PII core to overcome this weakness. (16k vs. 32k.)

 

[Josh_B]

Just a quick note: the first Pentium processor, which ran at 66MHz, had 3.1 million transistors.

I believe there was also a 60MHz Pentium

 

[Synful Serenity]

That's right, they were technically released around the same time, but it was really the 66MHz that was the first and the actual design specification, the 60MHz's were just recertified 66MHz examples that wouldn't run correctly at 66MHz but passed at 60.

 

[Rizen]

That's right, they were technically released around the same time, but it was really the 66MHz that was the first and the actual design specification, the 60MHz's were just recertified 66MHz examples that wouldn't run correctly at 66MHz but passed at 60.

I had Pentium 60. My first PC That fucker ran at 5v, passive cooling. Ouch!

 

[Duke3d87]

I had Pentium 60. My first PC That fucker ran at 5v, passive cooling. Ouch!

it wouldnt run that hot because it was based on a smaller core. If it was the 90 nm process and a 5 volt vcore, you would need some serious cooling

 

[FreiDOg]

I had Pentium 60. My first PC That fucker ran at 5v, passive cooling. Ouch!

yea, but was 3-4 times the surface area and 1/10th the power dissipatio of a high end prescott.

 

[Rizen]

it wouldnt run that hot because it was based on a smaller core. If it was the 90 nm process and a 5 volt vcore, you would need some serious cooling

Dude, it ran hot, trust me Not Prescott hot, but it got quite warm!

 

[jamestime88]

I heard rumor that they are going to be calling it "Pentathlon 64". Something about being able to convince more enthusiasts to buy...

are you serious?
you cant be serious.

 

[Nohpys]

http://www.intel.com/products/roadmap/index.htm?iid=ipp_desk_proc+roadmap&
http://www.intel.com/products/roadmap/desktop_feb_04.jpg?iid=ipp_roadmap+desk_corp_desk&
http://www.hexus.net/content/reviews/review.php?dXJsX3Jldmlld19JRD03NjM
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2118

The 4GHz Pentium 4, the 580, is scheduled for launch before Christmas this year on LGA775. Using current Prescott 90nm technology with a 1MB L2 cache, it's an 800MHz front side bus processor and won't need any new chipset (past Granstdale and Alderwood) for operation.

So we are also looking forward to Nehalem from H2 '05 of which I can't find any specs. Except this which made me chuckle somewhat:
http://www.theinquirer.net/?article=7481

I can see from here: http://www.septor.net/roadmap/intel/cpu.html that Prescott will go to 4Ghz though while Tejas was due to exceed that in H2 '05.

Interesting stats here though:

Code Name Socket Process L2 Max Freq FSB
Prescott LGA775 90nm 1MB >4.0GHz 800MHz
Prescott 2M LGA775 90nm 2MB >3.73GHz 1066MHz

 

[Duke3d87]

http://www.intel.com/products/roadmap/index.htm?iid=ipp_desk_proc+roadmap&
http://www.intel.com/products/roadmap/desktop_feb_04.jpg?iid=ipp_roadmap+desk_corp_desk&
http://www.hexus.net/content/reviews/review.php?dXJsX3Jldmlld19JRD03NjM
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2118

The 4GHz Pentium 4, the 580, is scheduled for launch before Christmas this year on LGA775. Using current Prescott 90nm technology with a 1MB L2 cache, it's an 800MHz front side bus processor and won't need any new chipset (past Granstdale and Alderwood) for operation.

So we are also looking forward to Nehalem from H2 '05 of which I can't find any specs. Except this which made me chuckle somewhat:
http://www.theinquirer.net/?article=7481

I can see from here: http://www.septor.net/roadmap/intel/cpu.html that Prescott will go to 4Ghz though while Tejas was due to exceed that in H2 '05.

Interesting stats here though:

Code Name Socket Process L2 Max Freq FSB
Prescott LGA775 90nm 1MB >4.0GHz 800MHz
Prescott 2M LGA775 90nm 2MB >3.73GHz 1066MHz

problem is that the article from the Inq was off. Intel cant get to 10.2 GHz anytime soon at least with the amount of power the prescott puts off. Tejas was officially canned too. Also with a3.6 GHz chip running so hot, what knows what a 10.2 GHz chip would be like...i dont think they could get that high even if they used tri-gate

 

[TheAcorn]

I don't think we'll see a dual-Prescott core, ever - Its too hot and not on the roadmap anyways. What kind of heatsink is going to remove that kind of heat? I'm sure it can be done, but lets all be glad that the Netburst is on the way out... (And I don't say that because I've got an AMD CPU in my sig.)

Intel (and AMD) have built their most recent cores from the get-go towards multicore CPUs. The Dothan and A64 are more-or-less drop-in solutions to whatever kind of CPU packaging and interconnects they'll be using to do this with from my understanding of it. How exaclt it works, well, who knows...

Sure is going to be neat, though, multiprocessing within one CPU. Now if we can have some 4-way boards with dual PCI Express x16 lanes...

Think 8 CPUS (And not in the hyperthreaded, virtual sense of things) with a pair of high-end video boards... Makes me all warm and fuzzy.