The Official Prescott discussion Thread

Originally posted by USMC2Hard4U
The Intel Pentium 5 :D :eek: :p

"Power.... Performance.... Perfection"

That should be in the P5 Commercial ;)
Click to expand...

Thats only 3 P's...what about 2 more? hehe
 
Originally posted by mjz_5
why not get an opteron, why faster!
Click to expand...

Get the hell outta this thread fanboy.....lol :D

I am also thinking the 3.2C is looking much nicer than a "pressy".
 
"Power.... Performance.... Perfection"
Originally posted by forcemac101
Thats only 3 P's...what about 2 more? hehe
Click to expand...
If it were SCO instead of Intel, the other 2 P's would be "Patents" and "Pay up"? :D
 
One *welcome* surprise concerning Prescott (for moi) is that ASUS was *right*...the 800-E Deluxe *will* swallow it with the 1014 BIOS (currently available), and that Prescott is available *today* at speeds as low as 2.8 GHz (decisions, decisions) and *1 MB* ATC (vs. 512K in Northwood-C) Prescott (Northwood-E) is indeed available now.

Second surprise: Northwood-E costs no more than Northwood-C. (Uh oh.) This makes the *motherboard* the deciding factor in which P4 you should get. (This is the first time in Intel history that something *other* than the CPU pinout or other *obvious* CPU feature changed the builder's decision tree.)

Third surprise: While Northwood-E does *not* perform well under ligh load (compared not just to Athlon64, but even Northwood-C), as the load goes up, Northwood-E starts taking command. (So far, every review has shown this.) Sounds like the E is the processor for desktops that run under heavy loads.

Now I know where *some* of my refund $ is going.
 
Fugger from xtremesystems.org has put up some OC'd numbers in the forum. 4ghz+ 300fsb on a 2.8E.

OC'd prescott benchies.

Tried to look for a close compare at about the same mhz in my 3dmark2001, this was the closest mhz score i could find. I am running my memory @ 1:1 while his is 5:4, same tight timings though.

His card is a 9800xt, mines 9800 pro @ 460 core/ 365 memory in this shot as you can see., and about 2 months old. So it seems the prescott isnt doing to bad since his vid card is at stock speeds..

http://mywebpages.comcast.net/tedinde/photogallery/XFpics/compare.JPG

Mines a 3.2c on the left, prescott on the right. The mods are kinda off in that forum so be easy on them!!!;)

compare.JPG
:D
 
Originally posted by FrgMstr
I do not know.
Click to expand...


I'm a bit confused bud, you had an Engineering sample right? Would the Retail perform a bit better than the ES?
 
The Retails are supposed to be a later stepping revision. And hopefully run cooler.

Fugger is using a retail 2.8E.

Mine arent shipping till tomrrow:(

I guess i'll have to answer the whole worlds questions myself on the retail if i can believe mine will ship! I Called them today, was told tomorrow.

I can see some resellerratings dropping like a brick if Spartantech/pcprogress doesnt get these out and they have been telling everyone they are there upon phone calls to them.
 
What's fugger been up to?
Last I spoke with him he wasn't quite the AMD basher he was in the past, although certainly not a fan of AMD, he actually could have an in depth conversation about the differences in architectures without namecalling...
 
well i told you guys this on like page 7, also hey ted, can i get one of those chips from you after you are done playing with them?? Shipping wont be expensive since you live 2 hours from me.
 
Originally posted by AthlonXP
well i told you guys this on like page 7, also hey ted, can i get one of those chips from you after you are done playing with them?? Shipping wont be expensive since you live 2 hours from me.
Click to expand...

You can have the 2nd best one!!:D Email me. gotta get them first though.

USPS priority you'll get it next day, you are close. Not as close as you were though, Did you move or do you commute now??
 
Originally posted by Tedinde
Fugger from xtremesystems.org has put up some OC'd numbers in the forum. 4ghz+ 300fsb on a 2.8E.

OC'd prescott benchies.

Tried to look for a close compare at about the same mhz in my 3dmark2001, this was the closest mhz score i could find. I am running my memory @ 1:1 while his is 5:4, same tight timings though.

His card is a 9800xt, mines 9800 pro @ 460 core/ 365 memory in this shot as you can see., and about 2 months old. So it seems the prescott isnt doing to bad since his vid card is at stock speeds..

http://mywebpages.comcast.net/tedinde/photogallery/XFpics/compare.JPG

Mines a 3.2c on the left, prescott on the right. The mods are kinda off in that forum so be easy on them!!!;)

compare.JPG
:D
Click to expand...

his fsb is way higher than yours.
 
I wanna see some sandra marks and some individual game marks at that prescott at 3.4, 3.6, 3.8, and 4.0Ghz
 
I stand corrected on the pipeline 'rumors'. /Eats crow.

How much cooler are the 3rd stepping production CPUs over the review samples? Another website seemed to show a 10c difference.

Also, I'm left wondering with all of the unaccounted transistors in Prescott (much more transistors over Northwood that aren't consumed by cache), PLUS the move to LGA 775... does Prescott have an on-die memory controller?

Edit for clarity: Meaning another disabled secret feature like HT. The Althon64 has the ability to use an off-die Northbridge controller. When Intel releases Grantsdale/Alderwood and LGA 775, it could mean a move to the on-die controller. That could explain the extra pins.

/end pointless speculation
 
Originally posted by pakotlar
his fsb is way higher than yours.
Click to expand...

Yeah your right on that, Hard to compare since he's @ stock.

Looked for something comparable fsb, but i dont have a 2.8, I have a 2.4 @ same fsb and almost same 3dmark score but with a OC'd Vid card while his is stock speed. But im 500mhz behind him on a 2.4c. Been looking through all my benchies for somekind of compare.

But no luck this one's 4 months old.

http://service.futuremark.com/compare?2k1=7511581
 
Originally posted by VVVVVVVVVVVVVVVI
I
Also, I'm left wondering with all of the unaccounted transistors in Prescott (much more transistors over Northwood that aren't consumed by cache), PLUS the move to LGA 775... does Prescott have an on-die memory controller?
Click to expand...
There's a lot of speculation but no real answers at the moment. One possibility is some sort of dual-core implementation, as the picture of the Prescott die clearly shows two integer execution engines, including 2 L1 data caches.
http://www.abxzone.com/forums/attachment.php?s=&postid=662865
 
Originally posted by SLee:

There's a lot of speculation but no real answers at the moment. One possibility is some sort of dual-core implementation, as the picture of the Prescott die clearly shows two integer execution engines, including 2 L1 data caches.
http://www.abxzone.com/forums/attachment.php?s=&postid=662865
Click to expand...
Another speculation is 64-bit registers. Too many rumours flying around lately.
 
Ted, post your info on your new Prescott's in the morning so I can look at them in between classes!!!!!!!!! ;) Try to OC those things ASAP, I'm still debating on buying one and I only have 2 days before I can't return my 865PE motherboard... :D
 
Originally posted by VVVVVVVVVVVVVVVI
I stand corrected on the pipeline 'rumors'. /Eats crow.

How much cooler are the 3rd stepping production CPUs over the review samples? Another website seemed to show a 10c difference.

Also, I'm left wondering with all of the unaccounted transistors in Prescott (much more transistors over Northwood that aren't consumed by cache), PLUS the move to LGA 775... does Prescott have an on-die memory controller?

Edit for clarity: Meaning another disabled secret feature like HT. The Althon64 has the ability to use an off-die Northbridge controller. When Intel releases Grantsdale/Alderwood and LGA 775, it could mean a move to the on-die controller. That could explain the extra pins.

/end pointless speculation
Click to expand...


[puts on dummy hat] intel is crafty, I bet they are throwing in those extra pins for no reason at all except to fool us into thinking they ARE being used for something extra. damn you intel!!!! (shakes fist)


In reality, I've heard some speculation that maybe they are including 64bit extensions with the 775 pin version? who knows.
 
when was hyperthreading announced?
wasn't it basically announced with the Northwood but not enabled?
 
I honestly hope Intel doesn't decide to flex it's dick and go with 64bit extensions. I'd rather get something that can benefit me NOW. By the time XP64 or Longhorn roll around I very much doubt any of us will still be on A64s/Northwoods/Prescotts. Save 64 for the Tejas and give Prescott something to make it truly competitive with the A64 NOW. At least that's my take on it.
 
Good point. I could really care less about a SSE3 64 bit instruction, that may or may not be implemented a year from now in select pieces of software. That is unless for sure it has a noticable benefit for all applications.

The specialized instructions are doomed to eventually die anyhow. Like MMX which is mostly integer based is pretty well going to be sitting idle with all the floating point DX9 games.

Specialized instructions have their place, but should never outweigh the core processing speed and ability.

I could only imagine how much better the Prescott might have been if they just kept the pipeline of the Northwood and tacked on the bigger caches.
 
I think the strained (stretched) silicon is letting current leak. The layers are so thin to start with, then they stretch them out.... to allow the electricity to flow faster. I think its like stretching a balloon. After you stretch it far enough gases leak through.

The should have just produced basically a Prescott core (all improvements EXCEPT longer pipeline) on the MATURE .13 process.

It would probably run faster & cooler, until they can get the .09 process straight. I know the die size would hurt them, temporariliy, but it cant be any worse than Williamettes were.

1Mb L2, improved HT, SSE3, etc.. various improvements on .13 would be the ticket at least temporarily.
 
Another long wait till the end of March.

I got my eye on that 939 socket A64 coming soon.

If not the good ole 3.4EE is about the best we can muster with current *STOCK* Intel Powah!

Psyko M.
 
Originally posted by chrisf6969:

The should have just produced basically a Prescott core (all improvements EXCEPT longer pipeline) on the MATURE .13 process.

It would probably run faster & cooler, until they can get the .09 process straight. I know the die size would hurt them, temporariliy, but it cant be any worse than Williamettes were.

1Mb L2, improved HT, SSE3, etc.. various improvements on .13 would be the ticket at least temporarily.
Click to expand...
The costs involved in manufacturing a P4-based processor with a 1MB L2 and all the other enhancements on a .13µ process would have been prohibitive for the market segment this chip is aimed at. Prescott OTOH, manufactured with the .09µ process is priced comparatively with Northwood. Moreover, there was very little additional speed left in the .13µ process to begin with. Beyond 3.4GHz and the process becomes unreliable. Northwood tops out at this speed for a reason. Lastly, the longer pipeline is necessary for ramping quickly and outcompeting AMD, as I mentioned in my earlier posts. It's the strategy Intel adopted and it has largely been successful over the past few years. Who can argue with success?
 
Originally posted by APOLLO
The costs involved in manufacturing a P4-based processor with a 1MB L2 and all the other enhancements on a .13µ process would have been prohibitive for the market segment this chip is aimed at. Prescott OTOH, manufactured with the .09µ process is priced comparatively with Northwood. Moreover, there was very little additional speed left in the .13µ process to begin with. Beyond 3.4GHz and the process becomes unreliable. Northwood tops out at this speed for a reason. Lastly, the longer pipeline is necessary for ramping quickly and outcompeting AMD, as I mentioned in my earlier posts. It's the strategy Intel adopted and it has largely been successful over the past few years. Who can argue with success?
Click to expand...

I said temporarily. in other words the next 2-3 months on 478 pin mobos, until they can get a new stepping probably, not just a new revision. B/c the heat/wattage is a major issue.

I know it would cost them more. But they have plenty of room in their profit margin to work with. Williamettes took up a huge amount of die space and were made on their 200m wafers.

They COULD do it and still make money no problem if performance was a higher priority.

But their # 1 priority is Mhz for marketing, #2 is increasing profit margin so they can have 1839108219383 TRILLION $, #3 stability of platform, #4 actual performance of platform.

(this is coming from an Intel fanboy - me )

I think we will see the real Prescott on 775 pins, when they enable the 2nd core (oops let the secret out)
 
Originally posted by chrisf6969
I think we will see the real Prescott on 775 pins, when they enable the 2nd core (oops let the secret out)
Click to expand...


the who-what?
 
Originally posted by 0ldman
when was hyperthreading announced?
wasn't it basically announced with the Northwood but not enabled?
Click to expand...

It was announced long before the northwood with the first willie but was not enabled like maybe 64bit is on the prescott.
 
Abit jsut released a BIOS flash for the IC7 board today for anyone that is interested. It allows support for the new Prescott chips. I haven't flashed to it because I don't see the Prescott as a worthy purchase at this moment. My [email protected] does everything I need....well at least until the new Socket Layout is released.:D
 
Originally posted by chrisf6969
I think the strained (stretched) silicon is letting current leak. The layers are so thin to start with, then they stretch them out.... to allow the electricity to flow faster. I think its like stretching a balloon. After you stretch it far enough gases leak through.

The should have just produced basically a Prescott core (all improvements EXCEPT longer pipeline) on the MATURE .13 process.

It would probably run faster & cooler, until they can get the .09 process straight. I know the die size would hurt them, temporariliy, but it cant be any worse than Williamettes were.

1Mb L2, improved HT, SSE3, etc.. various improvements on .13 would be the ticket at least temporarily.
Click to expand...

Let me explain to you why you don't work at Intel. You are not an electrical engineer. You do not know what you are talking about. In fact, Intel believes your thoughts on what they should and should not be doing are insignificant. They are right. Let me edumacate you:
http://www.eetimes.com/semi/news/OEG20020813S0012

"We have figured out a way of changing the silicon lattice structure to allow faster electron flow — a 1 percent change in silicon spacing to achieve a 10-to-20 percent increase in drive current. Intel is unique in that we can do this with no deterioration in terms of the short channel effect or junction leakage," the Intel fellow said.

http://www.xbitlabs.com/news/other/display/20030911140127.html


SOI and Strained Silicon appear to be the technologies that AMD and Intel pin a lot of hopes on during the next three to five years. Both technologies are intended to keep increasing the speed of current flowing through a microprocessor and to address the connected issues, such as power leakage. SOI adds a thin oxide layer to a silicon wafer in order to insulate the circuit against power leakage. Strained silicon deposites a layer of silicon germanium on top of a silicon wafer. This stretches the silicon atoms to let electrons flow faster through a circuit.
 
Originally posted by merlin704
Abit jsut released a BIOS flash for the IC7 board today for anyone that is interested.
Click to expand...
Code: 
1. Update Intel on screen branding (OSB) logo.
2. Update BBS code. This will fix some boot sequence issue with some SCSI
   card.
3. Fixed the problem that the system cannot boot into Windows if
   1) The OS is installed in the SATA hard disk when both the SATA and
      PATA hard disk are present.
   2) The PATA hard disk is removed after the OS installation.
4. Update CPU Micro Code.
5. BIOS compile date: 12/31/2003
i will be flashing it tonight - prescott or not.
 
Originally posted by pakotlar
Let me explain to you why you don't work at Intel. You are not an electrical engineer. You do not know what you are talking about. In fact, Intel believes your thoughts on what they should and should not be doing are insignificant. They are right. Let me edumacate you:
http://www.eetimes.com/semi/news/OEG20020813S0012

"We have figured out a way of changing the silicon lattice structure to allow faster electron flow — a 1 percent change in silicon spacing to achieve a 10-to-20 percent increase in drive current. Intel is unique in that we can do this with no deterioration in terms of the short channel effect or junction leakage," the Intel fellow said.

http://www.xbitlabs.com/news/other/display/20030911140127.html


SOI and Strained Silicon appear to be the technologies that AMD and Intel pin a lot of hopes on during the next three to five years. Both technologies are intended to keep increasing the speed of current flowing through a microprocessor and to address the connected issues, such as power leakage. SOI adds a thin oxide layer to a silicon wafer in order to insulate the circuit against power leakage. Strained silicon deposites a layer of silicon germanium on top of a silicon wafer. This stretches the silicon atoms to let electrons flow faster through a circuit.
Click to expand...

Intel needs SOI to stop the leakage. At .09 obviously they have problems with leakage, if they are operating at 1.3v and 103W at 60C+.
 
i doubt it's due to strained silicon. As AMD does not use it and they are having 90nm problems too.

There is a leak problem but it's not due to strained silicon. they are simply increasing the length of electron orbitals.
 
i doubt it's due to strained silicon. As AMD does not use it and they are having 90nm problems too.
Click to expand...
We don't have any 90nm AMD chips to compare to unfortunately. AMD has had 90nm problems in the past, but we have no idea whatsoever what their yields or power output numbers are going to be with 90nm A64 processors, or what they are currently.

There is a leak problem but it's not due to strained silicon. they are simply increasing the length of electron orbitals.
Click to expand...
Right, the leakage is probably less than it could be as a result of using the strained silicon. 90nm is so damn small that leakage is a huge issue.

Both SOI and sSi help to combat it by insulation and orbital deformation, respectively. The degree of efficacy of either technique is really not all that well established at this point though because I don't see too many independent analysts with good data discussing it.

Relying on Intel or IBM spokespeople to devulge the whole truth seems foolish. What does one expect them to say?

We will see how well Intels process matures at this process node. I'm quiet sure future stepping will cut down on power draw and heat output, but the degree to which this will occur and when these steppings will show up is currently unknown.
The EE Times article suggests Intel is combining sSi with a low-k dielectric, which is interesting.
 
Listening to these people you'd think Intel employees or microprocessor experts were in here.......NOT! Only in their own minds. :rolleyes:
 
Originally posted by batotman
Listening to these people you'd think Intel employees or microprocessor experts were in here.......NOT! Only in their own minds. :rolleyes:
Click to expand...


Pfffy, I work for Cyrix. That's why I know everything.:p :D
 
Originally posted by rayman2k2
I'm a bit confused bud, you had an Engineering sample right? Would the Retail perform a bit better than the ES?
Click to expand...


That is hard to say. But please look at it from our point of view. When you start making statements that will directly impact people's buying decsions, would you rather be commenting on a product you KNOW they can purchase themselves, or would you rather just HOPE your ES represented actual retail product? :eek:
 
Originally posted by pakotlar
cyrix is making a comeback! I swear! 10ghz cyrix processor!
Click to expand...


LOL Imagine the heat it would put off. Your case would be turned to lava.....just like the orginal Cyrix.
 
You must log in or register to reply here.
Back
Top