• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

How high do you think the Athlon XP will go?

soulm4tter

Senior member
I currently have an Tbird 1.4 with KT133A. The prices of DDR and KT266A motherboards has got me thinking about upgrading, but i don't want to get a new CPU yet. If i go to a KT266A setup, i'm wondering how fast the Athon XP will go a year from now or more.
 
if you want to wait for the die shrink to .13 you will see them push up to 2500mhz eventually. till then the palomino's are hitting a wall between 1600mhz to 1800mhz from what i am seeing


Jen
 


<< I still think we're going to see an official AthlonXP 2100+ (1733MHz) before Thoroughbred hits the streets. >>




i believe you are right about this ...........that i think will be it for AMD before the thoroughbred hits the streets



Jen
 
i believe you are right about this ...........that i think will be it for AMD before the thoroughbred hits the streets.

I hope so, I want a Thoroughbred processor because it'll run a lot cooler and overclock equal to or better than the NW. 😀
 
I actually think we will see a 2100+ and a 2200+ (1.8ghz) before thoroughbred...Many have talked at sites that 1.8ghz can be achieved with air cooling alone...Would be ice if they got these out by end of february to compete with 2.2ghz northwood before intel bust open gate with 2.4ghz and above...or a 533mhz fsb chip...
 
"I hope so, I want a Thoroughbred processor because it'll run a lot cooler and overclock equal to or better than the NW."

This might be a stretch of the truth. The Thoroughbred still has the 10 stage pipeline of the original Athlon, whereas the Northwood is running on a much longer one(20) . The end result will be that the Northwood should scale much faster and higher than the Thoroughbred. This is partially remedied by the fact that the FPU/ALU, and L1 performance of the Athlon core is much better than the P4.

I would still like to see the Thoroughbred incorporate 512k of L2 and a wider path to it. The current 64-bit path (9GB/s) to L2 is cripplingly small, especially when compared to the P4's 256-bit (48GB/s). A 166Mhz FSB would be nice as well. As Pabster has said, Palomino 2100+ will probably arrive, but I'm not so sure about the 2200+. The life expectancy of that model will be so short with the upcoming Thoroughbred as to be meaningless...unless AMD can't get Thoroughbred out the door in time. 😉


how much longer will AMD stay on the socket A path? will my 8k7a be able to take that Thoroughbred?

Until the end of 2002. The Clawhammer will change the AMD platform to another socket format. It might be a stretch for the 8k7a to work with Thoroughbred though. It's not a question of socket fit, but one of voltage. The .13um P4 runs with a Vcore around 1.4V at 2.2Ghz. The Thoroughbred should be quite a bit lower than today's 1.75V. If the Epox can run with a low enough Vcore or have a BIOS that can correct for this it should be fine.
 
i think only time will tell how far these palo's will go. myself i feel 2100 + will be the limit. as AMD had some thunderbirds that did go higher but didnt go beyond 1400mhz. i am hopeing that the thoroughbred will work with Epox 8K7A boards among others . Does anyone know if this is going to be true or are we looking at a new redesigned motherboard to support them?



Jen
 
I'm sincerely hoping that current KT266A boards will run Thoroughbred chips (and Bartons !?). Personally I use an Epox 8KHA+ and I'd love to know how much upgrade potential it has. Who wouldn't? AFAIK there's a number of factors that may affect the ability of current boards to use Thoroughbred chips...

1.Socket design. Last I heard AMD had confirmed that Socket-A was going to be used for Thoroughbred and Barton.

2.Voltage. The die shrink to .13um will mean a lower core voltage. Always has been the pattern in the past, so a fair assumption now. Limitations will be the voltage regulators on individual motherboards and also ability of BIOS to apply that correctly.

3.Signalling. Intel changed the signalling protocol for their Tualatin chips - the reason that Tualatins wouldn't run on then-current Socket-370 boards. Now I'm not saying that this will happen in the Palomino->Thoroughbred migration, but it could. Since AMD appear willing to keep the SoA form for so long, I'm hoping that they won't dick us around with a change like this. Besides, AMD have said (and no, I can't find a link to this quote, I'm sure I read it on the Inquirer, though) that it's a revision, not a new chip.

...so I'm cautiously optimistic that some KT266A's will support these cores. What d'ya reckon, guys? I'm no electronics engineer - are there flaws in these arguments?
 
thats sounds cool, hopefully it can hgandle it if not, ill be satisfied with 1.8ghz.. overclocked alil bit 😛
 
AMD's official statement is that current 266MHz FSB SocketA chipsets will support both the Barton and Thoroughbred cores, providing the motherboard will support the decreased core voltage.
Officialy they've stated they intend to scale the .18u Palomino core to 1.8GHz/2000+ by Q1 2002. The .13u Thoroughbred core is expected to take them to "at least" 2.07GHz/2600+ by Q3 2002. The Barton core is planned to emerge shortly thereafter and further scale the K7 to beyond 2.07GHz, though specific clockspeed have not yet been announced.
AMD is presently planning t continue production of the Barton core in the low end after the ClawHammer debutes in very late 2002/early 2003.

Officially there are planned to be no further changes to the K7 core beyond the expected die shrinkage, and SOI with Barton.
 
Whatever...BUt amd better have something like 2800+ or 3000+ by years end cause I have a feeling that is where intel will be by years end...

The fact intel boards run great minus the evil of rambus already makes them more attractive to me...If intel gets a 20-25 percent performance lead by end of next year and can get the price premium to just like 30 percent more then best amd I could see myself getting and intel system...

I think the recent review of northwood versus 2000+ was almost a dead heat...p4 northwood 2.0ghz is like 415 shipped (if not typo at pricewatch) and 200+ is 336.00 which is 23 percent premium now...However right now there isn't 23 percent performance boost...
 
ST4RCUTTER



<< The .13um P4 runs with a Vcore around 1.4V at 2.2Ghz. >>

the 2ghz northwood P4 runs at 1.5 volts, i doubt the higher speed chip would use less juice.
 
Rand...

Thanks - that's reassuring. 🙂 Where did you read the part about "AMD's official statement is that current 266MHz FSB SocketA chipsets will support both the Barton and Thoroughbred cores,..." ?
 
ST4RCUTTER wrote:

"I would still like to see the Thoroughbred incorporate 512k of L2 and a wider path to it. The current 64-bit path (9GB/s) to L2 is cripplingly small, especially when compared to the P4's 256-bit (48GB/s)."

I'm not so sure about that. If they do, I certainly hope they take the time to redesign the Athlon core to maximize the performance with 512K L2. Northwood's performance leads me to believe Intel simply added 256K. It doesn't do nearly as well in applications (such as SETI) where the additional L2 cache ought to significantly benefit.

The path to the L2 has been a topic of debate for years. I don't see it as being an essential modification.
 
Rand thank you for your input..........this is most reassureing to me and i'm sure everyone else


hugs


Jen
 
We'll see 2200+, then it's thoroughbred time. Just so you know, in spring Via is releasing the KT333, and SiS the 745. Both of these will support XMS2700.
 
The path to the L2 has been a topic of debate for years. I don't see it as being an essential modification.



Yes, but in the back of your mind you have to wonder what affect a 256-bit path to the L2 would do...would it be as effective as adding 256k of cache?
 
Back
Top