Something funny about T-bred

[BDSM]

We've seen the athlon xp 2100+ reviews. Most ppl seem to be able to get it up to about 1900 mhz. So the palomino core is capable of 2300+ speeds already! It's weird that the T-bred will be launched at 1800+->2100+ speeds only and roadmaps indicate that it won't scale that high either.

I wonder if AMD's .13 micron technology is very immature.. or if AMD is just very confident and doesn't feel any need to release higher speed grades.

Anyhow. an 1800+ overclocked to maybe 2 GHZ could be a very sweet processor for that money, and it would woop anything Intel could offer, save for supercooled set ups.

 

[Duvie]

considering from most amd users saying that 1800mhz on any chips from 1800-2000xp is tough with conventional air cooling alone...How do you think ppl are going to get to 2000mhz or 2500+ xp from a 1800+??? at 2500+ it would be a sweet system, but aircooling ...doubt it...watercooled maybe, but then figure some have gotten 2.2ghz to 3ghz p4's with water cooled and I don't think you will beat that...


Where are you seeing so much data to base the assumption that most 2100+ xp's will reach 1900mhz or about 2350+??? At .18micron?? Doubt it...Air cooling monsters maybe, but what amount of heat and voltage necessary???

Provide some links, please....

 

[BDSM]

Duvie..

Viahardware got theirs to 1872 Mhz
http://www.viahardware.com/xp2100_1.shtm

Tech-Report got theirs to 1885
http://www.tech-report.com/reviews/2002q1/athlonxp-2100/index.x?pg=7

Pc-Stats got theirs to 1890 Mhz
http://www.pcstats.com/articleview.cfm?articleID=998

Amdzone got theirs to 1924 Mhz
http://www.amdzone.com/articleview.cfm?articleid=999&page=6

Is that enough for you?
Most of these overclocks were made without unlocking he cpu.

The latest 0.25 micron based Athlon core released was the 700 Mhz. It had external L2 cache.
With the move to 0.18 micron AMD added the L2 cache and has since also added a few new instructions adding a few extra square mm's to the die size. The fastest available Athlon now is the 1733 Mhz 2100+.

This is a 247% speed jump! So the T-bred (0.13 micron) should be capable of alot more than it's launch speeds of 1800+ to 2100+!
And I do mean ALOT more! That's why I think it seems so weird that they are starting out at such low frequencies.

The p3 got a 60% speed jump moving to 0.13 micron. (1.6 ghz seems to be it's limit)
The p4 seems to get about as much, and it has also gotten twice as much L2 cache.

So.. If the Athlon shows as much gain as that it would be capable of about 2700 Mhz (1,5*1800) at LEAST.
This would yield a power rating of about 3100+. Intel may be able to compete with this but it's gonna be a tough race.

Though my question remains: Why the low launch speeds?

Do they have problems with the 0.13 micron process or do they simply feel confident with their current product line and therefore feels it is unnecessary to go higher, for now?

With this background maybe you don't think it would be weird to see an 1800+ T-bred go to 2 Gigs.

 

[alen1234]

Probably to make money. As they are taking market share away from Intel and building a better revenue base they are becoming more like Intel was some years back. And don't forget that AMD and Intel are conducting joint research projects at some of our National Labs so it's not all cut throat competition.

It's not personal. It's only business.

 

[Wuzup101]

Yes it's true that they might be able to scale that high. However you have to remember that the first chips usually aren't as good at overclocking as the ones that are produced after the core has been on the market for several months. (IE: the first 1ghz AMD chips didn't overclock like the AXIA stepped chips)

 

[Mikewarrior2]

I think part of it is due to a not-so effective .13 Mu process....

Translated ZDnet Germany Page with Thoroughbred heat output

seeing as how Tbred looks to be only 10% or so cooler than an XP chip(and that can be explained by the CPU Vcore drop), despite the drastically reduced contact surface....

Heat per sq MM is going to be 1.5X that of an XP chip, so cooling these buggers is going to be hard....



Mike

 

[BDSM]

Mikewarrior.. Yes, but the t-bred will have heat spreaders, right?

Although that doesn't change the fact that the heat/square mm is almost 1,5 times the pally's.

 

[Mikewarrior2]

Nope, no heat spreaders....

 

[SexyK]

I don't think anyone should fool themselves into thinking this is some kind of "strategy". If AMD had the capability right now, they would have .13u cores on the street, end of story. There would be absolutely zero negative side to having tbred out there right now, and a lot of older roadmaps show it having been released already. As Mikewarrior mentioned, AMD is reportedly having a very difficult time getting good yields on their .13u cores, so i still wouldn't be suprised if you can't get a Tbred in your hands until early H2 2002. Only then will we be seeing the overclockabillity i suppose, and what Intel has out. I hope they can get it out soon, because if not, based on the performance of the P4 on 533MHz OC'd busses, they are in trouble.

Kramer

 

[Lithium381]

Would air-cooling include those dudes that put their systems in refrigerators? I would think that would work? Then with a delta fan (ooo, already sound insulated from the 'fridge) it could do a decent job? or are these things going to be completely unweildy?