Next .13 Athlon rev to support 166FSB?

[Adul]

Rumor mill time


Well guys, this is some big news I'd like to share with everyone. After some phone calls to and from various sources inside the AMD underbelly.

I have news of the next Athlon XP processors to be released as the 2400+ and and 2600+ to combat the on going competition from Intel. While simple speed increases are not big news for a processor, there is more to this than meets the eye. Even before the introduction of the Athlon XP 2200+ processor, AMD knew they were having a small problem with the .13 micron cores and thus a second revision of the Thoroughbred core was put underway. This new revision, while not officially announced (and it may not be) is under going some testing with great results. The 2400+ processor is set to run at 2.0 GHz and the 2600+ processor is looking to be at 2.13 Ghz. As Intel has plans of releasing a 3.0 Ghz processor towards the end of this year, this kind of speed boost from AMD is desperately needed -- perhaps even more.

As if things couldn't get better, along with this news comes an update on the fabled 333 MHz FSB Athlons. I was told that the AMD labs are really looking into this and testing it out with the newer "rev 2" cores of the Thoroughbred. While AMD could take the current 266 MHz FSB up to the 2.13 GHz range, moving beyond that would be very tricky and I have it on good authority that AMD has a lot more life in the Thoroughbred than many are expecting. What does that mean? I think we are going to see the 333 MHz FSB processors and fairly soon. The inside tip also explained that a decision by the AMD authorities on whether or not to use the 333 MHz FSB will be finalized within the next three weeks.

While nothing was mentioned of the Barton core, we have two options as to its whereabouts. Either this "rev 2" is the Barton and the info about the extra cache was left out, or Barton will follow this "rev 2" Thoroughbred.

In either case, this is the kind of news that we need to know that AMD is working hard on keeping the Pentium 4 within its grasp of control. If more comes to me, you can be sure it will be posted here, on Amdmb.com or the AMD Forums!


Forgot to site the sounce

Source AMDMB

 

[Lezboy]

While AMD could take the current 266 MHz FSB up to the 2.13 GHz range, moving beyond that would be very tricky

Doesn't sound very credible to me. It doesn't make any sense. The FSB wouldn't restrict how high they could raise the multiplier.

 

[MadRat]

The XP multipliers actually go up to 20.5X, meaning the limit is 2727 MHz or a PR of 3600+.

 

[ToBeMe]

Don't know.............it's be nice to see, but I guess we'll have to wait.............what is true and announced by AMD is that they are cutting their top of the line XP's by 30% now, and again shortly after Intel releases the P4 2.8, and once more before the holiday buying season to combat the gap in speeds and the upcoming 3.06 Ghz P4 which is set to come by the X-Mas buying rush..........that announcement made this past weekend didn't sound like they had anything planned, but, these days things tend to change very quickly............

 

[XMan]

Why would they? Everything I've heard indicates that the jump from 133-->166 offers little to no performance improvement at the same clock speeds.

 

[kuk]

Originally posted by: Xerox Man
Why would they? Everything I've heard indicates that the jump from 133-->166 offers little to no performance improvement at the same clock speeds.

I dunno ... according to x-bit labs, improvements are around the 7 to 15 percent range ... somewhat what is needed for the Athlon XP to reach the P4.
What I think should be done is convert all Athlon processors to 166FSB after the Hammer is launched. The K7 line should be restructured, as the Athlon would eventually ocupy the low and mid-end market (bumping the Duron off from production). This could bring life to Socket A on 166FSB ... an avoiding competition with the Hammer family.

But that's just my opinion. I could be wrong.

Kuk

 

[MadRat]

166FSB would mean there are no upgrades for older motherboards, not that there are alot of top of the line upgrade choices, as the XP2200+ is not compatible with many older boards...

 

[Dulanic]

Basically any KT333 board will run 166, heck I do it now... even has the 1/5 divider... all VIA would have to do is whip out a KT333A which "offically" supports 166FSB... I think AMD would do it, as would VIA. KT333 already does 166, just not offically.

 

[Duvie]

Didn't anybody notice that a 2400+ would be 2.0ghz???

If a 2100+ = 1.733ghz and a 2200+ was 1.8ghz...It seems to me that a 2.0ghz would be a 2500+ according to amd's whacky pr rating....

Is this an adjustment to the crazy theory that every 66mhz of change resulted in 100mhz in performance??


Amd needs to just get to 333mhz fsb now and quit f'ing around

 

[cmdrdredd]

not like it matters Intel's 533Mhz FSb is still way faster than anything AMD could do unless they jump to a 533.

 

[wfbberzerker]

Originally posted by: cmdrdredd
not like it matters Intel's 533Mhz FSb is still way faster than anything AMD could do unless they jump to a 533.

why would amd need to jump to 533mhz? if what adul says is true, the next revision of the athlon xp will probably make it just as fast as the p4.

 

[Duvie]

That 7-15% will put it up to speed with the 2.53ghz but the 2.8 and 2.66 will be out in 4 weeks....I don't think we will see anything in the market before then...I see no timetable listed...

Amd xp with 333mhz fsb and 512kb cache could be quite the chip to tide ppl over until we see if hammer lives up to the hype!!!

 

[kuk]

Originally posted by: Duvie
That 7-15% will put it up to speed with the 2.53ghz but the 2.8 and 2.66 will be out in 4 weeks....I don't think we will see anything in the market before then...I see no timetable listed...

Amd xp with 333mhz fsb and 512kb cache could be quite the chip to tide ppl over until we see if hammer lives up to the hype!!!

The 7-15% would put the Athlon XP 2000+ up to speed with a P4 2.4ghz. A 2100+/2200+ could rival the P4 2.53.
If Adul's news post is true, then a 2400+ and a 2600+ (with 166FSB) could smack a P4 2.6 and 2.8 around a bit with quite a large trout.

Kuk

 

[Duvie]

Actually you may be right depending on with apps...in quite a few apps the 2200+ would need that to equal the 2.4 (533fsb) processor now...So lets not go there!!! I got all the last reviews printed out And I think you are stretching a bit....but just a little

I actually was just talking about the core revision adul mentions which seems more likley to already be still at 133fsb....I agree with 166fsb that would definitely help...I was commenting on just the 2400+ chip but I guess I took the 7-15 percent comment from a guy talking about the fsb boost.....the cpu speed would scale 11 percent, but we know that doesn't always scale like that in majority of apps...

I hope they do it...I never forget the amd's performance has forced intel to strive harder and keep the prices low...I like!!!!


As long as we are dreaming....Let pray for amd to make thier own chipsets and give via the boot....

 

[socketman]

I also have it on VERY good authority that AMD is considering a 166fsb.. So count this as a second vote for that happening. Indeed it would help AMD keep up a little better until Hammer arrives. Every little bit helps.

 

[Adul]

Originally posted by: socketman
I also have it on VERY good authority that AMD is considering a 166fsb.. So count this as a second vote for that happening. Indeed it would help AMD keep up a little better until Hammer arrives. Every little bit helps.

i wonder if my mb will run a barton... hmm

 

[Soulkeeper]

if anything the 166fsb will be a marketing gymic for AMD
when an uneducated buyer sees the 333 selling term then they will know it is something new from 266 even if they have no understanding of what it means
wonder how this will affect their model number system
i'm not against the move to the 166 fsb

 

[Soccerman]

it certainly would be a good move on AMD's part to do this, their current chip CAN (and obviously does) benefit from it. if they are having problems scaling their CPU in MHZ, this is a simple way to get around that while maintaining at least SOME competition with Intel.

btw, I'm assuming this change applies to the Thoroughbred core, not just the Barton?

 

[Swanny]

It would be nice.....

But I still think I'll wait for Hammer

 

[JellyBaby]

If Adul's news post is true, then a 2400+ and a 2600+ (with 166FSB) could smack a P4 2.6 and 2.8 around a bit with quite a large trout.

More like tickle the P4 2.6/2.8 with a minnow.

 

[Soulkeeper]

just wait for the reviews then we'll all laugh
heh

 

[ToBeMe]

OK, now there are a few sites saying that AMD is releasing a 2400+ and a 2600+ sometime this fall...........but..............NOT with the 333FSB. Seems the 333FSB will wait until Barton and that will now be around the first of the year.............so where does that put Hammer????

 

[Holmecollie]

Originally posted by: cmdrdredd
not like it matters Intel's 533Mhz FSb is still way faster than anything AMD could do unless they jump to a 533.

To my understanding AMD actually performs better to a certain clock. ie 1.6Ghz (2x133mhz fsb) performs better than a 2Ghz(4x133fsb).
Now that doesn't make intels 533mhz fsb supperior to AMD's 266mhz fsb to me, correct me if I'm wrong.

 

[Lezboy]

To my understanding AMD actually performs better to a certain clock. ie 1.6Ghz (2x133mhz fsb) performs better than a 2Ghz(4x133fsb).
Now that doesn't make intels 533mhz fsb supperior to AMD's 266mhz fsb to me, correct me if I'm wrong.

You're wrong. The processor's internal architecture allows it to keep par with P4 at higher frequency but it has nothing to do with the bus speed. A 266mhz bu can only move 2.1gigs per second between the processor and the memory where a 533mhz can double that. This is important when using a faster memory. It's obvious that DDR333 is currently bottlenecked on the athlon since the processor has no road to get that much data.

 

[Soulkeeper]

Nobody is wrong
but i would have to say that by virtue of the P4's design it requires more memory bandwidth than an Athlon system
to say that ddr333 is holding the athlon platform back would be a mistatement
there is no athlon chipset/cpu combo that currently takes official support of a 166 (333ddr) fsb
when overclocking to use a 166fsb there are tangible benefits but there becomes a point where you don't need anymore
if you were to run an athlon off a 533fsb you would barely use half of the available bandwidth at any one time
i suggest some reading on the P4's 20 stage pipeline and some other specifics on it's design