so whats up with Barton?

[texas_techie]

Heard a few rumors that Barton may have a 166 fsb.. That sure would be nice. In fact, I wouldnt be surprised AT ALL if it did debut with a 166 bus **wink-wink**

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

 

[bront]

That would definately make it a much more worthwhile buy. It would also help with the PR Rating.

If it is true, with the extra Cache, the 600 extra PR points could be somewhat legit.

Sucks that the speed will scale by 83.3 Mhz though. Ugly math.

English is phun.

 

[mbetea]

techie-
will the bartonMP be the same fsb? most mp owners now will probably look to barton for an upgrade and see how hammer turns out. there are no mp/mpx boards that will go to 166, unless there's another revision in the works? but it seems more like amd is just treading water until hammer debuts, it doesn't look like amd would want to go and do a revision on the mpx chipset, could be wrong though.

THG; the Jerry Springer of hardware review sites.

 

[texas_techie]

dont know about MP setups...

On another note: wouldnt it be cool if the t-bird scaled past 2.1 gigs (soon). again wouldnt surprise me at all **double wink**

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

 

[IIB]

So they've solved what ever was preventing its Mhz ramping?

This post is best viewed with common sense enabled

 

[lhgpoobaa]

at this point i dont think even AMD knows what barton will be released as... and all those rumors of hammer-esque optimisations sounds like spin.

definately a wait and see once again.


Ego: Check
Rose hued glasses: Check
Fanboy Button: Check
CPU forum, here i come!

 

[texas_techie]

iib,
the problem really wasnt preventing ramping, it was preventing good overclocks. And I doubt they fixed that. Im hoping they fixed it in barton. We shall see.

dunno about hammer-esque optmizations. But ill lay big money on the bus

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

 

[IIB]

I dont follow you... Overclock usally indecates the abilty of the chip to be ramped farther... you said you have some info on the specifics of the problem and to PM you if anyone wants it... well Im intrested for one.


This post is best viewed with common sense enabled

 

[Kemche]

Hi Techie,

I guess your **wink-wink** means that most likely they will put support for 166fsb in barton. But if you look at the MB makers there is no interest in the Old Athlon core. I don't know any chipset vendors releasing chipset which can support Barton @ 166fsb. So Why does it matter if barton support 166fsb or not since there will be nothing supporting that FSB in the market.

KG

"Artificial intelligence is no match for natural stupidity." - Sarah Chambers

 

[texas_techie]

IIRC any board that supports pc333 will benefit from a 166 fsb. Lets do the math. If the chip runs at a 166 bus DDR. Thats 166 X 2 = 333.
Chipsets like the KT333a support memory speeds of *gasp* pc333. So athlon's current fsb of 266 is wasted on a 333 board, since your still 67mhz shy of fully utilizing the memory speed. I have seen benchmarks of KT333 mobos overclocked to a 166mhz bus. It made a big difference. So if the chip natively supports a 166 bus, then you dont gotta OC to get the benefit of faster memory.

(or something like that). But yeah, there are already boards that can benefit from a chip supporting a 166 bus

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

 

[baldurga]

Are you saying AMD has implemented some protection to limit the overclocking potential? A way that, even if you unlock the CPU, it doesn't matter cause whenever it reaches more than 10% (p.e.) the CPU won't work?

Well, I don't know if it's posible or easy to do, but it good be a really great way of control. For our point of view, that would be bad news, but for AMD point of view, it prevents from selling 1700+ that are used as 2500+

I think this happens when really your processors can be better but because of marketing, strategic management, etc. you lower the CPU. Is what happens with P4 1.6A, 1.8A and 2.0A. The limit is not the CPU, it has been castrated, poor CPU :_(

DIY: read, buy, test, learn, reward yourself!

 

[ritesh_laud]

So if the chip natively supports a 166 bus, then you dont gotta OC to get the benefit of faster memory.

The motherboard must also support the 166 MHz FSB. On all current Athlon mobos, that means overclocking the FSB and possibly putting other components out of spec. So the mobo manufacturers would have to release new revisions of their products to "officially" support the new FSB.

Ritesh

 

[nja469]

According to a recent interview with an AMD spokesperson, the Barton WILL NOT have a FSB increase. This increase is being reserved for the upcoming Althon "clawhammer" and Opteron "hammer". The Barton will have some hammer optimizations, mostly SSE2 instructions, and the Barton will have a choc full of cache, 128K L1 + 512K L2.

Read about it yourself here http://www.theinquirer.net/?article=4403

"Opinions are like arseholes, everyone has one"

 

[ritesh_laud]

According to a recent interview with an AMD spokesperson, the Barton WILL NOT have a FSB increase.

Interesting, this directly contradicts Texas_Techie's "wink-wink."

This [FSB] increase is being reserved for the upcoming Althon "clawhammer" and Opteron "hammer". The Barton will have some hammer optimizations, mostly SSE2 instructions, and the Barton will have a choc full of cache, 128K L1 + 512K L2. Read about it yourself here http://www.theinquirer.net/?article=4403

I just read it and except for the cache, neither of your two statements above is supported by the <A HREF="http://www.theinquirer.net/?article=4403" target="_new">article</A>. Your first statement doesn't even make sense because the Clawhammer and Opteron don't have an FSB (the memory controller is on-die). And please provide a link showing evidence that the Barton will have SSE2 support. I've seen the rumors about Hammer optimizations on the Barton, so I won't call you on that one.

Ritesh

 

[jclw]

I don't think we'll see the Barton for a while.

<A HREF="http://www.overclockers.com/tips060/" target="_new">http://www.overclockers.com/tips060/</A>

- JW

 

[LED]

You're right there's something unbalanced and partial about this. It's called reality.

Even he is bracing for the inevitable flame mail from AMD Fanboys. The world is against you, AMD Chiperoos!!! LOL

This sig runs too hot.

 

[nja469]

Ok re-read the article, it states Barton WILL NOT get a FSB boost. Read SLOOOWWWLY the third paragragh.

As for Hammer optimizations in the Barton, read about here in this article. http://www.theinquirer.net/?article=4215
I do not recall where I read about SSE2, even then it was only hear-say. So yes there will be hammer optimizations in Barton, but I do not know what they are, however it was suggested somewhere that that includes SSE2.

"Opinions are like arseholes, everyone has one"

 

[ritesh_laud]

Ok re-read the article, it states Barton WILL NOT get a FSB boost. Read SLOOOWWWLY the third paragragh.

I wasn't talking about Barton's FSB, I was referring to the two statements you made that I quoted:

1. This [FSB] increase is being reserved for the upcoming Althon "clawhammer" and Opteron "hammer".

Makes no sense, Hammer doesn't have an FSB.

2. The Barton will have some hammer optimizations, mostly SSE2 instructions, and the Barton will have a choc full of cache, 128K L1 + 512K L2.

You state this like it's fact. Yet only the 512K cache is supported by the interview with Jan Gutter. The other two are purely speculation and The Inquirer is not a credible source (except when they link to another magazine like they did in this case on the Barton's cache and FSB).

Ritesh


<P ID="edit"><FONT SIZE=-1><EM>Edited by ritesh_laud on 07/19/02 02:50 PM.</EM></FONT></P>

 

[nja469]

Okay.... I didn't write these articles. A person can chose to believe them or not. The source has on occassion given misinformation, which usually the following day or so later has a follow-up with confirmed/denied info. A majority of the time they have already confirmed or are in the process of confirming something they write. This particular article dealing with Bartons cache size and FSB was an interview with an AMD engineer. No one here conducted the interview, and as far as I know no one posting here works for AMD. No one knows FOR SURE.

This entire thread is hear-say and I personally believe the interview with the AMD engineer. It's your choice to believe what you want. Sh!t I can sit here and say "I heard Barton is getting 166FSB (DDR 333) and hammer will get 200FSB (DDR 400) and 1MB of cache and will run at 3GHz :wink wink::"

I produced article(s) that contradicted what is wanted to be heard about a FSB increase and of course the source (an AMD engineer) is wrong. Hey though, someone else winked at a different number so that number must be right.

"Opinions are like arseholes, everyone has one"<P ID="edit"><FONT SIZE=-1><EM>Edited by nja469 on 07/19/02 04:55 PM.</EM></FONT></P>

 

[texas_techie]

hhehe..LOL

I understand your trepidation at not believing a "wink-wink". In this particular case, I would believe me over the article. here is why.
Hammer really doesnt have an FSB, so IF (ahem..cough)AMD was validating a 166 FSB.. which chip do you think its for?

Baldurga... where the heck did you get that idea? you CANT unlock a hammer. Its not like you unlock it and nothing happens. There is simply nothing to unlock. There is nothing to pencil, nothing to use a defogger kit on, nothing to buy a kit for. Will someone explain how Intel's stuff works to him, then maybe he will see what I mean.

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

 

[imgod2u]

First and foremost, there has been absolutely no movement in the motherboard community towards validating 166MHz FSB on the processor. It would make little sense as that would require AMD to push the motherboard makers into allocating enough resources towards their K7 line when they've hyped up their K8 line so much already. And as far as performance increases goes, <A HREF="http://www.anandtech.com/showdoc.html?i=1595&p=12" target="_new">Athlon 1.73 with a 166MHz FSB and 166MHz DDR memory at Anand</A> seems to indicate that an increase in FSB wouldn't help all that much anyway. Now while this may only be for a 1.73 Athlon and a 2.0 GHz (if Barton ever makes it there) may indeed need more direct memory bandwidth, I would doubt it would make too dramatic an improvement to tie up the corporate people at AMD to try to get the motherboard manufacturers and chipset manufacturers to go with it.

 

[texas_techie]

weird, ive read 3 articles that show a fairly nice performance improvement with a 166 fsb. Ill dig up the links if your interested.
I disagree with you about mobo makers though. I think once T-bred and XP become "value" processors (after hammer comes out) they (mobo guys) would like a means to increase the performance for that sector of the market.
Lets face it. Once prescott is out, northwood is still gonna be going head-to-head with t-bred and Barton, so AMD needs all the help it can get to compete. I would think moving to a 166 fsb wouldnt be insanely difficult to implement.
Oh well, guess we'll see

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

 

[nja469]

We'll just have to wait and see! )

"Opinions are like arseholes, everyone has one"

 

[baldurga]

I think we cross-talking about CPU. Probably my mistake, I reply to the incorrect post (working and being on this forum sometimes produces strange results )

"the problem really wasnt preventing ramping, it was preventing good overclocks. And I doubt they fixed that. Im hoping they fixed it in barton. We shall see."

I though here you were talking about the not too good overclocking results with Tbred, and that they intentionally prevent us from getting too far. My question was in that direction, if AMD could implement something that, whatever method you try for overclock (unlocking o raising FSB), the CPU wouldn't work when you exceed a "factory fixed limit frecuency". I don't know if I make my point clear.

Also, as far as I know, Tbred (and I pray Barton also) can be unlocked, and Hammer can't.

DIY: read, buy, test, learn, reward yourself!

 

[tbirdXPplus]

hey mine is 83.3 right now!

83.3 * 6 = 500(approx.)


Be nice.