Barton PR Rating.

lhgpoobaa · Jul 1, 2002

Well here it is, hot from the steamy side of info leaks

As the tbred is identical in performance Mhz wise it was logical that it shared the same PR rating to the XP.
Barton, with 512k cache, has +600 to the PR rating.
so:
1667Mhz = XP or Tbred XP2000+ = barton XP2600+
and
2000Mhz = barton XP3100+

see it here
<A HREF="http://www.xbitlabs.com/news/story.html?id=1025349410" target="_new">http://www.xbitlabs.com/news/story.html?id=1025349410</A>

Proud member of THG's Den Of Thieves :lol:

chuck232 · Jul 1, 2002

Seems kinda messed up. I mean, a Barton will have a PR rating close to the Hammer? That could screw AMD over. Considering Barton will probably be a lot cheaper than Hammer, a lot of people could flock to Barton instead of Hammer. I think they'll probably add 200 to PR (doesn't PR satnd for "performance rating"? so wouldn't PR rating be kinda redundant?) maybe even 400, but 600, that's kinda over the top.

Falling down stairs saves time

lhgpoobaa · Jul 1, 2002

maybe. depends on the performance increase of the 512k cache. that reminds me i MUST do some benchmarks of my XP cpu with the L2 cache disabled. see if we can get some ballpark predictions of barton.
though i must say the PR system is growing on me a little. specially when your dealing with complete computer n00bs of the highest order who just look at numbers. they can "see" that a barton is better than a tbred cauz its got a bigger number.

Proud member of THG's Den Of Thieves :lol:

chuck232 · Jul 1, 2002

Don't you find that this PR thing is kinda getting blown out of porportion. Whereas Intel at least has to make their procs run faster, AMD can just slap on a number and expect everyone to believe in it? I think not. AMD's gonna have to show us that these CPUs actually warrant these high PR ratings.

Falling down stairs saves time

lhgpoobaa · Jul 1, 2002

of course. agree with you there.
but so far they havnt done too badly. infact i think many people agree that the rating system of the XP was under that which it should have been.
and they are supposed to use benchmarks and whatnot to determine the numbers.

but with everything, we will have to wait until the processor arrives to determine if the are PR'ing thru their asses or not

Proud member of THG's Den Of Thieves :lol:

chuck232 · Jul 1, 2002

I was just thinking, right now with the new 533MHz FSB for the P4, a 2.26GHz performs about the same as a XP2200+. Now the 2.26 is only $50 more expensive than a AXP2200+ and with future price cuts, maybe the lower end P4s could become basically the same price as Athlon chips. I'm only speculating, but just from looking at the recent pattern of price/performance, if this continued, AMD may not have much going for it till Hammer at least. Also Hammer's gonna be delayed til December and Intel seems to have accelerated Prescott to 1H03 from Q303. Maybe Intel's got one or two things up their sleeve.

Falling down stairs saves time

eden · Jul 1, 2002

There is still the low end AthlonXPs, which have and will continue maintaining an EXTREMLY good price/performance ratio. The XP1700+ in Canada is roughly 80$ less, and as a Tbred may be able to get to XP2400, thus the value of a 1.6A to 2.4GHZ may be beaten, even with a better aircooler for AXP, it still won't cost 80$ tax inc.!

As for this PR rating, who knows, we all can speculate cache isn't that much the K7's best area, and that adding won't do much, but we could be wrong. LHG did you try the disabling? It sounds interesting so give it a whirl! It might as well do a big difference in Office benching and Internet Content for sure.

--

Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

lhgpoobaa · Jul 1, 2002

havnt yet. will do tonite when i get home from work.
i remember doing it on my old cellery 500mzh years ago though!
remembering back all those years... disabling the puny celleron L2 cache killed around 25% to 33% of performance.
then disabling the L1 instead killed around 50%.
disabling both was a bad idea :lol:
it took half an hour just to get to the win98 splash screen. LoL then i gave up

Proud member of THG's Den Of Thieves :lol:

Guest · Jul 1, 2002

maybe. depends on the performance increase of the 512k cache. that reminds me i MUST do some benchmarks of my XP cpu with the L2 cache disabled. see if we can get some ballpark predictions of barton.

what would you prove disabling the L2 cache?!?
the PR is related to the AXP cpu power but i don't think AMD engineers have benchmarked their processor without the L2 cache activated.

moreover this formula (Rating = 3 x Frequency/2 – 500) seems to fit with the AXP cpu models but from where this other formula (Rating =3 x Frequency/2 + 100) have been calculated? i didn't follow their hypothesis for the Barton.

if you know you don't know, the way could be more easy ...

Copenhagen · Jul 2, 2002

I think you will find that only some tests will obtain a "true" increase of 600 PR ratings and others will show a very modest increase. This will make the PR rating even more misleading than it already is.

AMD desperately needs more PR ratings. They can't squeze enough true MHz out of the T-Bred/Barton, so they just slap on a sticker with a sufficiently high number to match Intel.

/Copenhagen - Clockspeed will make the difference... in the end

Kemche · Jul 2, 2002

Just Maybe AMD is adding some cool features in Barton that we don't know of. May be 166FSB and 512k cache would give enough preformance to justify the PR they are speculating. Or may be they are just desprate. What ever it is we will just have to wait and See.

KG

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

texas_techie · Jul 2, 2002

I saw those PR numbers for barton and almost crapped. My thoughts were the same as everyone elses: If its rated higher than Hammer, then why buy a Hammer?
The initial Hammer PR will be around 3000. I guess the only advantage of Hammer(although a big one) will be the chipset. The ODMC (on die mem controller) and HT will make it a better performer than Bartom IMO. The question is will ya shell out a LOT more cash for that performance.
Im skeptical too. I have my doubts Barton will be THAT much better. AMD better watch their step. If they start inflating the PR numbers, it will hurt them in the long run.

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

lhgpoobaa · Jul 2, 2002

awww crap!
i was going to do some testing last nite but i got home and my beast was Dead in the water!

after an evning of testing the best i conclude is that my bios chip is screwed

the lights go on, fans spin, drives spool up, agp card lites on, then after 5-10 seconds they all turn off.
no beeps, no nuffin. whats most disturbing is my mobo has a LED display for indicating post errors. and it doesnt show anything

*sniffles*
well im off to the mobo forum to post a detailed account.

Proud member of THG's Den Of Thieves :lol:

eden · Jul 2, 2002

Man, and we thought your torments were over since the UPS!

--

Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

Hoolio · Jul 2, 2002

Remember the opteron is a SERVER chip, they are gonna charge premium prices for the opteron.
It maybe the case that the CH is not gonna come out for a while, and by the time it does come out it will be much faster than the barton.

The athlon core has nearly reached it's threshold, so maybe the hammer will offer the higher clock speed. for example
Athlon 3000 could be a barton at it's limit therefore the Athlon 3100 would be a hammer? (The CH will be the next stage athlon - THG article can't remember the date).

(Post edited after looking at the amd processor road map!)
Edited by Hoolio on 07/02/02 07:12 PM.

texas_techie · Jul 2, 2002

actually Hoolio, the CH desktop is coming out first. The server chip is coming second.

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

lhgpoobaa · Jul 2, 2002

yeah... im cursed.
anyway i did a big post in the motherboard forum if u care to look. all the sorid details are there. buggered if i know why.

Proud member of THG's Den Of Thieves :lol:

SammyBoy · Jul 2, 2002

Here's an little addition to the whole discussion. Instead of just showing the PR numbers of Barton, these slides, which are said to be genuine by the Inquirer... (take it with a grain of salt, I guess)... show both the server/workstation and desktop roadmaps up through Q3'02. Looks like the actual FSB is going to still be 266MHz, and the Claw' will debut at 3400+ with 256KB cache. The Barton will reach 2800+ by Q1'03 with the Hammer reaching 4000+ by the same time.

Whatcha guys think about that?

-SammyBoy

Matisaro · Jul 2, 2002

I think you will find that only some tests will obtain a "true" increase of 600 PR ratings and others will show a very modest increase. This will make the PR rating even more misleading than it already is.

AMD desperately needs more PR ratings. They can't squeze enough true MHz out of the T-Bred/Barton, so they just slap on a sticker with a sufficiently high number to match Intel.

Amds pr ratings have always been accurate, I will give them the benifit of the doubt, but if the pr rating barton does not perform on average the same as the equavalent pr rating tbred, I will never buy another amd chip again.

That is my vow.(much like the radeon quake cheat, I dont [-peep-] around with cheaters).

The Cash Left In My Pocket,The BEST Benchmark

nja469 · Jul 2, 2002

The reason for the revised PR rating is not only cache, and a revised core but it appears they are going to use optimizations from the hammer core in the barton.
AMD obviously hopes that Barton will go head to head with Pentium 4s at comparable speeds but, it believes, with better PR rating specifications. WE'll have to wait and see, but with hammer optimizations the 600+ might actually turn out to be accurate.

eden · Jul 2, 2002

This could explain AMD's strategy to make Hammer a direct Prescott competitor. Maybe their plan is to make Intel release the Prescott sooner as Barton competes a lot, then Intel does it, then CH comes after it and AMD makes a strike comeback. Sounds pretty logical no?

But they will have a hard time being creative for Hammer's features, and that will cause a big modification of the silicon which can waste some R&D money a lot...

--

Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

Kemche · Jul 3, 2002

Amds pr ratings have always been accurate

True, But when the PR rating first was introduced there was no Northwood, no 533Mhz FSB, None of the P4 beat AMD's Processor in Preformance. Only thing Intel had was the Huge Clock Speed nothing else. Intel still has the Huge clock speed but they are not loosing in proformance race.

I will give them the benifit of the doubt

Me too.

but if the pr rating barton does not perform on average the same as the equavalent pr rating tbred, I will never buy another amd chip again.

I will hold you up on that. jk.

More you look at it more it seems like this are just false rumors. The only way these PR Numbers can be true is if...

1. AMD is enabling 166Mhz FSB.
Most likely will not happen since it will require support from 3rd party chipset vendors and what not. No one is interested in Athlon anymore. Everyone is focused on Hammer.

2. Adding Hammer Feature (SSE2) on Barton.
Hey this can happen!!!!

3. More core advancements then we know of. Better Branch prediction etc.
This can also happen!!!!

Again, We will just have to wait and see. If This becomes true then AMD will end up loosing lots of their hardcore fans.

KG

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

eden · Jul 3, 2002

But again if this comes true, it also depends why, and it better have these enhancements. But indeed Mat's determination on this seems very strong, though I understand him, since the ATI's Quack 3 thing lol...
We'll see until then.

--

Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

Matisaro · Jul 3, 2002

True, But when the PR rating first was introduced there was no Northwood, no 533Mhz FSB, None of the P4 beat AMD's Processor in Preformance. Only thing Intel had was the Huge Clock Speed nothing else. Intel still has the Huge clock speed but they are not loosing in proformance race.

Pr ratings are compared to tbirds, not p4s. I mean accurate in comparison between axps.(ie they scale linerarly).

The Cash Left In My Pocket,The BEST Benchmark

Matisaro · Jul 3, 2002

But indeed Mat's determination on this seems very strong, though I understand him, since the ATI's Quack 3 thing lol...

I am so not biased on the amd intel thing, the last year imo amd chips have universally been the better buy, northwood changed that, I waited for tbred and it didnt cut it, so now I advocate oc intel, non oc amd.

But I dont deal with cheaters, if amd inflates its pr and cant handle it, I will not tolerate it.

AMd biased, eh whatever you guys say.

The Cash Left In My Pocket,The BEST Benchmark

Barton PR Rating.

Illustrious

Distinguished

Illustrious

Distinguished

Illustrious

Distinguished

Champion

Illustrious

Guest

Guest

Distinguished

Distinguished

Distinguished

Illustrious

Champion

Distinguished

Distinguished

Illustrious

Distinguished

Splendid

Distinguished

Champion

Distinguished

Champion

Splendid

Splendid

Share this page