PDA

View Full Version : More PR Rating Info To Discuss!


OC-Master
09-08-02, 12:14 PM
OK, so far we know the following,

AthlonXP Palomino's have a PR rating of 1500+ 1600+ 1700+ 1800+ 1900+ 2000+ 2100+ and 2200+ and corrispond to 1.33GHz, 1.40GHz, 1.47GHz, 1.53GHz, 1.60GHz, 1.67GHz, 1.73GHz and 1.80GHz. First generation T-breds use the same PR rating as Palominos do.

The AthlonXP T-bred B's have a PR rating shrink of 100+ which means using the Palomino PR, a 2GHz AthlonXP would be called 2500+ but now using the B revision, AMD has decided to go with 2400+ which by my records is a very smart choice.

Now, with the introduction of a 333FSB, AMD has decided to bring back the 100+ rating that it took away which means that a 2GHz CPU using a 333FSB has a PR rating of 2300+.

Barton is next and the most interesting of the bunch. AMD will add yet 100+ additional to the PR rating to make up for the additional L2 cache which was increased in Barton. All though a 2.27GHz 266FSB AthlonXP T-bred-b has a PR rating of 2800+, that same 2800+ rating on Barton would be called a 3000+, hence you add additional 100+ for the 333FSB and 100+ for the 512KB of L2 cache that Barton posses.

XP2300+ = 1.83GHz T-bred B 333FSB 11 X 166.66
XP2400+ = 2.00GHz T-bred B 266FSB 15 X 133.33
XP2500+ = 2.00GHz T-bred B 333FSB 12 X 166.66
XP2600+ = 2.13GHz T-bred B 266FSB 16 X 133.33
XP2700+ = 2.17GHz T-bred B 333FSB 13 X 166.66
XP2800+ = 2.27GHz T-bred B 266FSB 17 X 133.33
XP2800+ = 2.17GHz Barton 333FSB 13 X 166.66
XP2900+ = 2.33GHz T-bred B 333FSB 14 X 166.66
XP3000+ = 2.33GHz Barton 333FSB 14 X 166.66

I hope this helps to clear up that PR stuff!


DS-Master

OC-Master
09-08-02, 12:39 PM
Did you know that recently, AMD decided to remove T-bred A from there facilities which means that after this year, only the;

XP2300+ = 1.83GHz T-bred B 333FSB 11 X 166.66
XP2400+ = 2.00GHz T-bred B 266FSB 15 X 133.33
XP2500+ = 2.00GHz T-bred B 333FSB 12 X 166.66
XP2600+ = 2.13GHz T-bred B 266FSB 16 X 133.33
XP2700+ = 2.17GHz T-bred B 333FSB 13 X 166.66
XP2800+ = 2.27GHz T-bred B 266FSB 17 X 133.33
XP2800+ = 2.17GHz Barton 333FSB 13 X 166.66
XP2900+ = 2.33GHz T-bred B 333FSB 14 X 166.66
XP3000+ = 2.33GHz Barton 333FSB 14 X 166.66

CPUs will be produced on T-bred B. The reason behind this is of course there is no point on producing a less prone more faulty T-bred(A) which would leave AMD to requiring that much more space to be used to build these older cores in there factorys.

Instead, AMD brang out the AROIA core Palominos in hopes to keep the XP1600~2200+ line going. All first generation T-breds will no longer be built due to cost and marketing.

I guess it leaves us to wave goodbye to the XP1700~2100+ T-breds which will never see the light of day no more.


DS-Master

futura2001
09-08-02, 11:19 PM
Wait! Shouldn't it be
XP2700+ = 2.27GHz T-bred B 266FSB
XP2800+ = 2.27GHz T-bred B 333FSB
?
Futura

john240sx
09-09-02, 01:43 AM
wouldn't the XP2300 actually outperform the XP2600 because of the 333Mhz FSB?
i learned that FSB is way more important then cpu speed.

Arkaine23
09-09-02, 02:04 AM
It depends on how much FSB increase and total clock speed you can get with a particular motherboard/memory combo. A 2400+ may be able to get more speed but about the same final FSB as a 2300+ or 2500+ on a kt333- it would go from 133 to say 190, while going from 166 to say 190 would not be that much of an overclock. The 166's will need a 1/6 divisor or the abilty to increase the mutliplyer (above 12.5) or both to reach their potential.

I posted the about this stuff over here (http://forum.oc-forums.com/vb/showthread.php?s=&threadid=120614&highlight=barton+tbred+multiplyer), but had some different numbers. In some cases I was guessing. I would like to straighten out the clock speed conjecture and get it all straight and out in the open so I can start figuring what options I'll need in my next motherboard and what kind of memory speed I'll want. I'd also like to find out how all these new cpu's will unlock (including multiplyers above 12.5).

Edit: Here it is, copied and pasted:

XP 2200+ is 13.5 x 133 1800 Mhz
XP 2300+ is 11 x 166 1833 Mhz
XP 2400+ is 15 x 133 2000 Mhz (only one I guessed at and had wrong originally)
XP 2500+ is 12 x 166 2000 Mhz
XP 2600+ is 16 x 133 2133 Mhz
XP 2700+ is 13 x 166 2166 mhz
XP 2800+ is 17 x 133 2266 Mhz
XP 2900+ is 14 x 166 2333 Mhz
There may be more as the Tbred "B" cores mature.

Barton: Will these be similar in clock speed to their Tbred cousins? It could be they will be clocked lower but that their PR rating is the same or higher because the PR relates to performance and bartons should probably get a boost in PR rating because of extra L2 cache. My guess is that these will be right around 2000 Mhz with 166 FSB. Like the 2300+ and 2500+ in clock speed, but its only a guess.

2800+
3000+
And I think Bartons up to 3600+ will come out.

residentevil2
09-09-02, 02:55 AM
one thing i learnd from u ds-master is that u give out great info all the time good work :P

OC-Master
09-09-02, 10:37 AM
Originally posted by futura2001
Wait! Shouldn't it be
XP2700+ = 2.27GHz T-bred B 266FSB
XP2800+ = 2.27GHz T-bred B 333FSB
?
Futura

Never mind, I'm cracked up or something. Thanks for noticing. I also forgot the 1.87GHz Athlon, cant leave that one out!

XP2300+ = 1.83GHz T-bred B 333FSB 11 X 166.66
XP2400+ = 2.00GHz T-bred B 266FSB 15 X 133.33
XP2500+ = 2.00GHz T-bred B 333FSB 12 X 166.66
XP2600+ = 2.13GHz T-bred B 266FSB 16 X 133.33
XP2700+ = 2.17GHz T-bred B 333FSB 13 X 166.66
XP2800+ = 2.27GHz T-bred B 266FSB 17 X 133.33
XP2800+ = 2.17GHz Barton 333FSB 13 X 166.66
XP2900+ = 2.33GHz T-bred B 333FSB 14 X 166.66
XP3000+ = 2.33GHz Barton 333FSB 14 X 166.66

All corrections have been made. It threw everything out when I forgot to add the 1.87GHz 333FSB CPU into the table. EVERYTHING is now accurate! Thanks guyz for noticing.


DS-Master

Arkaine23
09-09-02, 10:48 AM
Over here (http://www.guru3d.com/files/), you said the 2300+ was 1833, not 1867, and that the 2900+ is 2333, not 2267. The speeds you posted for those two would be hard to get with a 166 FSB and standard multiplyers. You had it right in your other post. :)

Emericana
09-09-02, 06:17 PM
dsmaster where do you get all this info do you know someone from amd or something?

OC-Master
09-09-02, 11:53 PM
Originally posted by DS-Master


Never mind, I'm cracked up or something. Thanks for noticing. I also forgot the 1.87GHz Athlon, cant leave that one out!

XP2300+ = 1.83GHz T-bred B 333FSB 11 X 166.66
XP2400+ = 2.00GHz T-bred B 266FSB 15 X 133.33
XP2500+ = 2.00GHz T-bred B 333FSB 12 X 166.66
XP2600+ = 2.13GHz T-bred B 266FSB 16 X 133.33
XP2700+ = 2.17GHz T-bred B 333FSB 13 X 166.66
XP2800+ = 2.27GHz T-bred B 266FSB 17 X 133.33
XP2800+ = 2.17GHz Barton 333FSB 13 X 166.66
XP2900+ = 2.33GHz T-bred B 333FSB 14 X 166.66
XP3000+ = 2.33GHz Barton 333FSB 14 X 166.66

All corrections have been made. It threw everything out when I forgot to add the 1.87GHz 333FSB CPU into the table. EVERYTHING is now accurate! Thanks guyz for noticing.


DS-Master

OC-Master
09-09-02, 11:54 PM
Originally posted by Emericana
dsmaster where do you get all this info do you know someone from amd or something?

Nope, from the Register, Inquirer, X-bit and news groups. It all comes together. NOW I think I got it right this time.

OC-Master
09-10-02, 12:14 AM
I hope you guys/gals are happy, I spent over 1 hour figuring out the PR properly with a little news group help. Thanks again everyone.


DS-Master

OC Detective
09-10-02, 01:47 AM
If the XP1700 - XP2100 are not going to be manufactured would it make marketing sense to bring out a XP2300+ based on the B core when a XP2400+ albeit based on the A core is already in the market? To the unwashed masses this may seem like a backward step - whats the betting the B core starts at a higher rating?

OC-Master
09-10-02, 10:57 AM
Originally posted by OC Detective
If the XP1700 - XP2100 are not going to be manufactured would it make marketing sense to bring out a XP2300+ based on the B core when a XP2400+ albeit based on the A core is already in the market? To the unwashed masses this may seem like a backward step - whats the betting the B core starts at a higher rating?

Well, if you remember the old PR method, a 2GHz AthlonXP with the Palomino or Thoroughbred A core would have been called XP2500+ but now, AMD decided to go with XP2400+ due to the fact that the PR's were getting way to high for releative performance rating.


DS-Master

dropadrop
09-10-02, 12:06 PM
Maby there's something I'm not quite getting here, but I'll pitch in anyway.

I understand the pr ratings are amd's awnser to the p4's higher clockrate. The way I see it, the increased cache on the barton will make it more powerfull, and so it should have a higher pr rating against clockspeed. Atleast I would hope that it increases performance in certain applications? Maby even in everything.

I don't really see any reason why amd would release the xp2800 barton (512cache), and the xp2900 tbred (256cache), like mentioned in the list. This would only mean that the lower pr'd 2800 would kick the 2900's ass bigtime!

*edit* Ok, I just noticed you had left a bit of room for that, but isn't that too little?

OC-Master
09-10-02, 03:41 PM
Originally posted by dropadrop
Maby there's something I'm not quite getting here, but I'll pitch in anyway.

I understand the pr ratings are amd's awnser to the p4's higher clockrate. The way I see it, the increased cache on the barton will make it more powerfull, and so it should have a higher pr rating against clockspeed. Atleast I would hope that it increases performance in certain applications? Maby even in everything.

I don't really see any reason why amd would release the xp2800 barton (512cache), and the xp2900 tbred (256cache), like mentioned in the list. This would only mean that the lower pr'd 2800 would kick the 2900's ass bigtime!

*edit* Ok, I just noticed you had left a bit of room for that, but isn't that too little?

Nope, because 333FSB = 83MHz difference with = PR 100 points. Now also add the additional 100 PR from the Bartons extra cache and you get your difference.

Essentially, the XP2900+ is over 166MHz faster than the XP2800+ Barton.

Basically, 83MHz PR = the additional 256KB of L2 cache PR = 100

dropadrop
09-10-02, 05:28 PM
Damn, I was expecting a bit more from the increase of cache then 100mhz worth...

Someone somewhere threw the idea in the air, that the small cache limits the athlon at high clockspeeds. That this would be the reason the diferance between the athlon & P4 get closer and closer to each other around 2.5ghz and above.

I'm not a tech-head, it just sounded like the extra cache might keep that diferance wider. :rolleyes:

Will the cache help more in some specific kind of programs? Like audio software, where the computer is counting 5-6 advanced synths in reatime?

dropadrop
09-11-02, 06:33 AM
According to these numbers the xp2300 will have a 11.26 multiplier, and the 2800 will have a 12.831 multiplyer. These are the only two I checked, but it clearly states that these numbers are only speculation.

OC-Master
09-11-02, 02:56 PM
Originally posted by dropadrop
According to these numbers the xp2300 will have a 11.26 multiplier, and the 2800 will have a 12.831 multiplyer. These are the only two I checked, but it clearly states that these numbers are only speculation.


XP2300 = 11 X 166.66 = 1833MHz (1.83GHz)


Where did you get 11.26???

russian
09-24-02, 10:15 PM
http://www.angelfire.com/anime2/darkcustom/athlonxprating.xls

That is what I've got. I know the multipliers are strange, but atleast there is a pattern.

For example... 2700+ mhz should equal to 2600+ mhz.

Malakai
09-24-02, 11:59 PM
Barton's are gonna be yummy!


thanks for the info AXIA (cant get used to the new name, DS sucked ass:D)

i think upgrading my 1600+ to a 2+ghz Barton will be very worth it.

dropadrop:
i think the extra cache is gonna help out a lot more than you think;)

RoadWarrior
09-25-02, 05:49 PM
Yeah the extra cache might be like the K6-2 vs K6-3.

One thing though, although AMD is dropping Tbred A from it's fab, it was planning to get UMC to produce on a 13Micron process, so maybe the low end tbreds are gonna end up with UMC. Also the Duron line was meant to become the current XPs, and it was rumoured UMC might be making those.

I am quite looking forward to seeing some UMC produced chips because I have this great little UMC U5s 486 clone CPU that has freakin' amazing performance and overclcokability. It's rated at 33Mhz, at stock speed it gets integer scores as fast as a 60Mhz P5 pentium, or a 63Mhz pentium overdrive, and overclocked at 60Mhz (and it's the buses on the board limiting things here, the CPU might hit 70+ if the board would let it) it gets scores like it's a P90 or better. No FPU on it though, but still it's an amazingly faszt and cool chip for it's day, and I think AMD might have to clobber the UMC CPUs somehow, otherwise you're gonna be buying Duron 1800s that will run at 2.2Ghz, instead of their XP2600-2800s.

maybe

Road Warrior