XP 3200+ Reviews

Yup the 400FSB Barton is out :

http://www.lostcircuits.com/cpu/amd_xp3200new/
http://tech-report.com/reviews/2003q2/athlonxp-3200/index.x?pg=1
http://www.gamersdepot.com/hardware/cpus/amd/athlon/3200+/001.htm
http://www.amdmb.com/article-display.php?ArticleID=237
http://www.3dvelocity.com/reviews/Barton3200/Barton3200_1.htm
http://www.amdworld.co.uk/bar3200.htm

http://www.tomshardware.com/cpu/20030513/index.html ( Tom's usual Intel biased benchmarks )

Thanks for the links

I found this part interesting in one of the reviews:

I wouldn't have much of a problem with that if the Athlon XP 3200+ offered a friendly price-performance ratio, but as of today, that's not quite the case. The Athlon XP 3200+ will sell for $464 in lots of 1000 units. To compare, the Pentium 4 3.0GHz currently lists for $417. AMD may well push street prices for the 3200+ below those of the P4 3.0GHz, but the actual list price for the 3200+ looks mighty steep right about now. For that kind of money, I'd rather have a fast Pentium 4 with an 800MHz bus.

Click to expand...

Why would they price it so high? AMD has always had that price edge?

well the 3200+ is set to compete with teh PIV 3.2 Ghz and the 3.2 Ghz isnt' available yet so they have nothing to compare to... they porbaly made their best estitamte of the cost of the new piv and that's what they set their price according to.. atleast that's how i see it...

Tw00sh said:

I found this part interesting in one of the reviews:



Why would they price it so high? AMD has always had that price edge?

Click to expand...

The Amd now is a little more than the intel. but when intel comes out with a 3.2, it will be much much more expensive. by that time the Amd will have droped. Keep in mind, that the 3.06 has been out for a long time, the 3200 just came out

This CPU is a disappointment The 3.2C will destroy this.

Toscani said:

This CPU is a disappointment The 3.2C will destroy this.

Click to expand...

twice the FSB i would hope it does...

Toscani said:

This CPU is a disappointment The 3.2C will destroy this.

Click to expand...

??

The 3.2C will? I think the 3.06 and the 3.0C already did. When AMD first released the barton I thought it should have a little more clockspeed to justify its PR rating. Now they have raised the PR rating again while only increasing the clockspeed by 33mhz (xp3000 is 2166mhz, xp3200 is 2200mhz) and hoped that the 200mhz FSB would make up the difference. Back in the days of the Pally and Tbred the PR ratings were pretty comparable to the P4, each CPU would win roughly half the benches at an equal PR rating/clockspeed (ie. xp2200 Vs P4 2.2), now that a lot of people believe the PR rating AMD keeps raising it but they forgot to increase the performance of the CPU to match the rating.

Check out Tom's review (yeah I know some people don't like tom's), the xp3200 doesn't win one bench and is usually beaten by both the 3.0C and 3.06 and in a lot of benches it is beaten by everything from the P4 2.8 and up.

I am less then impressed with AMDs recent CPUs (basically all bartons but especially the xp3200+), all they needed to do was up the clockspeed a little bit more and it would have been a strong competitor.

edited to fix a typo...

Kinda makes me wonder if they are using thier resources to fine tune the 64 bit chips and just throwing something out there to claim they have something new?

More or less and PR (Public Relations) thing more than an actual product release.

I hope this drives down prices for the 2500+ in time for the 8rda3+ release.

DaddyB, You said it right.

Larger 512KB L2 cache in Barton and faster official FSB 200 MHz in 3200+ only help certain applications that requires larger cahce and are memory and FSB intensive.

And for many of us, we already are running 200+ FSB. And we won't see the advantage of the 3200+ since its clock speed is lacking compared to a 1700+ DLT3C oc'ed.

The raw clock speed would benefit everything and that is what XP/Barton needs to competite with P4. The Barton and the 3200+ have the same IPC (instruction per cycle for measuring CPU raw power) as the high clock rate 1700+ DLT3C. But ends up the 1700+ wins because it can be clocked to 2.4-2.5 GHz on air, whereas the current Barton and 3200+ max around 2.2 to 2.3 GHz on air. That is, the PR rating is increased (??) while raw CPU IPC kept the same.

In order to catch up with P4, any Tbred B, Barton has to have a clock speed = P4_clock_freq/1.28 to match the IPC for integer MIPS. And about 1.22 x P4_clock_freq for floating point MFLOPS.

The 1700+ at 2.5 GHz, will match with a P4 at 1.28 x 2.5 = 3.2 GHz (for integer computation).

A Barton at 2.2 GHz, can only match with a P4 at 1.28 x 2.2 = 2.8 GHz. That is why the Barton falls short in many benchmark to P4 due to CPU raw power and memory bus speed since effective QDR is 50-60% faster (not 100%) than DDR, assuming same bus clock.

We'd like to see a cross between 1700+ DLT3C and the Barton and the 3200+ for raw clock speed and CPU power, it would be much better then. Of course, the QDR for AMD, but maybe this is too much to ask for now!!!

The PR rating has become a joke, I was really hoping that they wouldn't resort to this. It's been rumoured for a while that the 3200+ would only be 2.2GHz, but I was really hoping that they'd smarten up and release it at 2.3GHz. Even then it would probably get dusted by the P4 3.2C though. I really hope AMD wakes up and stops playing these games before they lose all credibility.

Tw00sh said:

Kinda makes me wonder if they are using thier resources to fine tune the 64 bit chips and just throwing something out there to claim they have something new?

More or less and PR (Public Relations) thing more than an actual product release.

Click to expand...

Indeed that may be the case. Most people are buying 1700s, 2100s, 2400s, and 2500s now anyway, so the 2% they lose on their fastest chip won't hurt them as much and they'll make it all back with Hammer

Even on the Intel side of the fence, most people are buying 2.4s and 2.53s as opposed to 3.0s and 3.06es.

DaddyB ................. I also disagree with the numbering system since the barton's introduction but Tom is well known to pick well known Intel benches and programs which are optimised for Intel . In fact I remember seeing several reviews where the guys at Tom's actually brag about having the latest patches for hyperthreading or SSE2 installed on their rigs . But yet in this review they take great issue with AMD offering an enhanced dll . Where was the moral integrity when intel was releasing its special patches . Do a search on [H] , Anand or Tom and you will find that last year Sysmark 2002 was PROVEN to be a biased benchmark that was at the minimum a poor bench which didn't in any way represent real world processes but tested many obscure silly functions which are hardly used but 'happen' to do well on the P4 . At worst and likely in reality it is a cheating piece of software . Last year all the big reviewers were forced to admit that that was the case and even showed that they knew it before . But yet they continue to test with it , and then Tom's has the gall to object to an AMD patch .

Take a look at Tom's CPU reviews of late and see how many of his benches are intel biased or programs that no normal person uses , not even many professionals . And carefully look at the tests which AMD chips may win which they've been conveniently ignoring , Spec View Perf , Folding ( I dare them , since they use SETI ) , UT 2003 , Science mark , 3dsmax and just about anything which requires raw number crunching or FPU power . When that site reviewed the 3000+ they made it look like rubbish compared to the P4 3.06 . But Anand who did more tests, in particular games the XP looked alot better and in another review at ACE's ( I think it was there ) they had a much wider selection of benches and the XP 3000 + won at least half of them !!!

As a result I'm not too surprised if I see an AMD chip losing every bench at Tom's ,AMD chips have to be much , much better than their Intel counterparts to be successfull on that site . So I think the real world performance is much better than sites which use tests like Tom's would suggest , but the PR rating is still not properly in line , AMD is trying to creep upwards in GHz as slowly as possible and for as much profit as possible as well .

By the way good post ..................... hitechjb1

cowboyx:

" Fear of GOD is the beginning of wisdom "-psalms
Love of GOD IS wisdom - logos

anyway: we had a grand discussion on performance over in the intel forums, focused on the lack of pIV performance. dustybird posted some illuminating technical material which demonstrated how the pIV falls behind the AMD (even the pIII and Tualy) architecture in clock for clock performance. in fact the pIV gets whopped clock for clock.

i have always built intels but when i put together my first amd box i was laughing, and my brother asked if it was because it was so bad. i said no, it was because it was so rediculously FAST. Bios posted, programs installed, things opened, so fast that it reduced me to laughter in disbelief. for my sister i had built a faster clocked pIV which is so slow that even my sister, a complete noobie with no computer experience hates waiting for everything! i mean, it's total dog. but it benches well!

i agree with the higher pr being pr...it will appeal to lay users who won't know the difference, but the hobbyists and pros obviously are on to it. and we have been more attuned to the reality behind the marketing all along.

Correct buy a Tbred B 1700 , 1800 or 2100

Logos ............. Fear of GOD is the beginning of knowledge ........... another verse

Cowboy X said:

Yup the 400FSB Barton is out :


http://www.tomshardware.com/cpu/20030513/index.html ( Tom's usual Intel biased benchmarks )

Click to expand...

So tom somehow fixed the benchmark apps he used to make the intel beat the amd?..lol - benchmarks are benchmarks - they show it all - perhaps toms wording maybe in favour of intel, but again, look @ the benchmarks - i dont even read what they write anymore, same thing over and over - i look @ the numbers, what counts.

It wasn't that long ago when everyone was accusing Tom's of being bias in favour of AMD (which they were to some extent).

The new Intel chips are gonna pwn. People are hitting the 3.4 on the stock fan with the 2.6C. They are going to be the new chips for overclockers.

What I want to know is how well does it overclock (I'm to lazy to read all the reviews).