AMD introduces the Athlon XP 3200+...adds a whopping 33.3 MHz!
Okay, this is just getting to be completely retarded.
They already rescaled the Quantispeed curve when Barton was released, and now they're doing it again for their 333-->400 MHz FSB bump.
As it stands...
Thoroughbred Athlon XP 2800+ = 2250 MHz (333 MHz)
Barton Athlon XP 3000+ = 2167 MHz (333 MHz)
Barton Athlon XP 3200+ = 2200 MHz (400 MHz)
What's next, completely arbitrary Quantispeed Ratings? Those nonsense Opteron model numbers? (242, 244, etc?)
They already rescaled the Quantispeed curve when Barton was released, and now they're doing it again for their 333-->400 MHz FSB bump.
As it stands...
Thoroughbred Athlon XP 2800+ = 2250 MHz (333 MHz)
Barton Athlon XP 3000+ = 2167 MHz (333 MHz)
Barton Athlon XP 3200+ = 2200 MHz (400 MHz)
What's next, completely arbitrary Quantispeed Ratings? Those nonsense Opteron model numbers? (242, 244, etc?)
Comments
Originally posted by der Kopf
Maybe there are other things going on as well, no? That FSB bump is already sure to speed things up a lot. At the risk of sounding uninformed, it sounds like someone being disphoric while hearing the 970 introduced at 1.4 GHz. "But we have a proc running that fast already...". Isn't there more to this chip that meets the eye, and that makes it a good deal faster?
http://www.tomshardware.com/cpu/20030513/index.html
http://www.extremetech.com/article2/...1081507,00.asp
http://www.tech-report.com/onearticle.x/5126
Bad AMD. What an embarrasment.
http://www.anandtech.com/showdoc.html?i=1826
Oh, no! We're DOOMED!
DOOMED
AMD has to develop faster athlon 64 if he want to stay in the competition.
2003 will be an hard year for them.
Originally posted by Gizzmonic
Hyperthreading is HYPE. It causes maybe 1% speed increase. people forget that the first hyperthreading capable P4s also had biggest caches and higher clockspeeds.
HT can be disactived, benchmarks with HT disactived or actived where published in many newspapers. In some case HT bring an extra 20 % bonus, in other case it slow down the computer.
But your 1 % number is wrong, generally it's near 15 % .
If HT was a failure, IBM will not include one in the power 5.
Originally posted by Gizzmonic
Hyperthreading is HYPE. It causes maybe 1% speed increase. people forget that the first hyperthreading capable P4s also had biggest caches and higher clockspeeds.
Your ignorance speaks for itself. No further comment needed from me.
Personally, I think AMD does a decent job with keeping their ratings close to the P4's MHz ratings (ie. a XP 2000 is about the equivalent of a 2GHz P4, a XP 2200 about equivalent to a 2.2GHz P4, etc.). At least, it seems that way according to these Tom's Hardware benchmarks.
Originally posted by Eugene
What's next, completely arbitrary Quantispeed Ratings? Those nonsense Opteron model numbers? (242, 244, etc?)
It's true than the 3200+ rating is a too optimistic, but it is really becoming difficult to compare these different architectures. Intel and AMD have chosen two so much different approach for there chips. It is even becoming worse with the K8. The choice is now so much based on the application: content creation -> P4, complex FPU computing -> AXP, games -> it depens which one,...
How can they define a QS ? And if they don't define a QS, wich Joe buyer will buy AMD-based compputer ? This is clearly a big problem for AMD. But Intel is now facing the same problem with their P-M, what will they do ? Don't know \
About the Opteron model numbers, in wich way are they nonsense for you ? The whole idea or that they are begining at x40 ?
But the problem for AMD is that they need to sell chips.
The average consumer is buying by chosing the biggest number, if he see P4 3GHz or AXP 2.2GHz, he will never chose the AXP.
The even biggest problem is that he know the word "Pentium", but has probably never heard "Athlon" or AMD.
The other solution is to launch a huge campain to promote the Athlon brand name and to educate the average buyer. But AMD has no money for that.