was anyone else disappointed with the XP2600+ benchmarks?

[Alex]

The XP 2600+ is supposed to be as fast as a 2.6 ghz p4 processor, yet it manages to underperform the 2.5ghz p4 in a great deal of the tests... bah

 

[jwo7777777]

The reviews I read at THG and HOCP were more interested in trumpeting the OC'ability of the processors. I am not disturbed if a 2.6GHz processor is slightly or even mildly faster than a 2600+ one. I am instead encouraged that AMD and Intel are trying to slug it out toe to toe. It keeps them on their technical toes, prices down, and variety up. The downside is that this competition stifles mass innovation and rewards momentum, but everything has a price.

What I find interesting is that AMD keeps piling on the metal layers while Intel deepens their pipeline in the race to dominate performance. It will be interesting to see what happens. The question one may ask is "What performance would AMD have garnered if they had gone to ten or eleven or more metal layers?"

I'm guessing this revamp of the XP core was to tide them over until ClawHammer.

 

[Alex]

true... well from that point of view the release was a pretty good move... im really eager to see what the 166mhz fsb can do for the athlonxp and hammer...

 

[Athlon4all]

Originally posted by: franguinho
The XP 2600+ is supposed to be as fast as a 2.6 ghz p4 processor, yet it manages to underperform the 2.5ghz p4 in a great deal of the tests... bah

I have to admit, I was initally somewhat dissapointed, and I had a feeling we may have been seeing for the first time, the athlon's "old" 266fsb was limiting a otherwise faster CPU, but it did increase about as much as could be expected too. In any case, the XP 2600 should've been 2500+ agreed.

Anyways, I don't think anyone should be disappointed with the XP 2600+ launch. Remember how badly the XP's were defeated by the 2.53 only a few days ago? Well now the 2600 is very competitve with the 2.53. Just as it was between the 2.26 and the 2200+ before this launch, both the 2.53 and the 2600+ win some and lose some and tie some. The XP wins in non-SSE2 3D Rendering, MP3 Encoding, and in Professional 3D Worksation apps (ie SPEC Viewperf), where as the the 2.53 owns in SSE2 3D Rendering, MPEG-4 Encoding and Content Creation, and Comanche 4. Also, they are for all practical purposes tied in most games (2.53 leads by 3% in JN2, 2600 wins Serious Sam 2 buy 2%).

I certainly am not disappointed. It has been a long time since AMD has had such a large clock speed increase. I was expecting it to do a tad better than it has but I am not disappointed with its performance, what I am disappointed with is, for the first time, AMD's PR Rating is saying its faster than it really is.

EDIT:

true... well from that point of view the release was a pretty good move... im really eager to see what the 166mhz fsb can do for the athlonxp and hammer...

Well, Ace's Hardware did oc their XP 2100+ to 166fsb, not a big improvement.

 

[mfavin]

With the 2.53 Ghz P4 getting a whooping 62% price cut next week, I think that's the sweet spot for me. This should mean the processor will be well below $200. If the XP 2600 were 20% lower in price than the P4 2.53 I might consider it, but otherwise no way.

 

[Actaeon]

I wasn't disappointed at all, though it does lose to the 2.53 P4 in some benchmarks, in others, the XP2600+ beats the P4. Also the newer PR rating is also more accurate now then it was before. It shows AMD is still "in the game", lowered the price tags on the Intel and AMD CPUs, and gives us more options as the consumer for purchasing new CPUs.

I'd like to type more, but this keyboard is acting all funky on me.

 

[apoppin]

Yes and no.

Not disappointed that AMD is again competing with Intel with a o/c'able CPU (again). And LOTs to look forward to with Barton and Hammer!

A little disappointed in that the 2600 acts more like a 2500+ (even though the PR was adjusted down). I would have preferred that AMD give the 2600 a more conservative PR.

Overall, 'thumbs-up' AMD!

 

[spanky]

i'm not disappointed. clock for clock (throw the pr rating out the window), it's a pretty nice chip. at least that's what i think.

 

[ToBeMe]

Yeah, it should be the 2500+ I suppose, but at least they are closer. I'm also skeptical as to whether the increased FSB will acheive much in and of itself for AMD. With the Intel 2.8 basicly here, AMD had to do something, but, the 2600+'s aren't even available yet........they need to get it in the channel and available NOW. AMD needs the Hammer ASAP and they are going to need to make adjustments to their "PR" schemeIMHO the Athlon core, no matter how they attempt to give it boosts, is not going to scale with the P4 in peformance and the performance gap will widen with each release...........Come on Hammer........

 

[Athlon4all]

Originally posted by: ToBeMe
Yeah, it should be the 2500+ I suppose, but at least they are closer. I'm also skeptical as to whether the increased FSB will acheive much in and of itself for AMD. With the Intel 2.8 basicly here, AMD had to do something, but, the 2600+'s aren't even available yet........they need to get it in the channel and available NOW. AMD needs the Hammer ASAP and they are going to need to make adjustments to their "PR" schemeIMHO the Athlon core, no matter how they attempt to give it boosts, is not going to scale with the P4 in peformance and the performance gap will widen with each release...........Come on Hammer........

Good points ToBeMe. We will see about the faster fsb. I think it will be like the 533fsb launch with the P4 a few months ago, nothing big, but down the road, it will be necessary.

I think you are certainly right, that AMD is just not going to scale well at all vs the P4. Thats the picture we have seen all year until today with AMD increasing clock speed by only 200MHz for the past 9 months before today (where it increased 333MHz:Q). Honestly, I think that if AMD has to they can compete with the 3.06 Northwood in the same sense they are competing with the 2.53, but that would mean practically maxing out the Athlon XP core at like 2.4-2.5GHz.

 

[CrazySaint]

I think that unless AMD knows they can get Hammer out (in retail channels in mass production) by January, they need to hurry up with the 512k barton to boost performance, or else they'll have to come out with a new core revision every 3-500MHz in order to scale with the Northwoods. I do, however, think that the 2600+ PR rating is pretty accurate, as it seems to beat the 2.53GHz P4 in about as many benchmarks as it loses, according to Anand's review, and not counting specperf.

 

[DeschutesCore]

Originally posted by: franguinho
The XP 2600+ is supposed to be as fast as a 2.6 ghz p4 processor, yet it manages to underperform the 2.5ghz p4 in a great deal of the tests... bah

The Athlon PR rating is not based on P4. It is based on the performance of a Thunderbird had one been made at 2.6 GHz. AMD made quite a big ordeal about making sure we do NOT spread ideas like this.

And secondly, no, there's nothing to be disappointed about, the XP 2100+ has dropped to ~100, AMD is still producing chips people can use on older boards at excellent speeds. I, personally, can't find one damned thing wrong with this.

DC

(Edit: Damned typos)

 

[Markfw]

OK. One more time, a quote from Tomshardware review , and which I remember from AMD's site when the XP rating started:

Quote from Tomshardware review of Xp2600+:
With the launch of the Athlon XP with the Palomino core, AMD also introduced its new model number system, called P-rating. The idea behind this was to create a direct comparison to arch-rival Intel, which markets its CPUs using "real" megahertz figures. This is a language that the less tech-savvy customers can understand: megahertz and gigabyte. As we've already determined in numerous articles, a processor's performance capabilities cannot be analyzed in terms of pure clock rate numbers.

 

[DeschutesCore]

Originally posted by: Markfw900
OK. One more time, a quote from Tomshardware review , and which I remember from AMD's site when the XP rating started:

Quote from Tomshardware review of Xp2600+:
With the launch of the Athlon XP with the Palomino core, AMD also introduced its new model number system, called P-rating. The idea behind this was to create a direct comparison to arch-rival Intel, which markets its CPUs using "real" megahertz figures. This is a language that the less tech-savvy customers can understand: megahertz and gigabyte. As we've already determined in numerous articles, a processor's performance capabilities cannot be analyzed in terms of pure clock rate numbers.

Did he really say gigabyte or did the translator screw up AGAIN? Either way, I think I'll trust AMD over Tom on this one.

DC

 

[John]

The Athlon PR rating is not based on P4. It is based on the performance of a Thunderbird had one been made at 2.6 GHz. AMD made quite a big ordeal about making sure we do NOT spread ideas like this.

I disagree. The AMD tech tours I have attended do not back up the above statement either.

Read this

 

[Siddhartha]

Originally posted by: franguinho
The XP 2600+ is supposed to be as fast as a 2.6 ghz p4 processor, yet it manages to underperform the 2.5ghz p4 in a great deal of the tests... bah

No I am not disapointed:
1. From the reviews I have read the AMD cpu wins about half of the benchmarks:
A. The P4 2.54 gHz excels in memory bandwidth intensive tests
B. The XP 2600+ excels in CPU intensive tests
C. Right now the XP 2600+, compared to the P4, represents a better value.
2. Right now both XP 2600+ and the P4 2.54 gHz would meet my needs nicely.
3. The XP 2600+ is competitive with the P4 2.54 cpu which will hopefully drive prices down.
4. The XP 2600+ overclocks which means there hopefully will be less breast beating on hardware website forums.
5. I have always considered the AMD's XP "PR" to be nothing more than model numbers and have only paid the cpu's actual performance.

 

[DeschutesCore]

Then I shall concede victory to the new theory. AMD was very solid about this stance when we first rolled out the XP. Why they've backed out is beyond me.

DC

 

[CubicZirconia]

Why they've backed out is beyond me.

Maybe because comparing the xp to the p4 makes a whole lot more sense. If some random person walks into a store and sees something called "2600+" and another called "2600 mhz" they will assume they are very similar. Its much easier to say "yes, they are similar" than "No, actually 2600+ is a relative number figured by approximating the speed of a cpu based on the thunderbird core running at 2600 mhz." Besides, benchmarks show that it is relatively accurate.

 

[Markfw]

DeschutesCore: I don't want to start a flame, but my whole point here is that this has ALLWAYS been their stand (AMD). No, I don't have the text to copy from their site, because I can't find it anymore, but I remember it as if it was yesterday. PLEASE believe me that this was exactly their stance at the time. Their whole idea, was the "megahertz is not all", and they wanted to try to even the scales, and even today are trying to get a "CPU performance rating system" that is standard among all processors.

Edit: I don't suppose there is anybody out there that was anal enough to copy their white paper or text from their site when they first started this rating system ?? If so, PLEASE post it here to stop this argument once and for all.

 

[KhoiFather]

You people shouldn't be dissappointed at all. I mean, how can you be dissappointed when a xp processor that is only 2.16ghz can come close and also beat what the P4 2.53ghz. I mean, for all I care, I'm happy, because you know why, all I know is that if for once you would benchmark the same speed, the XP would kill ass severely. Nothing is greater then a cpu that is a lot cheaper and can deliver the same performance.

 

[bcterps]

I'm not disappointed in the benchmarks, I'm disappointed in the availability! If these chips were available now it would make a difference, but with the P4 price cuts coming Sept 1, they'll only make the 2400/2600 less attractive. I dont know why these are available only to OEMs, they arent going to make the best use out of them.

I dont see how anyone can really be all that disappointed, the consumers are the ones that will gain the most from fierce competition.

 

[ALstonLoong]

The review is pretty good. xp2600+ perform very close or ahead p4 2.4-2.5 mhz and also come with lower voltage. The reason i dont choose xp over p4 is because i need to buy alots of overclocking fans to cool it down. Hopefully amd will introduce their new xp with fancy heat spreader like p4. Until that time ...i will buy a amd with lower price and better performance compare to p4.

 

[BD231]

I think the PR rating needs more work than AMD put into it, it's not very reliable now.

 

[PlatinumGold]

am i the only one that saw these benchmarks differently?? i thought the XP's handily won all the CPU Intensive Benchmarks and that the P4's won all the memory bandwidth intensive benchmarks.

I would say that the xp has better upward mobility. AMD is more likely to resolve the Memory bandwidth problems.

 

[Random Variable]

I'm disappointed that it doesn't have a 166MHz DDR FSB.