AMD XP 3000+ Benchmarks

RCPilot · Feb 10, 2003

Well I suppose I ought to wade in here & put in my $.02 worth.

WTF is going on? I bewildered! Both of my favorites are coming out with less than stellar performances. I love a underdog. The FX & now the Barton! It bum's me out I'll tell you.

I don't exactly know what to do at this point. I've got some machines that I've been waiting for both of these to build for people, so now what do I do?

Am I going to have to cross to the other side (GULP) dare I say it, Intel & ATI & have to go on another learning curve?

I guess that I'm going to have to let all of this sink in.

Bummed out though I'll tell you I am. 2003 dosen't look got for either of them, dam!

If it ain't broke, take it apart & see why not!

knownalien · Feb 10, 2003

you know, i have been reading this site for years. And as far as being biased, I can truly say that THG really is a site that wants AMD to do well. And with that in mind, I think the qualification for that is that it might drive down CPU prices all together, something I think we are all for. I never got the impression that THG is overly critical of Intel. Certainly there are questions posed relating to manufacturing, development, marketing, implementation. But guess what, these same critisisms are leveled at AMD too!
Now back to the matter at hand. I am a novice, no doubt. And I am constantly dismayed at the idea that one need even tinker with anything in a computer to get great performance out of it. Overclocking should be a fun passtime, but now it seems like something you have to do to make the playing ground even. I am afraid I am getting nervous at this. For the regular like me who only knows enough to put a system together up to "specs", it seems as if I am being dared to go further and void any warranty I had. It seems to me that Intel at least puts its product out there aiming their product at "joes" like me who have no desire to tinker with the system. Isn't that the majority of people! Not everyone is super computer literate! But it seems to me that everytime I see AMD, "overclocking" is mentioned as if they go hand in hand. Listen, if there is such a performance boost, then let AMD overclock their CPU's and then sell them to us as such. Now it is getting to the point where buying an AMD CPU makes me feel like I am getting duped because it will never attain the numbers that sites like THG achieve.
For the record, my next purchase within the week is:

AMD 2800+ (old school, if I can find it)
Asus A7N8X (deluxe)
Antec case SX635BII (350 watt)
Taisol CPU fan
Maxtor 40 & maxtor 120
Corsair XMS PC-2700 C2 (2 X 256)
ATI All in wonder Radeon 9700 Pro
2 Antec 80 mm Case smart fans
Windows XP

It will be a fast system, but I don't pretend that it will beat the faster Pentiums.

sonoran · Feb 10, 2003

Just my $0.02 worth... An even better question is why does Tom's include overclocks in their standard benchmarks at all? After all, everyone who buys a processor is guaranteed to hit the stock speed, but no one is guaranteed to reach the same overclocks as Tom's. I think the overclocks ought to be in completely separate articles. Do the benchmarks at rated speeds.

RCPilot · Feb 10, 2003

Sweet system your putting together.

I'm like you. I've never OCed & really don't see a need to. I run pretty much stock out of the box myself & usually under 1 year old stuff. I know how to do it, but haven't seen the need to. I build for myself once a year usually & I'm fast enough. I always get good money out of my parts or system when I do build, that way.

I build a lot for other people, at least 2 machines a month or more. When they leave here there not OCed. If they want to OC I show them in BIOS how to do it. I've seen some burnt up machines that I put together, but I didn't do it. Hey, people like to play is all I can say.

The only real reason I build for anybody else is that way I can put together the latest stuff, see how it runs & buy what I have liked since the last time I built for myself. Keeps me on the edge & costs me nothing.

I have nothing against ATI or Intel. I'd rather put together AMD & Nvidia, because that's what I know best. I may know ATI & Intel a lot better soonr is all I can say right now.

I built in Nov. just before the Nforce. My rig I built then:

A7V8X W/everything
XP2400
Samsung PC2700
Ti4600
WD 80 gig JB
Antec 1080 AMG
ViewSonic a90f +

So I'm a few months off the top now. All I really do is game. I've got a whopping 6 or 7 gig's on the HDD.

BTW Look at the 1080 AMG, it comes with 3 smart fans in the case & a 430 w TruePower, might save you a buck or 2. Maybe not. Course you know what your doing though.

It's fun for me & that's the way I like it & what to keep it that way. It's not a way of life for me is all I can say.

If it ain't broke, take it apart & see why not!

slvr_phoenix · Feb 10, 2003

Just my $0.02 worth... An even better question is why does Tom's include overclocks in their standard benchmarks at all? After all, everyone who buys a processor is guaranteed to hit the stock speed, but no one is guaranteed to reach the same overclocks as Tom's. I think the overclocks ought to be in completely separate articles. Do the benchmarks at rated speeds.

In an ideal world I'd agree fully. That's exactly the way I think it should be done. Have one article strictly stock. (Stock speeds, retail heat sink, etc.) Then have a second article stricktly on how well it OCs.

The thing is, I've seen that done some times at other places and a lot of times in the OCing article, they don't properly compare the OCing results with the competition. So now you've got an article on how the new retail X compares to the competition's retail Y, and an article on how the oc X compares to the retail X, but there's no way to properly compare how the oc X compares to the retail Y and the oc Y. And so you end up hoping that all of the benchmarks were done with as much of the same hardware as possible with exactly the same software settings just so that you can actually compare all of your options. But unfortunately, you're just usually not that lucky.

So because of this and because of time spent involved in writing additional articles, I can see and generally live with having the OCing and stock info all in the same article. Yet I think that the OCing information should be seperated from all retail comparisons and article comments when it's done this way. It should, at most, be an additional segment about OCing on the end of the article. Just as importantly (if not more) the competition should get the same amount of effort and expertise to OC it as the reviewed product does. After all, comparing a skilled OC of a chip to a retail chip or a half-arsed OC of a chip is hardly fair.

PC Repair-Vol 1:Getting To Know Your PC.
PC Repair-Vol 2:Troubleshooting Your PC.
PC Repair-Vol 3:Having Trouble Troubleshooting Your PC?
PC Repair-Vol 4:Having Trouble Shooting Your PC?

knownalien · Feb 10, 2003

re: non-OC's ONLY!!!!!

I've gone back and studied the benchmarks a little more with the XP 3000. Now since I am going to get the XP2800(1st version) I was most interested in that. It seems that it is very often equal to a P4 2.8/133. That P4 costs $381. XP2800+ (1st version) costs $395. On a very few tests, mostly memory tests, the XP does rather poorly coming close to even a P4 1.8/100 or a P4 2.0/100. But what you must be paying for is a CPU that can nearly match a P4 3.06 (twice). One time was on the SiSoft-Sandra 2003 CPU Bench under Win XP! That's an important test, no? Mind you I went back and did NON-OC's only.
However, there is absolutely no justification for the price of the XP 3000 right now at nearly $650! Overclocks aside, XP2800 (1st version) is still the best AMD chip to get.

p.s. when I saw them breaking the wire out and tying together pins, that's when I knew they might be "overcompensating" a little.

imgod2u · Feb 10, 2003

But my point was that this end user will for example, talk to his buddy with the fastest amd and say "i have the better cpu" well no he does not. He has the fastest CPU but not the better CPU.
Because if amd made a athlon at 3ghz it would be faster.

That's only if you define "better" to be more work done per clock. But then, why is clockspeed made the standard factor in which to judge "goodness"? I can say that a processor is better because it's cheaper to produce for the same performance, that's an actual reason, because money is a real resource, hence it is better to conserve it while maintaining the same net result, that is efficiency. Is clockrate a resource? Do you get a limited amount of MHz when you buy a processor and therefore, need to conserve it and get as much "bang" for the MHz as possible to be "better"?
What are your throughts on Hypertransport? It transfers less data per MHz than standard FSB buses but it's very narrow design allows it to scale so high in MHz that it's able to surpass conventional design in bandwidth. Not to mention be much more flexible. Shouldn't that be what matter? The performance you get? Does it matter how that performance is reached?

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

slvr_phoenix · Feb 10, 2003

Wow imgod2u. You're being nice, replying to ctbfalcon and all. It took me a whole one glance and I decided that ctbfalcon wasn't even worth replying to. I'm just not feeling very generous today I guess. Good luck. I have the feeling you're going to have one long and pointless debate. (That is, if you get any intelligible response at all.)

PC Repair-Vol 1:Getting To Know Your PC.
PC Repair-Vol 2:Troubleshooting Your PC.
PC Repair-Vol 3:Having Trouble Troubleshooting Your PC?
PC Repair-Vol 4:Having Trouble Shooting Your PC?

imgod2u · Feb 11, 2003

It was either this, or study for differential equations.....

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

knownalien · Feb 11, 2003

i wish to add that the difference in price of the mobos is quite drastic.
A7N8X + XP2800 + 2 sticks ram = $733
$150 + $395 + $188 = $733

P4G8X + P4 3.06 HT + 2 sticks ram (or do you need 4) =
$230 + $655 + $118 (x2) = $1121

for a difference of $388.

there's you ATI All in Wonder Radeon 9700 PRO!

Edited by knownalien on 02/10/03 08:45 PM.

imgod2u · Feb 11, 2003

An equivalent comparison to the 3.06 P4 would be the Barton "3000+" which costs $630 last I checked.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

knownalien · Feb 11, 2003

God2SomeoneElse,

go back and read all of my posts in this thread to try and understand why I would close with the cost comparison of a 3.06 and an XP2800 (1st version).

Then get back to me "last time you check".

I'll make it easy for you: the XP3000 wasn't much better than the XP2800+ (1st ver.), so the cost effective approach as of now involves the XP2800+ . . . . . last I checked.

Edited by knownalien on 02/10/03 09:05 PM.

juin · Feb 11, 2003

PR rating end up to be only a big lie to codumer and the like when 800 FSB will be introduce a 2.6 will be faster that a XP 3000+ does AMD will reduce there PR rating no.

Anyway i have allready forseen all of this on 3000+.

Just next to the lab and the bunker you will find the marketing departement.

Schmide · Feb 11, 2003

Being that the XP3000+ takes at least a third of the benchmarks, I think it lives up to its PR. It may be a bit skewed, but its not totally backwards.

Who are these codumers and why are they getting screwed?

Dichromatic for your viewing plesure...

eden · Feb 11, 2003

at Juin's weird word!

Yes, the PR is now officially destroyed IMO, as Slvr would indicate. I won't delay my post and make my usual "new post-NEW-CPU review* rants that take a half page, but I will just say that I am also losing lots of respect to AMD, especially their marketting dept. It's far worse than Intel now, and they keep doing more and more screw-ups that can be outlined in Slvr's thread. I really don't get companies these days, nVidia, AMD, just WHO MANAGES them, seriously! I feel like even a newb to this forum and technology could be chosen and do a better job, it's just, too much shameful.

Anyways, the only aspect I think Barton will definitly be good at, is overclocking. It seems the die surface increase may be helping. Though THG did not report any temps yet for the XP3000, nor power consumption, which is weird. Anyways, if the XP2500+ is a downbinned CPU, it can most likely overclock to 2.33GHZ easily with unlocked multis, and perhaps a small FSB boost, and it'd likely be an excellent CPU for its price, and value.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden.

knownalien · Feb 11, 2003

maybe AMD will issue a revised version than what THG tested / / /

or

the mobo makers will find a better bios or chipset adjustment

or

the XP3000 will not really be nearly $650!!!!

I'd say it is really worth $500, but maybe the XP2500 is too since it can be OC'ed really high. Maybe the question that should be asked is if the XP3000 will be any faster on applications of the future much as the P4 HT's will have to wait for apps. compiled for HT technology. But really, I don't know jack about computers . . . so who knows.

imgod2u · Feb 11, 2003

I'll make it easy for you: the XP3000 wasn't much better than the XP2800+ (1st ver.), so the cost effective approach as of now involves the XP2800+ . . . . . last I checked.

By that argument, you should've used a 2.8 P4 in your listings, which is $366 right now on pricewatch. It's more of the same performance range as the "2800+" t-bred.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

Schmide · Feb 11, 2003

PR rating end up to be only a big lie to codumer

Ah I get it now, let me help you play with your words

I guess you were trying to mix the words

consumer with dumber (maybe con-dumber)

However, by transferring only the co of the first syllable of consumer and leaving the b off of dumb. (Cod-umer) You were left with cod, which is a fish, and umer which is an acronym for Ultrasonically Modulated Electron Resonance. A pungent-acousticly-adjusted-atomically-vibrating-fish. I get it now a week later and you thought PR was hard to grasp.

knownalien · Feb 11, 2003

Concerning NON-OVER CLOCKED CPU'S - XP2800, P4 3.06 & 2.8:

The XP2800 has the ability to match the P4 3.06 in a couple of benchmark tests, and match the P4 2.8 (as well as lower P4's), so right now the XP2800 seems somewhere in the middle of the P4 2.8 and 3.06. I think the XP2800 is a superior chip to the P4 2.8 except on maybe memory tests, so why use those two in my price comparison (go here if you have any further questions on the XP2800 vs. P4 2.8: <A HREF="http://www.tomshardware.com/cpu/20021001/index.html" target="_new">http://www.tomshardware.com/cpu/20021001/index.html</A>). With that in mind it seems plausible to compare the two [XP2800 & P4 3.06] as I have, though I am quite sure the P4 3.06 is superior. $235 (for OEM's only) superior??? However, the performance gains by HT technology have yet to arrive, and I remember waiting like hell for MMX encoded applications, which I eventually forgot about.

p.s. newegg.com has the P4 2.8 retail for $381.

juin · Feb 11, 2003

Ok it was consumer but i was lacking of time to re-read it.

I got some power consumation and temp
48 CEL T-bred 50 CEL Barton (2700 vs 3000+)
it was 68 watt vs 7X
I will get the link when i will have more time.
The increase in the L2 dont increase too much power cosumation that why Banias run with 1MB of L2.

I hope for AMD that increasing the L2 will resulte in better scaling for K8 that what we saw on a K7.I will like to see new L2 cache lantency figure.

400 FSB was a posibility look toms dont agreed with this.I agreed to myself also.To some extend the 800 FSB will be include.

Just next to the lab and the bunker you will find the marketing departement.

tlaughrey · Feb 11, 2003

Well, I've been waiting for the Barton for over a year now and it's finally here. The result is definitely mixed, but overall I think it's a good chip. I already bought my nForce2 motherboard anyway, so I'll be getting a Barton sooner or later as a last upgrade before the K8 arrives.

I suppose we can all argue about the minutiae of Barton vs. P4 vs. Thoroughbred and whether the PR is justified. I pretty much ignore the PR now anyway since it's gotten so convoluted. The bottom line for me is that it does perform at or near the level of the fastest currently available P4, and it does it while running about 900MHz slower. That's pretty impressive to me. A price of $650 for a 3000+ is a little steep, but it should come down. If not, I'll just get whatever Barton I can afford when the time comes.

The Barton's definitely not perfect, but then every chip that's released by Intel or AMD seems to fall short of expectations. But it should keep AMD reasonably competitive until they can release the Clawhammer. It's fun to argue about everything, but after all is said and done the Barton will give you excellent performance. If you don't like it, don't buy it.

Money talks. Mine always likes to say "goodbye."

Makaveli · Feb 12, 2003

I wasn't to impressed by the benchmarks. But I don't think its a bad processor at all. For a processor with a 900mhz handicap it held its own. Also considering alot more software is optimized for P4's, also add SSE2 into the mix.
But I guess alot of AMD fan were expecting some kinda magic and for the athlon to best the p4 in the majority of the benchmarks. But anyone that knows there stuff would know that doubling the cache was only gonna add so much performance 5-10% @ best. And the reduction in mhz didn't help it much either.

The positive things I saw in the benchmarks was that it was faster clock for clok than the Tbred. also it has bit less power consumtion. Most likely due to improved 0.13 process.

All in all the Athlon XP is almost at the end of its life. And it amazes me that AMD is still squeezing more juice out of it. I commend them on that. But the Athlon XP3000+ is not a P4 3.06 killer like most expected it to be.

In the end if I had to choose a new XP right now it would still be a T-bred (B) but mostly cause of cost. The bartons will bring more price drops to the rest of the athlon line and I welcome that. Also alot of people are so caught up in whats the fastest CPU these days and forgeting about whats best for them and there budget. Will you see a difference from an P4 3.06 to an Athlon XP3000 in game playing @ 1024x768 or higher. I prefer 1280x1024 on my Radeon 9700 pro.

AMD still has a major price/performance ratio in Canada where I live from medium to low end Cpu's. And well the high end cpus are priced similar. which is resonable since most sales are in the medium to low range and not the highest.

P.S toms benchmarks are servely lacking these days. I only check them out these days cause i've been coming to these forums for awhile now. I prefer Anandtech,tech report,Hardocp, Firingsquad reviews as of now.

I found this interesting also in the content creation winstone benchmarks. Would like to get feedback on that
<A HREF="http://www.tech-report.com/reviews/2003q1/athlonxp-3000/index.x?pg=6" target="_new">http://www.tech-report.com/reviews/2003q1/athlonxp-3000/index.x?pg=6</A>

imgod2u · Feb 12, 2003

The XP2800 has the ability to match the P4 3.06 in a couple of benchmark tests, and match the P4 2.8 (as well as lower P4's)

The P4 2.66 has the ability to match and beat the Athlon "2800+" in a "couple" of benchmarks as well. That doesn't make it exactly comparable as it looses more than it wins.

so right now the XP2800 seems somewhere in the middle of the P4 2.8 and 3.06. I think the XP2800 is a superior chip to the P4 2.8 except on maybe memory tests, so why use those two in my price comparison (go here if you have any further questions on the XP2800 vs. P4 2.8: http://www.tomshardware.com/cpu/20021001/index.html). With that in mind it seems plausible to compare the two [XP2800 & P4 3.06] as I have, though I am quite sure the P4 3.06 is superior. $235 (for OEM's only) superior??? However, the performance gains by HT technology have yet to arrive, and I remember waiting like hell for MMX encoded applications, which I eventually forgot about.

In tomshardware's review, the 2.8 P4 won 14 out of 30 tests. In <A HREF="http://www.anandtech.com/showdoc.html?i=1718&p=6" target="_new">Anand's</A> test set, the 2.8 won 11 out of 15 benchmark sets. Superior? Hardly.
Again, the same argument could be made about why buy a 3.06 instead of a 2.8 P4. The price difference is great for what, to some, too little a gain. But then again, people running a 3dsmax design studio may like it a lot.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

ozmol · Feb 12, 2003

Once again Tom's has pitted P4 Rd-ram against Athlon DDR, and the P4 DDR results are nowhere in sight. I'm having trouble believing that I just read four pages of posts in this thread and no-one has commented about this.

I've said it before and I'll say it again. The vast majority of P4 buyers are buying DDR systems (I'm a system builder/retailer in Australia). P4 buyers who read an article like this and then go out and buy a P4 DDR system thinking that they're are buying the superior CPU are being misled; albeit by their own lack of understanding of the situation.

Having said that, I agree it's a shame that the new Barton has been given the XP3000+ tag which it clearly doesn't deserve, but this doesn't change the fact that AMD's performance rating system is still very sound in general when comparing Athlon DDR with P4 DDR. In fact, in most cases I'd say AMD have been kind to Intel in this regard.

juin · Feb 12, 2003

We pit fast cpu again each others so the same goes for the platform

Just next to the lab and the bunker you will find the marketing departement.

AMD XP 3000+ Benchmarks

Champion

Distinguished

Distinguished

Champion

Splendid

Distinguished

Distinguished

Splendid

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Champion

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Splendid

Distinguished

Distinguished

Distinguished

Share this page