Slowest Athlon 64 humbles fastest P4 in gaming

zerodamage

"GERMAN HARDWARE Web site Hard Tecs 4U has published one of the first reviews of the AMD Athlon 64 3000+ processor. This new model runs at the same 2 GHz frequency as the flagship product, but does so with 512 KB of level two cache, which is half of what the 3200+ model has. Cheapest Price Watch price has it for $214, which is $156 less than the $370 3.2 GHz P4. The Athlon 64 won eight of the nine benchmarks, and one of them by 27%. For those who need superior gaming performance than a 3.2 GHz P4, but at less cost, these benchmarks indicate that the Athlon 64 3000+ is the way to go.

It was a smart move on AMD's part to release this device at the same frequency as its sibling, but with half the level two cache, as it still betters the P4 in the very important world of gaming. AMD deserves some credit here. It has brought to market a processor that has half the theoretical memory bandwidth and 37.5% less frequency than the 3.2 GHz P4, and it is still a gaming winner. For those that have been waiting for the cheaper Athlon 64s to arrive, you can now have your cake and eat it"

Source: Neowin

craig588

WOW!!!!!!

spark rain fire

but yet the kids still buy intels, because the clock speeds are higher.

__________________

----
[color=red]http://www.invalidsystemdisk.vze.com[/color]

craig588

the kids? What kids have more than 50$?

people bought intels because they were faster. Now its not the case. (dont delude yourself, an overclocked intel is faster than an overclocked barton)

The_Neon_Cowboy

hah hahaaa i been telling the intel fanboys this for quite a while lol

mhz isn't everything it comes down to design and performance wich amd has the advantage

__________________

@/2ct!<

I am glad to see what relatively small impact overall the cache has, ie the next gen 754 newcastles with 512 are still gonna be pretty attractive.

__________________

Markius

from the looks of things it wasnt the P4 EE that was used...

but i agree with the general message they are saying...

my next cpu will be an AMD 64 because its great performance wise and doesnt cost an arm and a leg like the way over priced P4EE

__________________
DriverHeavens Aussie...

"Lifes tragedy is that we get old too soon and wise too late"

Optimummind

The reason I prefer to buy Intel is similar to why I prefer to buy nVIDIA GPUs. It's not about how fast the processor or GPU is but rather how well it integrates with the rest of the system and how stable it is.

For CPUs, I want stability in my chipset. For GPUs, I want stability in my drivers and compatibility with old and new games.

Intel is known throughout the world for its highly stable chipset and mobo architecture. On the AMD side, however, there are frequent chipset-related issues from VIA, SIS, and nVIDIA. Of course, for the last generation of AMD CPUs, the nFORCE have been very stable and reliable. But, it still did have up and downs. As for Intel, I haven't heard any negative reports about its chipset that is flaking out.

For the GPUs, I prefer nVIDIA because of their awesome drivers. ATI, although much, much improved, is still not at the same or above level as nVIDIA. If you want to know just how many driver-related issues plague ATI and games, go to the Rage3D forum and look for the known incompatibility thread.

Although it is still too early to how good the chipsets for the Athlon64 series will turn out to be, I will continue to buy Intel CPUs for their stability and reliability.

AlternateVirus

you do realize only 1% of users who have ati cards go to rage3d forums? I'm sorry but if you go to any nvidia forums, you'll see there a huge shitload of same threads about problems related to the nivida drivers. nvidia isnt perfect at all. same with amd chipsets, you basing everything because you hear it online but you never hear shit from system intergrators or anybody using the same motherboards. 90% of the problems are caused by n00bs.

AMD is as stable as Intel processors and ATi is as stable as Nvidia. theres nothing you can prove to me that tells me otherwise.
Intel great for overall performance, AMD great for budget and gaming. Period
ATi great for overall performance and Nvidia is great for basic gaming that most users only need.

nForcer

But sit back and ask yourself 'does one REALLY care about benchmarks' I direct that question to the open minded individuals. Those who replay 'yes i do care' are not open minded. I've said it before and I'll say it again, I'm a die hard Intel fan but AMD is giving Intel competition it has never seen - which is a good thing. Problem is, the two chip giants are growing in different directions. AMD is pushing 64-bit technology and Intel is working to better 32-bit. Each have thier justification but in the end, they both get the job done. To most, including myself...thats all that really matters. (I'd like to thank the Academy, my parents, .... )

craig588

64 bit is the future because, well, its...... um....bigger.


seriously, in the future 64 bit will be very important for addressing huge amounts of ram/HD space, not to mention the fact that twice as many instructions can be fit into the same space.(sorta)

in games we'll be seeing huge textures that probbably are 80-90 megs each, and each scene will probably have a few dozen of them. We all will have like 128GB of ram and 4-5TB HDs. we need 64 bit to address all this info. (and if you dont belive me think of what gates said saying we would never need more than 640K of ram, if you want a more recent one think of a dual 12M voodoo 2 setup and everyone said 24M of ram was crazy, even more recently dell released a custom GF2 with 64M of ram and people said they were crazy)

-={420}=-

Yeah I was thinking since Longhorn is gonna be 2006+ by then they'll probably only make it for 64-bit processors...I doubt they're gonna waste that time to make a 32-bit version.

__________________

mainman

Dont think so bud, theres still people using win98 in our days so, im sure they will make a 32bit edition when longhorns out.

__________________

@/2ct!<

Even Intel plans to be 64 bit mass market by then though, so most likely 64 bit only OS.

__________________

WaltC

I have to say I disagree--you're wrong on both counts... I dumped Intel a couple of years ago when they pulled the WilliametteP4/Rambus fiasco, and although I initially felt at the time as you relate above, I've been pleasantly surprised to have had few, if any, reliability issues which have exceeded the reliabiity issues I had previously with Intel (When I used mostly Intel and Intel-chipset Asus motherboards.) In short, I've experienced nothing in the last couple/three years that has at any time caused me to consider going back to Intel--even once. OTOH, I've been pleasantly surprised by several things by going with AMD. Intel hasn't made anything in a long while that would cause me to switch back. In sept. '02 I owned two GF4 Ti4600's in my home machines, and at that time one was displaced by a R9700P--which frankly was as stable as anything I owned from nVidia--and which I liked so much that in May of '03 I replaced the other GF4 with a 9800P. Ditto there, too--nVidia's given me no reason to come back on the product front, and ATi has given me plenty of reasons to stay with them. So I'm afraid there are many who'd disagree with you on this topic. Good advice is not to rely on gossip and hearsay--but find out these things for yourself. Like you, I was a bit over cautious when dumping both these companies--but I've never regretted it since.

And my goodness, if you want to take a gander at people having trouble with Intel cpus and chipsets, you should check out the forums at www.intel.com ! Heh... You'll run across *thousands* who are having problems. Just like the dozens who report problems from time to time at Rage3d. The thing to remember, though, is that it's only a tiny fraction of the people having problems with these products who actually post to these forums, and you'll never hear from the many, many more who have no problems--because of course they don't post. Many more people have few problems with Intel cpus and chipsets than have a lot of problems with them, but you should give yourself some perspective and check those Intel forums out... There are several Intel-product newsgroups, too. You can try comp.sys.intel for starters--just do a newsgroup search for "intel" through your browser, and you'll find literally tens of thousands of posts by people having their share of Intel problems in various areas.

No_Style

And I <3 My AMD. They're both great processors.

__________________

mainman

Theres still a long way to go so lets just wait and see.

__________________

DriveEuro

Yes, I've fallen Intel. By the end of the week I'll have a P4P800 with a 2.4c. But let me show you all this. After I spend my $205 (yes, I found a HELLUVA deal) all you people with 2500s and NF7s will crumble to intel. You people need to realize that the XP chips get raped by the 800fsb P4s. It's only the 64bit chips, which most of us can't afford that win.

__________________
[color=orange]ASUS P4P800 Deluxe [color=black]-[/color] 2.4C --> 3300mhz [color=black]-[/color] Mushkin PC3200 at 220 5-2-2-2 (cpu limited) [color=black]-[/color] BBAti 9800Pro at 430/375 [/color]
[color=black]3[/color][color=gray]D[/color]M[color=black]a[/color][color=gray]r[/color]k[color=yellow]01[/color] [color=black]3[/color][color=gray]D[/color]M[color=black]a[/color][color=gray]r[/color]k[color=yellow]03[/color] 1600*1200 1280*1024

Optimummind

Yeah? And how exactly did you get that figure and how will you substantiate your statement?

In keeping with that kind of attitude, then I can say the same thing. There's nothing you can prove to me that tells me otherwise. Dumbass, get a brain.

Optimummind

The issues I have mentioned in my first post regarding ATI driver problems are not about isolated and random errors that users experience but rather documented and official errors. The kind of errors that ATI knows about and is working on a fix in their continual set of drivers.

Here is the thread at Rage3D where real ATI engineers go to check out user-reported errors with games. As you can see, these aren't isolated and random events. Most are acknowledged errors that ATI engineers are working on to fix...eventually.

BWX

"Watch price has it for $214, which is $156 less than the $370 3.2 GHz P4."

__________________

FluffyChicken

Not particularly a good selection of games used....
but for a better review (unfortunatly without the 64-3000+)
includes games that are not fsp
http://www.aceshardware.com/read.jsp?id=60000262

__________________

WaltC

I think the term you are looking for is "bug fixing." I'm not sure where you get the idea that ATi has more "bugs" to address than nVidia in its drivers. If you have such an idea, I'd really advise you to discard it because it simply isn't true--not in the least. When I stopped using nVidia products more than a year ago, I was using the 30.82's, because even though the 40.xx Dets had been released, and I had tried them, I had problems with them in certain games and had to go back to the 30.82's. In fact, all year long, with each succeeding 40.xx Detonator release, I read post after post after post from nVidia users who did exactly what I had to do--go back to the 30.82's because many of the newer drivers, riddled with "optimizations" and other things I could use less friendly terminology to describe, were simply too unstable for daily use for them.

Here's what you probably don't understand about "bug fixing"....(as I notice you erroneously use the term "errors"):

All 3d game engines are different from each other in various fundamental and substantial ways, even though they may support the same API's. In fact, games which make use of the same game engine--for instance, Q3-engine games or UT2K3-engine games, often do some very different things. The APIs allow for a wide range of approaches in software, and there is no "narrowly defined set of rules" that developers must follow to be API-compliant in the sense you apparently think. Here's an example of what 3d-card IHVs have to do:


(1) They release a set of drivers in January, a new set of Forcenators or Catalysts let's say
(2) In Februrary a new 3d game is released in which the developers did something slightly out of the ordinary that nobody else has done in exactly the same way. And in certain hardware configurations, under certain conditions, an IHV's driver--released a month earlier-- doesn't function as intended. The IHV must then look at the situation and make changes in his drivers so that they will handle the situation.
(3) In March, another developer with a popular game already shipping issues a game patch which makes some fundamental improvements and changes to its 3d-engine in the game, changes which cause problems for certain hardware configurations, under certain conditions, for an existing IHV's driver set. The IHV should then look at the situation and issue a code change, if applicable, which will solve the problem for his customers who are experiencing it
(4) Conditions 2 & 3 above describe what is *continuously* occurring in the 3d-gaming marketplace--those conditions never end, and are always ongoing, in other words.

These conditions apply no less to nVidia than they do for ATi. The difference in the current situation between the two companies as regards driver development is that while ATi has stepped up its official driver release schedule to reflect increased support for a growing and expanding 3d-gaming marketplace, nVidia has actually complained about its current support level of quarterly driver releases, and has even made the ludicrous suggestion in public that it might drop back its support to a single official driver release per year. Well, conditions 2 & 3 as I've listed above don't really care what nVidia decides to do, they are going to exist regardless of nVidia's official driver release schedule. Simply put, if nVidia was to actually drop back to a single driver release per year they would soon be out of the 3d-gaming gpu reference design business, because their products would quickly become useless to a wide number of 3d gamers who might buy them. For the record, should nVidia ever actually *do* what it said it might do, and drop back to a single driver per year, you can take it for granted that they are withdrawing from the 3d-gaming gpu reference design markets and will no longer compete in them.

So why doesn't nVidia *want* to emulate ATi and provide better support for its customers? Simple: the kind of driver support ATi is currently providing its customers *costs a lot of money* to provide. It's an ongoing, resource-intensive process, and what ATi did when it stepped up its official driver release schedule to one every month was to apply pressure to nVidia's quarterly release schedule, making it suffer by comparison. What did nVidia do? IMO, the "one driver a year" remark was nVidia arrogantly thumbing its nose at ATi and saying that it doesn't matter how good ATi's driver support gets, or how bad nVidia's gets, ATi won't affect their business. Heh... It's classical delusional thinking on the part of nVidia, IMO, but I suppose that ATi should be gratified that someone high-up in nVidia is closely wedded to the crack pipe... nVidia is quite the Humpty-Dumpty in this case, I fear.

Last--yes, you can visit a number of forums around the Internet and you will find ATi engineers and driver developers often candidly discussing their products and any problems they may have as well as future directions they might take, on a *personal* basis. You will *not* find this true of their counterparts at nVidia, however. Apparently, nVidia has an internal policy such that if you are a technical person involved in either driver or hardware development you may not talk about your work in any public forum. That's what I think, anyway, as the only people from nVidia who do any talking publicly are nVidia PR people, mainly Burke, Perez, or Kirk. Generally, too, also very much unlike ATi, the only talking nVidia PR people do is in controlled interviews--you won't find them out in forums engaging their customers in dialogue and discussion. nVidia employees just don't do that, and I think it is because they aren't allowed to do it as a condition of their employment.

I think there's a very good reason for that, though, that helps us make some sense out of the very different approaches to so many things between the companies. ATi clearly seems to understand that when customers buy an ATi-based product from one of its AIB partners, that those people are their customers, even though in a slightly indirect fashion. Also, ATi sells it's own ATi-branded products directly to end users, and of course understands that the individuals who buy them are their direct customers.

nVidia, though, does not directly sell *anything* to end users, and over the years nVidia has apparently reached the conclusion that its customers are strictly its AIB partners (those few still remaining who have not also yet picked up ATi at this time), and that end users like you and me should be addressed by nVidia's AIB partners, and not nVidia itself. And that's why I think you don't see nVidia employees in the forums like you see some key ATi personnel in them--nVidia simply doesn't see the end users who visit those forums as its customers. You and I don't count, when it comes to nVidia, as persons worthy enough to engage in dialogue--nVidia's stance seems to be that we are merely sheep to be led by the appropriate level of public relations propaganda. So that's why, in my opinion, nVidia does not engage in discussion in public forums--it simply does not see that it has any customers in those places.

When you really take a look at the errors in judgment nVidia has been making for well over the last year now, their current problems in relation to their competitors become all too understandable. This is the way I see the situation you've described. The key cause of business failure is "mismanagement," which is a polite business acronym for "faulty judgment at the top"...and the corporate graveyards are littered with the tombstones of once high-flying companies who mismanaged themselves into oblivion. Also, some companies simply can't function in a competitive environment--3dfx had a very hard time with that--and up to the past 14-15 months nVidia has had a pretty clear field for itself. I think it will be interesting to see whether nVidia can be as successful in identifying its internal faults and fixing them as we've seen happen with ATi for the last year. Next year should be very interesting...

zerodamage

Stay On Topic

WaltC

Heh-Heh...(sheepish grin)...I did get carried away, didn't I?

Anyway, on the topic, I did find the remark earlier in this thread that a P4EE wasn't used in the comparison to be interesting, considering that the P4EE is $1k and the A64 used in the comparison is close to being 25% of the cost of the P4EE. And that disparity is exactly what the person making the comment noticed himself. If this pans out as reliable, although it may spell trouble for Intel from a competitive standpoint, it's also very good news for us consumers of those cpus from either company, as it will hold down pricing on both. The only way for Intel to maintain such pricing for the P4EE in such an atmosphere is if yields for it are so low that it has no choice but to price it @ $1K--also, though, the P4EE pricing issue is compounded by Xeon pricing, as noted in the primary post in this thread. It's really interesting to watch the way competition works.

I can't help but breathe a sigh of relief when I think about the fact that just a few years ago the high-end x86 desktop cpu was a 600MHz PIII, and because of the lack of any competitors in that market space people were paying $1200 for them--each...

Crash Override

Personally I find it interesting that the reviewer states that the 3000+ is running at the same clock speed as the 3200+

Completely untrue. The 3000+ is a 1.8GHz CPU if you look really closely. So it wasnt just removing half the cache that made the renaming necessary


Also Optimummind I have some benchies for ya from Sandra which is a good baseline comparison tool

My non-barton 2200+ which has been slighly changed from stock. I have upped the FSB on it to 166 but dropped the multiplier on it so its still pretty much at a stock speed. It is running on a Via KT400 motherboard with PC2700 Corsair XMS.

My Arithmetic scores were as follows:

Dhrystone ALU: 6962 MIPS
Whetstone FPU: 2885 MFLOPS

For comparison I took the baseline for what a stock 2200+ should be at:

Dhrystone ALU: 6876 MIPS
Whetstone FPU: 2885 MFLOPS

P4-C 3.2GHz w/800MHz FSB:

Dhrystone ALU: 7869 MIPS
Whetstone FPU: 2365 MFLOPS

Now I am comparing the FPU to the FPU here. I was not adding in SSE2 instruction as my 2200+ does not have them. SSE2 on the P4-C 3.2GHz scores 4325 MFLOPS.

That does give it a fairly impressive lead over my lowly almost half the speed CPU in the Whetstone portion of the Arithmetic test. My CPU also has a lower FSB than the P4-C which also helps explain the difference.

So lets see what happens when we test out a barton 3200+ vs that same P4-C 3.2GHz.

Barton 3200+:

Dhrystone ALU: 8404 MIPS
Whetstone FPU: 3465 MFLOPS

The SSE2 speed for the P4 is already listed above so no need to post it again. The Barton also does not contain the SSE2 instructions which were reserved for the Opteron and the Athlon 64.

Now onto the multimedia part of the test.

My Athlon XP 2200+:

Integer aEMMX/aSSE: 16821 it/s
Float aSSE: 17241 it/s

P4-C 3.2 GHz:

Integer iSSE2: 18653 it/s
Float iSSE2: 21969 it/s

In this test it seems that Intel has pulled farther ahead of my little CPU. Lets see how it holds up to the Barton 3200+

Barton 3200+:

Integer aEMMX/aSSE: 20181 it/s
Float aSSE: 20449 it/s

Whoa, wait a minute. What just happened to Intels lead?

It seems in the floating point test they are just barely floating by on their merits but in the Integer tests both Arithmetic and MultiMedia its getting its ass handed to it.

The Barton 3200+ is only a 2.2GHz part though.

If clock speed is all that matters then why is Intel not winning this one?

Also if you want to get that speed back youd have to buy an EE or overclock the P4-C. So at stock Intel is losing.

Anyone with any benchies from a 3.2c who wanna post their results if they contradict mine feel free so we can see what others score on those tests so we have a baseline for comparison

DriveEuro

But you're stupid if you buy a 3.2ghz P4. They don't overclock high on the FSB. You need a 2.4C or a 2.6C cuz they can pull 3.6ghz+ with a 600mhz FSB.

__________________
[color=orange]ASUS P4P800 Deluxe [color=black]-[/color] 2.4C --> 3300mhz [color=black]-[/color] Mushkin PC3200 at 220 5-2-2-2 (cpu limited) [color=black]-[/color] BBAti 9800Pro at 430/375 [/color]
[color=black]3[/color][color=gray]D[/color]M[color=black]a[/color][color=gray]r[/color]k[color=yellow]01[/color] [color=black]3[/color][color=gray]D[/color]M[color=black]a[/color][color=gray]r[/color]k[color=yellow]03[/color] 1600*1200 1280*1024

Optimummind

LOL, you're funny and stupid.

First of all, I did not state in any of my posts above that I choose Intel because they have a faster clock rate or because Intel has higher benchmarks. You don't believe me? Scroll up and see for yourself if you can find me mentioning that Intel is better b/c of higher benchmarks or Intel is better b/c they have higher gigahertz ratings. Can't find it, can you?

Why, then, did you even bother to post those?? Your brain can't, perhaps, stay focused on the topic??

I'll state what I said before again just for you, Crash: I choose Intel because of its stability and compatibility.

Hopefully, you won't be mucking that up as well.

Optimummind

One more time, Crash, here is what I originally said. Don't forget, ok?

BTW, I'm not going to trust any benchmarks run by individuals like you. I'm gonna trust sources like Maximum PC, Computer Power User, etc. These guys know the complexities of running benchmarks by carefully factoring in all the variables and statistics. I don't think YOU or most people who run benchmarks at home are fully aware of the big picture.

BWX

Yeah that's true. But you know as well as I that most people are. LOL

__________________