Just one simple question...

Did you use 1x512MB RAM, or 2x256MB?

Atleast nForce2 wont use dualbank features, unless you have 2 RAM "modules" inserted in the right slots...

test

I'm impresed. The 1.6GHz Athlon 64 pre-production sample shows some definite strengths. However, considering the imminent projected speed increases for Barton (3000+ and 3200+ at 400MHz memory) and Northwood (3200 at 800 MHz memory), it is clear that AMD needs to debut this processor at more than 2.0Ghz to be competitive. With such a 25% processor speed increase over the 1.6GHz model, and assuming linear performance scaling with processor speed, a different performance picture would emerge.
1. PCMark CPU - 2nd place and just ahead of Barton 2800+
2. Business Winstone - 1st place and way ahead of Barton 2800+
3. Sysmark - 2nd place just behind P4 2.8, and ahead of P4 2.53, Barton 2800+
4. MP3 encoding - 2nd place, with Barton 2800+, ahead of P4 2.53
5. WinRAR - way further ahead in first place.
6. MPEG4 - 1st place ahead of P4 2.8, Barton 2800+
7. WME - 2nd place, ahead of Barton 2800+, P4 2.53
8. Sciencemark - 1st or 2nd place...
9. Lightwave Raytrace - 1st place ahead of P4 2.8, 2.53
10. Lightwave Sunset - 2nd place with Barton 2800+, behind P4 2.8, and ahead of P4 2.53

Can AMD have the FP unit run asynchronously at a faster clock, since it is a major performance module? Lets say 25% faster than the rest of the processor? (ratio 5/4) With such a change the Athlon 64 would have dominated more of the benchmarks and more meritted its 2800+ rating.

AMD perhaps needs to review its rating system for this processor. It probably would more realistically be a 2500+ to 2600+ and a 2.0 GHz would be a 3000+ The current Barton 3000+ should probably be rated no higher than 2800+, and the 2800+ no higher than a 2500+ to 2600+. The 2500+ is more of a 2300+.

"Unfortunately, 64bit operation systems and applications supporting x86-64 are not available yet."

Never heard of Linux???

It would have been nice to see AMD make the Opteron able to run at a higher frequency in 32 bit apps.

I think your missing the point

The point is that Intels plan is to leave AMD with 5 to 7 % of the personal PC market.
What Intel needs is AMD to make the least amount of money they can and still stay in business. Just like Bill Gates helping Apple the fear is that the company will be left alone in the market and risk being split up.

Intel is getting closer to reaching there release goals which is to be one major release ahead of AMD, which will give them the ability to get those few more percentage points of market .

AMD has a strong desire to get in the server market and this 64 bit processor will fall far short of that and that's going to hurt. In fact bankruptcy concerns are beginning to pop up . If they really get into trouble Intel will slow down or find other ways to keep them in business.

The Intel 64 bit processor which will be originally released almost exclusivily for the high end server business market will operate at a substantially higher rate then AMD , 2 to 3 times higher .
AMD may be able to bridge the 32 to 64 gap for consumers (the game players, these are AMD's staunchest allies ) which will help bring personal users in to the 64 bit field. and this may help them keep above water. But for now and the future things are very tough for AMD . This news is bringing some AMD stock drops.
The question is we know there is an AMD P4 but will there be AMD Pafter

Not so good...

That's promising results for a cpu running at only 1.6ghz.

Isn't Athlon64 supposed to have 256KB cache? Or this is some more deluxe version...

this is not going to go well for amd.

its a revamped xp core. with more l1 and a memory controler....
?

did i miss somthing? i really dont see any innovation here. its just a logical step for amd to migrate to a on board memory controler to feed thier chip.

thats not innovation, its just a ploy to try to squezz that last bit of performance from a aging core. the xp chip (which is what this thing really is) is losing ground. it cant scale to much higher and greater memory bandwidth doesnt do much good, the chips desgin pretty much prevents that.

it was desgined from the get go to use as little as possible as much as possible and it has finally caught up with them.

yes its fast in dbase and memory intense tasks. but so is alot of other chips. and frankly none of them are that great either.

optron / xp64 is going to be at best a mediocre addition to the faimily unless they can do somthing to DRAMATICLY increase the execution rate. a high ipc is nice, but at some point it begins cutting your own throat when the returns on being a low clocked chip come calling.

example is barton chips, thier rated fatser for lower core speeds, but cycle intense apps lose to thier older higher clocked brothers...its going to be the same here. its just a xp core with hypertransport and alot of memory. im not impressed. this thing was comapred to old school xp, not even a barton and it lost alot of ground.

sure its games good...but its still a 2.8ghz rating..and i dont sit here all day making zip files....

what market is this thing for?

2 years for what i wonder. i have a huge suspicion that amd tried to bite of too much and may have made a whole different chip that failed or was inferior.....why you ask...cause this xp64chip simply feels "rushed" it doesnt feel like a real chip amd would make, like when the k7 first appeared. it doent feel new...it feels......well desperate... like a backup plan

and given that this thing was supposed to be a "huge revolution" i feel VERY let down.

sure it may run 20-30% faster in 64 bit, but i have to point out its not a real 64 but chip. its got the plumbing, but its just added registers, address units etc....yamhill has had those for 2 years now. but they never implemented them for the same reason. its complex and hard to produce and wont scale very well. you have to debug 2 levels of chip now....thats a tough thing to do.

secondly, this is a windows world weather i like it or not. untill dell/hp starts shipping windows xp64 as a standard os.....noones ever going to care or understand the *possible* benifits of 64bit computing.

nix users will see more benifit from this core than anyone else, and thats fine. more power to you.

but if the success of this chip is based on corporate adoptation and nix users...this is going to be a abysmal failure captital wise...it will take years to recover. and if they stict to thier curreent statagy, amd will simply turn into another apple. and good but obsucre nitch product

i like amd, but this was a shocking failure.

Using 64-bit mode has more advantages then just 64-bit number crunching and increased addresss space. In 64-bit mode the additional 8 GPR are available. This should provide a good perforance boost.

ROLFMAO!

"During Athlon’s architecture life-time Intel managed to shift from Pentium III architecture to absolutely new Pentium 4 architecture, and then enhanced it significantly by increasing the L2 cache size and speeding up the system bus. Athlon also underwent certain enhancements, although, they never were that drastic. The most AMD did, included only system bus overclocking, L2 cache size increase, implementation of SSE support."

Does this guy read what he writes? Comparing what he says happened to the P4 and Athlon, the athlon had 1 more "enhancement" than the P4. Yet he says that it is not as drastic.

Try to be a little less ambiguous!

If there would be so many 64 bit optimised application and working, famous OS amd wouldn't do a 32/64 bit processor...
Of course this test focused on 32 bit applications only.

Just a simple remark on results...

After reading Aceshardware review of opteron it appears that they have tested Opteron with dual DDR333 memory. So the memory bandwith they got are much closer to Xeon ones (even higher in their simulated mono-processor test)
Tom's review dispite what he says appears to have used single channel DDR as they have numbers that are quite close to yours.

I would be really surprised if the final version of athlon 64 get to the market without the dual DDR controller as all the top performing platforms benefit from heavy memory bandwith.

Correct me if I'm wrong but all tests can be somewhat RAM bandwidth limited (via limitation of cache pre-fetching for example) ???

Just wanted to leave comments if this will let me without registering.

It seems like the power of this review was only what you could glean about the CPU's intended applications, and its relative lacks and strengths therein. Putting the Hammer on a gaming test, or MS Office content creation test is like using a backhoe to do your garden-weaseling work. It just doesn't work right. And while in some instances it'll perform well, it was meant for other purposes. I'm not big into chip names, and roadmaps, but (I'm sure you'll all correct me if I'm wrong here) we have the Barton to place in those tests now, and ClawHammer when it comes out in the future.

What would be a more applicable test of this chip? Firstly, there are a couple of 64 bit operating systems that might (maybe?) better leverage the design of the chip. There are also a couple of applications which would be better suited for its intended environment. Think corporate. I know AMD is.

Thanks for your reading.
Daniel

ummm here's a better one, VIA's solution is not a via ible one since, a major part of the the design is to get control of the memory to the CPU and get rid of the northbridge, and VIA put's in a northbridge limiting the frontside bus, and well as creating a bottle neck that we had finnally gotten rid of! Though starting it at 1.6Ghz is tooo slow.

Please, please, please guys.

When is someone going to overclock one of these babies?? If not the athlon64 what abt overclocking a Opteron? This surly is the question enthusiasts want to see answered. I want to see the performance gains which will be atained from a higher clock speed!!!

You have the tools, and probably the financial backing to pull this off. Please do it!!

Thanxs
Mongoled

Anandtech has done a similar comparison review of the Opteron 1.8GHz (versus 1.8Ghz Barton, 3000+Barton, 1.8GHz Athlon XP, and Pentium 4 3.0)\

Check it out guys... Evaluating "Desktop" Performance likely to be seen on Athlon 64.

http://anandtech.com/printarticle.html?i=1818

ctually, that article makes my point. Barton at 1.8GHz is not noticeably faster than Athlon XP at 1.80Ghz for most things. We're glad to have the improvements, but I'm not sure they merit calling a Barton at the speed of a 2700+ XP, (2.167GHz) a 3000+ processor or a Barton at the speed of a 2600+ XP (2.083GHz) a 2800+ processor.

AMD made a slight mistake. Hope they correct these mistakes over the next few releases...