“After reading the round-by-round account of our dual-core desktop CPU prizefight, it should come as no shock that AMD’s Athlon 64 X2 chips are the runaway victors here, laying out the Intel Pentium D and Pentium Extreme Edition 840 chips pins up. If we had to call out one chip, AMD’s Athlon 64 X2 4400+ is an outstanding bargain given the competition, but as our results show, any AMD dual-core CPU will serve you better than its similarly priced Intel equivalent.”
I doesn’t state what speeds the intel chips are.. are they using 3ghz p4 ext vs a 4400+?
Pentium Extreme Edition 840 is 3.2 ghz,800 FSB, and supports Hyper-Threading.
They didn’t explicitly mention any clock frequencies, but that doesn’t mean they’re unknown.
Here are the frequencies of the D8x0 CPUs:
http://www.intel.com/products/processor_number/index_view_pd.htm
Here are the frequencies of the X2 CPUs:
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_94…
Does it matters?
They also compared price. The comparison is about perfomance/$ and for me doesn’t matters the speed.
Intel Pentium EE – 1000 $
AMD Athlon 64 X2 4800 – 800 $
200 $ of diference and AMD’s still kicks butts !!!
It doesn’t clearly state which chips are they testing. Extreme edition with 4400+? Because no matter what, I will not pay that much for a processor. Maximum in my book is 300-350 eur. Is really enough for all the tasks I do. Anf for that price I can get quite nice cpu’s Intel or AMD.
All dual-core Pentiums are LGA775 and therefore do not have pins, so they can’t be “pins up”
On the lower end, the cheapest Pentium D I can find costs about 240,- Euros, where the cheapest Athlon X2 costs about 100,- Euros more than that. Am I missing something?
Edit:
I hate that all these reviews always assume you’re only interested in the high end. I don’t feel like spending a grand on a CPU, and there are barely any comparisons available on the lower end.
Edited 2005-11-28 15:11
> On the lower end, the cheapest Pentium D I can find costs
> about 240,- Euros, where the cheapest Athlon X2 costs about
> 100,- Euros more than that.
That’s right, but the Athlon 64 3700+ costs about the same and beats the D hands down in almost all tests.
That’s right, but the Athlon 64 3700+ costs about the same and beats the D hands down in almost all tests.
How about apps that fully support multithreading?
How about apps that fully support multithreading?
This is an interesting comment… I’m having trouble reading the motive. I can’t tell if it’s simply to point out Intel’s marketing hype about hyper-threading, or if it’s a genuine question.
This is an interesting comment… I’m having trouble reading the motive. I can’t tell if it’s simply to point out Intel’s marketing hype about hyper-threading, or if it’s a genuine question.
It’s a genuine question, because I am currently looking for a new system that will run mostly multithreaded software. I couldn’t care less if a single-core Ahtlon smokes a Pentium D in Quake3, what I’m interested in would be rather Cinebench or make -j2 compiles.
> I hate that all these reviews always assume you’re only
> interested in the high end. I don’t feel like spending
> a grand on a CPU, and there are barely any comparisons
> available on the lower end.
Sure there are! Those comparisons were done when those CPUs were the best ones on the market. 😉
There are also many reviews that focus on current budget models.
Sure there are! Those comparisons were done when those CPUs were the best ones on the market. 😉
There are also many reviews that focus on current budget models.
Of course there are some ; just not enough. And not only about CPU reviews : graphics cards, RAM and such mostly review high-end hardware. And that just makes sense : hardware makers need to show they’re the best, and sending low- or mid-ranged hardware for review isn’t the best way to get the hardware crown, IMO.
We need a http://www.hardwarereviewsforthecommonman.com
http://www.anandtech.com/
They do low,middle and high range comparative review every month or so…
Intel has been trailing AMD in performance ever since the original Athlon was released.
It’s why the P4 was created. They didn’t have a chance of reclaiming that performance crown, but at least they could capture the mhz crown.
That’s why the P6 core is coming back in the form of pentium M’s.
Not really. The Athlon was able to catch up and beat the Pentium III because Intel was already focusing on the Pentium4. Unfortunately, they realized too late that NetBurst was a failure, but they couldn’t just drop the product line.
You are right that this is the reason the P6 is coming back. It’s a much more solid design that will be much better than the pentium4, and maybe even the newer Athlons.
I don’t know about the D8x0 CPUs, but the X2s have already developed a reputation for their amazing overclocking potential. Most X2 cores are able to reach about 2.7GHz with just fancy RAM. This means that buying cheaper X2s with good memory will get you the same speeds as much more expensive processors. I think that most of the cores are identical, since they seem to only get to 2.7GHz before having stability problems. AMD currently offers two kinds of chips: one with 512KB cache per core and one with 1meg.
I would recommend buying a 3800+ to anybody who wants to run dual core chips. Unfortunately, there is no 4000+ X2 (which would be the 1meg cache version of the 3800+), so if you want that kind of cache, you have to get a 4400+. Once they are overclocked, you get better performance than two FX-55 chips and, if your programs are multithreaded, you will likely best the FX-57, too. The X2s are the most exciting things in the enthusiast market since the 300As or the Barton cores.
How is it on the Intel front?
I’ve heard that Dell may start using AMD chips. I believe the reason is Apple. By Dell selling AMD chips, they will be able to market the higher CPU horsepower of the Dells over the lower CPU powered Intel Apple computers.
just my 2 cents…
– Mark
“I’ve heard that Dell may start using AMD chips.”
This has been a rumor and source of speculation for at least five years. I will have to see the Dell press release stating their intention to believe it at this point.
Dell has used AMD as leverage in price negotiations with Intel over the years, but I wouldn’t jump to the conclusion it was anything more than bargaining.
AMD may also be somewhat reluctant to “dump” a huge quanity of their processors into the hands of someone like Dell who are going to demand discounts of untold proportions. AMD may well be selling all of their processors without discounting them.
The wamperdyne multi-core chip makes a lot of sense in the server context. But have we not seen stories about forthcoming chips with integrated DRM? Seems that, irrespective of the server, slower chips will target the desktop possibly stifle fair use. If Intel smells more money in DRM, they may not be too interested in the multi-core server chip shootout.
the only thing I don’t like about AMD’s is having to use Nvidia’s closed source MB chipsets. Their onboard devices all require closed source drivers.
Then use VIA. What’s the problem?
(VIA releases all their specs
and is in general opensource friendly)
That’s not true, I’m using the a Biostar nForce4 SLI motherboard, and it works great with the forcedeth driver for ethernet, open source driver (in the 2.6 kernels) and the sound works well enough with the generic i8x0 sound driver. SATA support works as well, so I don’t know what you’re really complaining about, I never have to download those drivers to get my motherboard to work.
> the only thing I don’t like about AMD’s is having to use > Nvidia’s closed source MB chipsets. Their onboard
> devices all require closed source drivers.
Heh? Ever heard of VIA? :/
Then use a MB with a chipset from Via or SiS. You don’t have to use Nforce just because it’s the best one
(seriously, the differences aren’t that huge)
The memory controller being on the CPU makes chipset preformance much closer for AMD than in the past. But still, I always advise people to look at the products benchmarks before buy it.
Here’s some other dual core benchmarks which test other applications such as Mental Ray.
AMD Athlon 64 X2 vs Intel Pentium D 840:
http://www.gamepc.com/labs/view_content.asp?id=a64x2&page=1
AMD Opteron vs Intel Xeon Paxville:
http://www.gamepc.com/labs/view_content.asp?id=paxville&page=1
its sad.. intel is much larger and has more resources than amd.. why on earth would intel let amd do this?
“its sad.. intel is much larger and has more resources than amd.. why on earth would intel let amd do this?”
Intel made some seriously bad choices over the past few years. The first one not paying attention to what the competition was doing.
AMD simply grew up……
in the “bang for your buck” part. Out of 100 performance points (X2 4800+ = 100), the Intel D820 scores 77, meaning that even if the D830 and X2 3800+ scored the full 100 points (which they don’t), due to their price they would still lose to the D820 in the absolute bang/buck department. Higher models just get worse bang/buck ratios.
Test 7 should have gone to Intel, given that their question was “Which CPU has the most juice per dollar?”
Not that a 6-1 loss is anything to boast about…
In all cases they’re basing their comparisons on top performance. That may not be the best metric for you, but it doesn’t mean they’re cheating. Top performance is certainly a concern for professionals where a one-time cost can easily be regained in increased productivity.
There are three dual core processors from AMD that achieved a better score than the best performing (by CNET’s metrics) Intel processor, and every single one of them costs less than the Intel processor. Even if you ignore the EE chip, there are still two AMD processors that outperform the best Intel chip while costing less. That seems like a pretty good argument for AMD to win the price/performance crown at least at the high end.
One last thing to note…
The prices listed are for the CPUs only. Once prices for supported motherboards and memory are taken into consideration the advantage for AMD generally tends to increase. So, even the Pentium D 820 doesn’t win by as much of a price margin as it at first seems to.
Intel should change the name of the Pentium to ‘Ricer’. All the appearance of looking fast (including the big numbers, the ludicrous fans and cooling systems and the name stickers!) but they simply get their backsides kicked to a red pulp every single time.
Pentium Extreme Edition 840
Ooooh, er. Extreme Edition eh? Oh well yer, with a name like that it must be fast, mustn’t it?!
Somebody’s going to set up a site on this sooner or later, or maybe somebody already has!
First of all they tested on crappy Nvidia chipset board instead on real Intel chipset.
Over the years as a computer tech I’ve seen it again and again that Intel has the best chipsets. None of the Via, Nvidia, or SiS (shudder) ever exhibit the same stability and reliability as Intel chipsets.
They obviously didn’t know what they were doing.
I’ve done my own benchmarks with Photoshop and Intel beats AMD easily.
You can run your own if you don’t believe it.
Here http://www.driverheaven.net/reviews/dualcoreintel/photoshop.htm
is an Intel D820 review and the benchmark script can be downloaded from there too.
I’ve stopped using AMD thanks to crappy Via hardware back in 2001 after seeing time and time again that although the AMD cpu’s are fine the chipsets are usualy not.
And even this year when working on a few AMD64 machines I’ve seen Windows XP throw error messages that I’ve never seen on Intel based hardware.
All these reviews and wild claims how AMD is so much better than Intel will not convince me since I’ve seen the opposite with my own eyes.
Besides, the D820 is a real bargain compared to even the cheapest AMD X2 and when coupled with a quality Intel chipset based mobo (like the Asus P5LD2) then you have a rock solid and super stable system.
On the other hand, competition is good. At least AMD keeps Intel prices in check.
I’m a hardware reviewer by profession … and sorry, but you’re wrong.
That whole “Intel is more reliable and stable than AMD” is a long-gone myth that came to life in the days of the early Athlon when VIA was more or less the only chipset maker. NVIDIA’s nForce chipsets, while usually suffering from one or two noticeable bugs in the first revision, are very stable and very fast in the rehash releases.
When comparing the performance of the Intel platform to anything else, you don’t want to be using anything but NVIDIA chipsets. Sorry, but that’s just the case. Intel’s chipsets might give you a feeling of greater stability, but they’re not exactly the fastest of the bunch. Last time I checked, they weren’t even in the top 3 in terms of performance (i955X vs. nForce 4 for Intel vs. some VIA one vs. an SiS one).
If you compared equivalent P4 vs. Athlon 64 systems, there’s no chance in hell that the Pentium 4 won anything but benchmarks that have been specifically optimized in the P4’s favour — unless you botched the benchmarks intentionally or due to ignorance. If you look at DriverHeaven’s Photoshop results, the Task Manager screenshots are extremely suspicious to me. Regardless of the architecture, the two CPU graphs should look more or less the same — unless one of the systems had a (dis)advantage.
“the Task Manager screenshots are extremely suspicious to me”
Yes, for one thing the OS and application footprint are anything but the same.
The Intel system has virtual memory commit of 828 MB (nearly all of which is working set) with a process count of 37. The AMD box, however, has a vm commit of 262 MB (all of which is working set) and a process count of 25. VM usage graph would seem to indicate that these system are running very different workloads. If I had to guess, the screenshots were manufactured (the Intel shot looks like a benchmark session, SiSoft to be exact, while the AMD looks like cpustress).
@ Linux Is Poo
I’m a hardware reviewer by profession …
What does that mean? That you are automaticaly the ultimate authority on hardware?
Well, I am a computer builder and technician by proffesion and my job involves a little more than hooking up a mobo and running a few benchmarks for 5 minutes.
I’m responsible for computers that are meant to be running stably and reliably 7-24 all year long.
So forgive me if I don’t immediately drop on my knees and bowe to you.
That whole “Intel is more reliable and stable than AMD” is a long-gone myth that came to life in the days of the early Athlon when VIA was more or less the only chipset maker. NVIDIA’s nForce chipsets, while usually suffering from one or two noticeable bugs in the first revision, are very stable and very fast in the rehash releases.
You prove my point exactly here.
When I deliver a system to a client’s office I can’t tell them “Oh sir, sorry, but we’ll have to wait a few month for this mobo to run stably after a few BIOS updates. So for now just hit the reset button when the box locks up to restart it and pray that your files are not corrupted when it boots up again”
I don’t have to worry about anything like that when I use Intel chipset on a quality mobo. I know I’ll get stability and a reliable system.
If you compared equivalent P4 vs. Athlon 64 systems, there’s no chance in hell that the Pentium 4 won anything but benchmarks that have been specifically optimized in the P4’s favour — unless you botched the benchmarks intentionally or due to ignorance. If you look at DriverHeaven’s Photoshop results, the Task Manager screenshots are extremely suspicious to me. Regardless of the architecture, the two CPU graphs should look more or less the same — unless one of the systems had a (dis)advantage.
Well, I don’t care about the DriverHeaven’s Photoshop benchmarks, although in this case I agree with them because I was able to verify the benchmark myself.
As a matter of fact, my P4 2.4Gz with hyperthreading scored quite a bit better than a brand new AMD64 3500+ system I was just setting up for a client.
I just don’t agree with the article and the benchmarks listed because they are so heavily skewed toward AMD and I know from my own experience that in reality the perfomance differences are a lot more balanced between the two cpu’s depending on what apps, and under what conditions you run them.
Well, I am a computer builder and technician by proffesion and my job involves a little more than hooking up a mobo and running a few benchmarks for 5 minutes.
You’re a screwdriver operator. You screw motherboards, drives, and power supplies into cases and then connect up a few cables. BFD. That you think that benchmarking systems is five minutes of work shows just how little you know about the art and science of reviewing hardware.
You prove my point exactly here. When I deliver a system to a client’s office I can’t tell them “Oh sir, sorry, but we’ll have to wait a few month for this mobo to run stably after a few BIOS updates. So for now just hit the reset button when the box locks up to restart it and pray that your files are not corrupted when it boots up again”
The only thing proven here is that you have reading comprehension problems. He didn’t mention anything about BIOSs and I’m really wondering if you know the difference between a chipset and a BIOS. The original poster was referring to a few third-party chipsets that, several years ago, had problems.
I don’t have to worry about anything like that when I use Intel chipset on a quality mobo. I know I’ll get stability and a reliable system.
Yeah, that’s why Intel had to recall the i810, i820, 915, and 925X chipsets for serious stability problems, right?
If you use an nVidia or ATI chipset on a quality motherboard for AMD CPUs, you will get at least as much stability and a whole lot more performance.
As a matter of fact, my P4 2.4Gz with hyperthreading scored quite a bit better than a brand new AMD64 3500+ system I was just setting up for a client.
An AMD64 3500+ that is properly configured will stomp a 2.4ghz P4. I recommend that you direct your clients to more experienced vendors if you aren’t seeing such results on systems that you build and/or configure.
@fmaxwell
You’re a screwdriver operator. You screw motherboards, drives, and power supplies into cases and then connect up a few cables. BFD. That you think that benchmarking systems is five minutes of work shows just how little you know about the art and science of reviewing hardware.
Well, nowadays it’s more or less true since hardware generaly work well out of the box. But building a well working system was much more difficult 5 years ago with all the cheap junky Via and Sis mobos out there. However, I can’t brag I built any systems like that because I simply avoided that kind of hardware. But I did have the displeasure of troubleshooting countless systems like that.
I fail to see how reviewing hardware is art and science. Specially the art part unless you’re trying to replace some hard data with fake one.
He didn’t mention anything about BIOSs and I’m really wondering if you know the difference between a chipset and a BIOS. The original poster was referring to a few third-party chipsets that, several years ago, had problems.
Well, if you had been past your Nintendo stage back then, you’d know that some more reputable mobo manufacturers (Asus, Abit etc) tried to fix the chipset issues with frequent BIOS updates. Good example was my own Asus A7V mobo that took serveral months and close to a half dozen BIOS updates to finally make it work stable.
AMD had for a while their own chipset and mobos but vast majority of mobos and chipsets were Via and Sis.
And since then ALL chipsets for AMD processors ARE “third-party” chipsets. Not A FEW, not SOME, but ALL are third party.
Yeah, that’s why Intel had to recall the i810, i820, 915, and 925X chipsets for serious stability problems, right?
Well, that may be, but I never used those chipsets in systems I built. I stuck with BX chipset as long as I could, although I did use i815 based mobos for a short period during the transition from P3 to P4, and after that it’s been smooth sailing from i845 on.
Research and review is very much required when building quality systems. And finding out what to use and what to avoid is cruicial.
Don’t know about 915, but I had no problems with the systems I built using 915P based mobos. It’s possible some of the 915G* may have been problematic but I rarely use onboard video.
Never used 925 but I’m quite happy with 945.
I think you’re overstating the minor problems Intel may have had with a few of the chipsets.
An AMD64 3500+ that is properly configured will stomp a 2.4ghz P4. I recommend that you direct your clients to more experienced vendors if you aren’t seeing such results on systems that you build and/or configure.
Yeah, one would think so eh? I was surprised myself.
Unfortunatelly a lot of software, in this case Photoshop 7, seems to like the P4 much more than the AMD64.
And for your information I did not built that system. A client of mine bought it in a store without consulting me first. The system works fine but the Photoshop 7 benchmark was disappointing.
Untill very recently most commercial apps were simply better optimized for Intel than for AMD.
So, yeah, a mediocre, two years old P4 can wipe the floor with the latest and greates Athlon64 when it comes to running a lot of everyday software.
But it’s nice to see AMD being hyped so much since it will keep Intel prices down and competition is always good. Both for the AMD fanboys, kids building Athlon sytems to run Halflife2, and people like me, building computers for everyday business use (using Intel hardware of course). :o)
well good point about the chipset, but when tou compare costs you most take power consumption in to account (last time i checked power was not free) and who wins ther? TCO is not jost hype, if used wisly it can be a good tool.
well good point about the chipset, but when tou compare costs you most take power consumption in to account (last time i checked power was not free) and who wins ther?
In terms of power consumption, it looks like AMD has a good advantage:
http://www.tomshardware.com/cpu/20050509/cual_core_athlon-19.html
Could it be that these Photoshop filters use instruction sets like SEE/SEE2?
Could it be that these Photoshop filters use instruction sets like SEE/SEE2?
SSE and SSE2 are supported on all Athlon 64 chips. SSE3 is supported on all dual cores and single cores since Venice.
Benchmarks have been rigged in the past by looking for GenuineIntel rather than checking the appropriate CPUID bit to determine whether these extensions are available.
apple should optimize os x for amd — the benchmarks speak for themselves
The Turion isn’t great competition for the Pentium M, for better or for worse. And we’ll see how Intel’s Crusoe fares compared to AMD’s competition this time next year.
Holy macaroni!
I am getting a Athlon X2 3800+ in about 2 weeks!
That will be some holy macaroni, I tell you.
Dual boot WinXP and Linux, guess what that will be?
Some 1,5 month ago I read on the DragonFlyBSD forum archives that this very processor finishes a “make -j 8 buildword” in less than 23 minutes!! With build times so fast I may finally get to explore the BSD’s a bit better.
Then there will be pasta all over the place.
I bought a PC with an AMD 3800×2 CPU and I am not disapointed.
I was really surprised by how well Windows balances the load over two CPUs on a dual core.
Most of the time the load is even across them.
Now and then you see something max out at 50% which means it is stuck on a single thread.
Even then it is still good to be able to do something else at the same time.
Competition is good… plus!
http://www.swallowtail.org/naughty-intel.html
Thats more than enought reason for not to trust Intel…
Intel has tried to directly cripple other CPU manufactures (yes… even those small ones) with the obove tactics and some strong arm monopolistic moves…
Intel loses all of my benchmarks… I plased an Intel Pentium 2 on my keboard so that it would complete my “GIMP” bechmark… and even thoug it had all of its pins free to manage the keaboard and mice… it just stood there… I will need to revice my testing mesures! Maybe it needed the fan to cool it!
-iMoron
right now is the X2 4400, I agree 100%. Pair this baby with the nForce4 and you have an economical powerhouse that will serve you well for many years. Downgrading to the 4200 or moving up to the 4600 gives comparable performance if your budget is outside the nice 4400 middleground.
AMD is amazing. What a rise to power. I love em!
“apple should optimize os x for amd — the benchmarks speak for themselves
”
You fool PC users are overstimating benchmarks. 🙂
I think Steve Jobs doesn’t read (and doesn’t want to read) benchmarks. Intel can offer to Apple Centrino, VIIV, financial support for the transition and it’s strong production partner. AMD can’t: not wi-fi solutions, not media center solutions, not enough money, it isn’t a motherboad productor. It’s simple. 🙂
… is that the relative gap between the CPU (registers), L1 & L2 caches, memory, HDD, etc. just seems to increase. It does not help that much if the CPUs get increasingly faster, so long as the memory and HDDs don’t keep up. I have an X2 myself. And while it’s extremely fast by all means, it’s not as much utilized as my CPUs were 10 years ago.
The overall experience of speed as time passes is *not* exponential. It’s linear, at best.
And please bear in mind that I’m not talking about computer games, that’s another story.
This seems to indicate it’s much closer than you suggest”
http://www.tomshardware.com/motherboard/20050629/intel_955-18.html
Well, Toms Hardware is known for being very Intel-biased and using crippled benchmark suites, so I would recommend comparing their benchmarks with benchmarks from other testers
@dylansmrjones
Well, Toms Hardware is known for being very Intel-biased and using crippled benchmark suites, so I would recommend comparing their benchmarks with benchmarks from other testers
Care to back that up with some hard facts, amd fanboy?
I’ve been reading Tomshardware since probably it’s beginning and have seen many benchmarks and reviews there that were sometimes in favour of Intel and sometimes of AMD, and as far as I can remember they always reflected the reality.
In fact back in the early AMD Athlon days they clearly stated that AMD was faster and my only gripe with them was that they didn’t state more strongly that the chipsets from Via and Sis were utter crap.
I’d go as far as stating that all these current benchmark and hardware websites that cropped up in recent years are just Tomshardware wannabes and have very little if any credibility.
AMD Fanboy?
Huh?
It’s fact that Tomshardware has been using benchmark suites crippled by Intel and using them to show how good Intel was. So has others, and they have apologized for it – Tomshardware hasn’t. They instead keep using them.
In regard to Tomshardware I agree with AOD.DK that Tomshardware is biased towards Intel.
I’ve seen tests too on Tomshardware showing AMD was fastest, but they’re sure doing their best to avoid such accidents to occur.
I’d like to see how Gentoo Linux perfoms on these beauties
“It’s fact that Tomshardware has been using benchmark suites crippled by Intel and using them to show how good Intel was. So has others, and they have apologized for it – Tomshardware hasn’t. They instead keep using them.”
You got nothing to back this up. ONLY an AMD PC gamer fanboy would make such claims as they threaten the beliefs of the FPS crowd. Show us some proof of this or stop spreading disinformation.