Slashdot Mirror
Athlon XP1900+ -- Faster Than A 2GHz P4?
doormat writes "AMD releases their AthlonXP 1900+ Processor today, thats 1.6GHz. And it seems like its enough to topple the P4-2.0GHz, even in Quake 3 Arena!! AMDMB has a review of it." Ian Bell points out an AMD press release on the new processor. I love watching my old Athlon get slower every day ...
http://www.tomshardware.com/cpu/01q4/011105/index. html
Wait 20 seco....SHUTUP!
One thing I've noticed over the last year or so when talking to non-techie friends/family is that many people with relatively little knowledge of IT as a whole are starting to realise that the processor speed, however it's being measured, is far less important than the vendors want them to think.
The end result of Intel and AMDs battle of "my processor's faster than your processor" seems to be that people are saying "I don't care" - as they realise that there 'obsolete' PII is actually perfectly capable of doing all the things they use their PC for and that only graphics people and the hardest of hardcore gamers actually need 1.5 to 2GHz.
I don't doubt that the XP1900 is faster than
the P4 2GHz, but at amdmb there's only a test against it's smaller brother the XP1800.
Wheres some real tests comparing it to the P4 : )
I'd rather they waited a a little in between releases, rather than every couple of megahertz.
I'm not saying that speed is bad, but do we have to have a release for every chip?
-- Don't believe the megahertz myth!
Now, I know have the hype comes from the readership here, but I just keep wondering why AMD and a few other companies like Transmeta get covered here so lovingly. Is it because Slashdot readers don't like frontrunners ? Is there something inherently open-source-dogma-friendly about the corporate philosophies about AMD and Transmeta (though I doubt it, I am sure their lawyers are or would be as agressive about patents and infringements as Intel) ? Surely it can't be just about performance - Transmeta lacks sorely, and I cannot imagine the day when Slashdot posts an article crowing with glee about how the P8 trounces the AMDXP6400 or whatever.
i just love how quake 3 has become the benchmarking programming for any video card that comes out now...
maybe they should design the cards to work better in quake 3 and ignore the other...oh wait, i think someone may have already done that...
Argh!
:/
I had just finished downloading the athlon xp 1800 yesterday.
Great now I have to download the new one over my 28.8k modem
What? You mean it's not a piece of software?
!
^_^
Nice editorial work guys.
I saw a review comparing an Athlon 1800 and a 1900.
I didn't see a single thing in there that mentioned the P4 being outperformed or toppled.
Just unsupported speculation.
Originally I bought a Slot A motherboard.... Then the chip makers decided to go socket. I could buy a new mother board, but everytime a New chip comes out.... I will have to buy a new mother board. I buy a new motherboard but it's only good up to a certain MHz. So when I buy the new chip. I need a new board again.....
Someone stop the insanity!
Linuxrunner
www.slightlycrewed.com - Because aren't we all?
We're not talking a couple of MHz, we're talking 130MHz for the AthlonXP 1800+ over the TBird 1.4GHz and 70MHz of the 1900+ over the 1800+. When you consider we're still barely in the GHz range, MHz still matter! If they released on every few 100 KHz that'd be different, but until we get up to say 15GHz or more MHz makes a difference, especially considering AMD's IPC over Intel's. But I'll step off the soapbox before I slip ;)
I guess you do have a point though... for bleeding edge people they won't care, but Intel and AMD are competing businesses in a big market, so they can't afford to slip behind each other, it's a vicious game.
...they didn't test it on quack3.exe.
Have fun: Join D.N.A. (National Dyslexics Association)
If this Ghz race remains the main focus for AMD and Intel, we might need Quake 25 Thunderdome just to keep up benchmarking. I am not convinced that a few Mhz is instantly better, but you can never have enough Quake engines :)
I intend to live forever, so far so good.
Oh yawnny yawn yawn!
;-)
A few more MHz here, a few more there, so what?
Admittedly with one of those new Athlons and 0.5G of RAM, Nautilus might be half-way usable
Seriously, apart from a few more FPS while fragging how much of a difference do those few extra clock cycles actually make?
[My pondering could be said to be jealousy, as my main computer is still a P166, but my XFCE/Evolution/Galeon combo works like a dream on it...]
Listening for the sound of the coming rain...
Back in the days of modems (anyone remember those?), US Robotics was a company that you could always go to when you wanted the fastest modem on the market. In a way, you could say "nobody got fired for adopting US Robotics". An ISP that selected US Robotics as their vendor knew where they were going, and they'd have the best speed. Customers would stick with that ISP because they knew that they'd have the fastest connect rates. (Okay, mind you, locked into a propriatary format and vendor.)
AMD is known for having the lowest cost. Period. Rarely ever are they more expensive than Intel. But I get confused about Athlon's strategy. They're not going to have the fastest CPUs for long periods of times, so for something like computer manufacturers, you're not going to select AMD for performance machines (even though they may currently be "on top") because you know it isn't going to last.
I suppose I'm getting far off on a tangent here, but I think AMD would be far more successful if they could continually be known for creating the best performance processor. Then, hardware vendors would be far more likely to adopt their processor and chipsets.
But I don't have my finger anywhere near the pulse of this market. Am I just plain silly?
. . . AMD lost so much street cred using that PR-rating like scheme. If they wanted to deemphasize clock speed as a measure of performance, picking "model numbers" that look a lot like clock speeds in MHz wasn't the way to honestly go about it.
CEE5210S The signal SIGHUP was received.
We just like things that don't suck. AMD's processors suck far less than Intel's because they go much faster at the same clock speed as Intels do, and they cost a lot less. Transmeta's processors don't suck because they are implemented with some really cool technology with potential that we have barely begun to explore. Intel's rather passe unless you're talking about the Itanium in which case the alpha was at least as cool a 64 bit processor a decade ago.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Why, oh why, do you people insist on using something like Quake as a measure for how powerful your system is?!
Because it's a benchmark that is *tangible* to most people. One doesn't get excited because a processor can do a 500x500 matrix multiply really fast, but run Q3 at 130fps and people start flipping out. Sort of like saying how fast a car can go 0-60--not that everyone is drag racing, but it's easier to understand than the amount of torque an engine puts out.
Where are we going and why am I in this handbasket?
"Games, by definition, are not suitable as real performance meters"
Err, you do know what most of these chips are going to be bought for, don't you? Benchmarking on the application you intend to use is the only sensible thing to do.
By all means, if you want a processor for doing matrix multiplication then test it with that, but most of the people after the new ninja AMD chip will be wanting an extra couple of frames per second.
"I Know You Are But What Am I?"
Hell, that's why I bought an athlon.. AMD chips have a reputation for being faster, on average, than Intel chips, especially for 3D gaming. Now, factor in the bang for the buck factor along with a little bit more screwing with to make work nicely, and the choice for me was a no brainer.. athlon..
..don't panic
There other reviews of this 1.6GHz processor at AnandTech and at AMD Zone and at VIA Hardware. Check them out.
===> An eye for an eye makes everyone blind - MG
FWIW, the UK mag PC Pro says in the Dec 2001 issue (this is with the 1800, not 1900, Athlon-XP):
"Is there any bad news? Yes, but only for Intel. The Athlon XP 1800+ sees AMD back as the undisputed king of the performance castle. With a score of 5.24 in our 2D application benchmarkss, the MESH is comfortably ahead of any 2GHz Pentium 4 machine we've seen. In 3DMark 2001, meanwhile, it achieves a monster score of 7,611 at 1,024 x 768 in 32-bit colour. The fact that it can still delIver 5,271 3DMarks at a resolution of 1,600 x 1,200 speaks volumes - Athlon XP and GeForce3 Ti 500 Is a potent 3D combination."
The same issue has a review of a 2GHz P4 which benchmarks at slower than an Athlon 1.33GHz!
Ok, Yet Another Lame/Appropriate Analogy:
Considering the last time a topic such as this compared the Intel's best P4 to AMD's best Athlon.
The Car/Engine analogy was used to no end and many valid points were made, but noboday really put it into a conclusive and easy to understand "package" that the Average Joe User could understand.
Recall, if you will, the movie "The Fast and the Furious" as the analogy of Intel vs AMD saga.
Remember the scene at the end with the race between the souped up Honda and the Toranado?
Intel's P4 is akin to the Honda, as it has a lot of "high-RPM's" and "high-tech" under the hood (i.e. 2.X Ghz and Rambus et al).
The Athlon is like the Toranado(?) and American Muscle car that had the "High Torque" and "lower-tech" that relied on brute force (i.e. 'superior' FPU and Large cache and the blower is similar to DDR-SDRAM in a way).
The end result of the race at the end of the movie was that they (for the most part) tied.
The current Intel/AMD debate is very similar, in that you have all this high RPM/low torque (intel) vs old school High Torque/mid RPM's (AMD).
If it is not on fire, it is a software problem.
So...
Does the XP chip require a new m/b or will it work in what I have?
Flame away, I don't care, I have better things to do than monitor every change in the PC world.
VIA released their C3 cool processor video that showed a C3 lasting 24 hours playing Quake with no heatsink or fan on. That was 800MHz. A similar speed Celeron hung after 5 seconds.
My world is in ruins..
Nothing to see here...
Remember, there are no stupid questions. But there are a lot of inquisitive idiots.
Anyone seen a comparision between P4 and Athlon when it comes to raw number crunching?
I remember a few years back when AMD was starting to pick up steam, the K5 (I think) had horrible FP performance. But with the Athlon, the FP pipeline has been greatly improved.
I did a few benchmarks a year or so ago for 3d graphics ops (matrix, texmapping, vertex computation) on a few machines--an Athlon, Intel, Sun, and SGI, and was quite amazed at how the SGI (while way slower in terms of Mhz) beat the AMD and Intel CPUs handily.
But in terms of price/performance, the Athlon chips have always a great deal
Where are we going and why am I in this handbasket?
Hi.
Large matrix multiplies are in fact used as benchmarks by many sites, and are part of specFP, sysmark, and other benchmark suites, many of which you'll see quoted (though not specifically the mat mult part) in these reviews.
The problem is that matrix multiply only tests your floating point/simd unit and your i/o bandwidth. Not a very comprehensive test, and unless you actually plan to do large matrix multiplies, quite synthetic.
As for your web page serving idea, it's called specweb, and anyone who is catering to or buying in the web server market cares a lot about this benchmark. It is a more comprehensive test than just multiplying matricies, but still only targets certain aspects of the cpu (I/O again, cache size).
Games actually make a good addition to these benchmarks. A modern game engine can tax a cpu a great deal, and will use a mix of integer and floating point applications, plus put pressure on the memory subsystem. If the performance isn't limited by the graphics card, then you can use games quite effectively as CPU benchmarks.
It's funny that you mentioned "real". If you're running sicentific apps that multiply lots of matricies, then matirx multiply is "real". If you're running anandtech.com, web serving is "real". And, if you're a gamer, games are real and the performance you see in them is what is the "real" performance limited.
It almost sounds like you want the "max", as generated through some kind of synthetic test. As in the performance you'd never, ever get in a actual application that you'd want to run once you'd bought the system. Which was how it used to be done, and it sucked, so everyone stopped. Let's not suggest we go back, hm?
The enemies of Democracy are
It's most web people don't understand that generating a page once and serving it 10,000 times is more efficient than generating it 10,000 times and serving it once each time.
I know no web designers are going to read this, and I know 99.9% of them are too arrogant to listen if they did, but here it is anyway: If your web page doesn't HAVE to be dynamically generated, then DON'T make it dynamically generated!
This means you should have dynamic pages in TWO situations:
1) You web pages are customized for each visitor (slashdot home, my.yahoo.com, google.com)
2) It is different virtually every time it is viewed (slashdot comments, or a page of stock quotes)
*sigh* Oh well. Hopefully the IT shakeout will help get rid of all these hack web designers.
The enemies of Democracy are
Indeed. AMD need a clear message to send out to people.
"60% the cost, 90% the performance"
Pretty much as fast, a heapload cheaper. They are not going to be 'the fastest' long term in the foreseeable, but they ARE pretty much up there now.
If I bought a PIII 600 or 700 back a while it makes NO difference now, they are both old machines way behind the curve. 20%, even 30% better performance is only desirable at the time. Pretty soon the machines are seen as roughly equivalent as they both start to offer less than half the performance of new machines.
This is a more complex message than 'FASTEST' but a whole lot more useful
Yeah that's right an astounding leap of 4.9% is a questionable bump.
Back in the PII/200 days that would have been like releasing PII/210. Yeay!
to help you heat your house for the winter...Those AMD guys are ok with me.
(+1 Funny) only if I laugh out loud.
Well, I don't know. I work at a PC Spezialist store in Germany, not the tech dept., though. We've been selling Athlon XPs from the day they were out. Before that, we've mostly been selling the TB 1000-1400 and some Durons. I've noticed that AthlonXPs are actually cooler than the TBs. You know, when you use the really cheap ones, the passive cooling part below the fan will get hot. Not with the XPs, with those it just gets a little wormer than room temperature.
About customers feeling tricked I can only say the following: Not our customers. There's generally two categories, Type A looks at the numbers and it makes perfect sense to him: 1200, 1300, 1400, 1500, 1600, 1700, 1800... You can't make it easier for him. Type B knows about the whole trick but generally also reads Tom's Hardware Guide/c't magazine/$YOUR_FAVORITE_SOURCE_OF_DECENT_INFORMAT
Um... I didn't do it!
AMD Zone gives this summary at the end of its review: "No architectural or marketing changes with this release ... expect the previous CPUs to decline in price ... expect a bit higher performance and power consumption."
Anandtech agrees, saying the chip will not offer any significant extra performance over the 1800+, so early adopters need not sweat too much about being left behind. The site believes that AMD is currently the performance leader on desktop processors.
VIAHardware.com reckons users could be just as well off picking up the 1800+ at 1.53GHz and simply overclocking it to 1.6GHz. Users already owning a high-speed XP chip are better off waiting for the next upgrade on the platform to significantly increase performance.
Tech Report has some extensive benchmarking, putting the 1900+ slightly ahead of Intel's P4 2.0GHz in most of them, while SimHQ.com gets very excited about the new chip.
Amdmb.com also has a piece showing the expected five to six per cent performance increase.
What on earth did you do with your home computer to make the heat sink fall off your graphics card? (Or was it not attached properly to begin with?) I'm sorry, but it looks to me as though Tom's Hardware was pretty desperate to make any sort of anti-AMD point it could (BTW, it was running Intel ads on its site back when the 2 GHz P4 came out)--rather like Car and Driver running an exposé of what happens to your car if you take the fan and radiator off and then go drag racing. Duh...
I thought we were going dual 1800 XP's, since the MP only went to 1.2. But this morning, checking the price of this chip, I discover 1800MP's all over the world, for a small price increase.
So what do I do??? Dual 1900 XP's? Dual 1800 MP?
This system is for smashing numbers and making sure that my code doesn't bring down the system before running it on the heavy iron (hey, bring down the beowolf or the SP2 too often and they get mad
or is a 1900 MP likely to be hot on the tail of all this?
hawk, who has to have the money spent by year end
To put this in perspective, when I got hear a year and a half ago, that difference is the entire speed of the machine sitting on my desk when I got here a year and a half ago, a 133 pentathingy with 160M. (yes, I know speed isn't linear, but when you consider the memory subsytem is four times as fast, uses less cycles as well, and is 2 or 4 times as wide, as well as cache & fpu . .
hawk, now using a 1G laptop and waiting for his dual athlon workstation
The MHz race is getting a little rediculous at this point if you ask me. Processors are coming out faster than we can keep up with. Compitetion used to be good for the consumer, now it is abusing the consumer. People will want to stop upgrading at all for fear of missing a newer processor.
did it really take only two minutes to compile the latest linux kernel? that's just insane, not that I compile kernels all that often, but still two minutes for a kernel compile is quite impressive. I forget how long it takes on my 1Ghz, probably 10 or 20 minutes and I thought that was fast.
Things you think are in the Constitution, but are not.
AMD used to be getting by on the "just as good as Intel but costs less" line of marketing.
In the last year AMD has been going on the "As fast if not faster and still costs less than Intel" marketing.
The marketing tack AMD appears to be taking now is "we're our own company with our own product and it's great" (without so much as the incidental mention of that other processor company)
This is the direction AMD has to go, to get out of Intel's shadow. The upcoming Hammer line of processors is a bold move in that direction -- with the advantage of having built in backwards hardware compatibility -- which departs clearly from the 64 bit architecture Intel has chosen. With ~20% of the market, though probably mostly in Europe and Asia, AMD should be making testing these waters.
All that aside, you as a wise consumer, should choose the CPU that's "right" for you. By "right" I mean speed, efficiency at your primary task, with reliability and support to meet your standards. A difficult decision, really, considering most buyers get suckered by a minimum wage salesman on a commission and make important decisions truly uninformed. Lucky for most of them that you really will not miss the mark by too much, whatever you buy, though customer support is usually where people meet their grief, so consider that a primary factor over speed, etc., unless you're a bold, devil may care, geek who provides your own customer support and get the rest off the net.
A feeling of having made the same mistake before: Deja Foobar
While Intel has a fab on the ethnically cleansed land of Al Faluja I will never buy a Intel CPU or knowingly buy a product with any Intel parts in it.
Some may go on about the fact that AMD's fab in Texas is built on Indian or Mexican land, but those Indians or Mexicans weren't driven off while the Geneva Convention, Hague Convention, the IDHR or the UN exited.
The fact is that until Israel permits the return of Palestinian refugees (to both Rump Israel & the Occupied Territories) & returns all illegaly expropiated lands its in contravention of the Geneva Convention (A49P6), the Hague Convention (1906C), the IDHR & dozens of UN resolutions.
Now as Intel did not lease or purchase the land its Israeli fab is on, from the people with the internationally recognised legal title deeds to that land (Palestinian refugees mostly living in Egypt) its an illegal fab on ethnically cleansed land. So I'm not ever going with Intel.
BTW a good percentage, if not most P4s are made in that Israeli fab.
A better question is when is the last time a fan failed? Not everyone monitors their fan speeds/temperature, and it is pretty hard to know if a fan has died if you have other fans in the box whirring away.
You'll find that a 1 deg C/W heatsink-fan becomes a heck of a lot worse than that when the fan stops. If you are cooling 30 watts at 1dC/W you are probably looking at 50 deg C on a mild day. Without fan, you could easily be talking 110 deg C. I'd hope that in this situation my computer would halt as opposed to frying my new $300 processor.
The sink doesn't have to "fall off" to be in the failure range.
In this case the processor heats up slowly, and this means the processor will shut itself off, or slow itself down.
When AMD gets headlines by introducing a new processor chip that runs only 70 Mhz faster than the previous chip, you can bet their marketing has been/will be successful.
AMD has an advantage. Unlike the old Cyrix PR ratings, these chips really do outperform their intended Intel counterparts. Maybe its just me but I don't think this would be news unless the 1800+ 1900+ etc. rating system was working its way into the minds of the consumer
That's it. I've had all I can takes, and I can takes no more. This is a message to all of you that write reviews and occasionally do up graphs.
GRAPHS THAT DON'T HAVE A BASELINE OF ZERO ARE MISLEADING.
In the VERY FIRST GRAPH, the numbers show a 6fps difference, but the bars seem to indicate a 100% performance increase of the 1900+ over the 1800+.
If you don't start at zero, your proportionality is lost. You can no longer eyeball the graph and get a rough feeling of what the difference between the test subjects is. You have to read the numbers to be sure, and that defeats the whole purpose of the graph!
You should be able to roughly analyze performance (or whatever) WITH NO NUMBERS ON THE GRAPH. This is why pie charts are useful. A small slice is small. You don't have to look at the number to see that it's a small piece of the pie.
In conclusion: do the damn graphs up right, or don't bother with them. You aren't conveying any actual information if you do it wrong.
No one ever wants to flat out say that the motherboards for AMD chips are a lot less well supported than the motherboards for Intel chips because they're so busy cheering for the underdog.
s p4-15.html)
But if you dig deep into, say, Tom's Hardware Guide: Another factor is the stability and product quality of a system: while all Athlon processors suffered from occasional instability in our tests, the Pentium 4 platform ran without a glitch. (http://www6.tomshardware.com/cpu/01q4/011031/xpv
Now, for me and I'm guessing a lot of people, system stability is far more important than a few percent performance increase. Since these machines are so closely matched and overpowered anyway, I'd like to see more emphasis on other factors like stability. More than a single sentence buried in one review, anyway. If these things are crashing during the tests, I want to know about it with a big red X on the graph...
Or just the chance to stop having to download freakin' 4-in-1 drivers for my KT7A...
its been faster than every 18 months for a while... Moore's law is better expressed/replaced by The Law of Accelerating Returns
"I would say that 99 per cent of what my father has written about his own life is false." - L. Ron Hubbard Jr.
Yes, I was in the cabinet taking the heat sink off at that time. But it didn't take very much force to break that ear. I can imagine that much force might be generated by a tall, heavy fan receiving a sideways G shock, of the kind you might cause by transporting the cabinet improperly.
You have to figure that at some point in the future, these mobos will be expected to last more than three years before being replaced. Believe it or don't, but there are people out there who don't upgrade their systems annually.
John
John
I guess this means my GPF screens will pop up that much faster, so I will lose less work when my wordprocessor/spreadsheet/morpheus pr0n sessios gets whacked. When they finally reach 10ghz, I wonder if the CPU will tell me "Don't even bother firing up Word, I'm going to crash in 7 seconds".
Seriously, why pump out faster cpu's when they provide nil benefit ? Yes, I do have an Athlon 1000 running anywhere between 1200 and 1466, depending on my mood. I have no idea what to do with it, I actually bought it just to out-clock my buddies (until one smartass bought a water-cooling system - that's cheating). My Geforce2 is still maxxed out, even my previous Celeron was able to push it to the limit. My hard drives are still slow, and I have better things to do than buy more drives to widen my raid-0 stripe. It's already quite clear that the CPU is no longer the most important part of the computer, yet they still bust their asses trying to produce bigger numbers just to bleed us dry of our hard-earned money. We need better memory, better hard drives, better cd-roms, better video cards, better everything, but not CPUs.
I think that AMD and Intel should help out Micron, NVidia, Maxtor, etc. We've reached a point where faster processors just don't yield much more performance, but if they would be wise enough to pitch in and actively work on the other functional parts of a PC, the entire system would become more efficient, not just some over-hyped core that overheats 2 zillion times per second while waiting for an i/o transaction.
-Billco, Fnarg.com
So stop keeping up with them!
Keeping up with processors isn't any different than memorizing any other sort of ephemeral trivia. Just start ignoring the press releases, and it'll be just like giving up TV: you won't miss it. Unless buying/speccing PCs is part of your job, you do not need to keep up-to-date with the latest trivia.
Then, when the day comes that you need a faster box and you think upgrading the hardware is the way to go, and you want to stay within x86: here's what to do:
Surf the net for a couple of hours, to get familiar with what is currently out. Then:
Buy a $200 to $250 processor. The dollars are everything. The clock speed, model number, etc. is trivia that you don't really need to know, except for purposes of motherboard compatability.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
Well, probably because having such a number across the benchmark set would be pretty meaningless. An average of FPS in 3 different games isn't going to be very useful, because the games may be very different. FPS in Quake3 are 2-3 times those in Max Payne. For an individual benchmark, normally the reviewer makes several runs to ensure that they get the same number, or at least statistically insignificant variation.
But on the other hand, suites like Sysmark, 3dmark, etc DO produce a single score that represents a number of benchmarks. It uses some arbitrary weighting scheme to come up with this number. Which is why I tend to ignore those scores, because you don't know what they really mean.
If you want to know how the CPU performs on all benchmarks, just look at all of them.
The enemies of Democracy are