Slashdot Mirror
AMD's Duron Birthed
maniack writes "The AMD Duron, the "Celeron-killer", has finally been released and lives up to the hype. According to these reviews from Ace's Hardware, Gamer's depot, Anandtech, and Tom's Hardware, the Duron thrashes the Celeron clock for clock and even hangs with the P3 in a lot of the benchmarks. Looks like AMD has another winner in the value market to go along with the Athlon."
Some good links.
n /duron_cpusink.jpg
Looks like the Duron runs pretty hot too, http://www.aceshardware.com/articles/reviews/duro
Ace had some detail on overclocking the Duron, but I want to know, is it going to significantly reduce my heating bills.
I could get one for my granny to keep her warm during the winter. I could write it off on the tax too.
But seriously, is it not a case of too little too late??
$110-$130 is a good price but we've had the celeron for a long time, is there room for another low end chip in the arena, or should they be aiming higher with the performance. Perhaps as a cheap upgrade option for the celeron??
Search and destroy, lunch is served
I'm confused how you can be against one monopoly (I assume by your nick) and for another (once again your nick). If you want the best performance (that is if you can actually find a 1GHz coppermine from Intel in a retail outlet, unlike Athlons which are plentiful and much cheaper, and then spend $800 for Intel's great motherboards that only support RDRAM, thanks to those nice investor options!), get a Thunderbird (an Athlon with on die L2 cache running at CPU speed_. Duron was meant to beat Celeron as a low cost alternative and does that easily. It isn't a high-end CPU.
You are compairing list price vs. street price. Price Watch has them listed at $89 (600Mhz) and $159 (700Mhz). I expect the prices will be lower when the get released for real.
Wait a few weeks and check for Duron prices on pricewatch.com.
While you're at it, compare with PIII prices, because that's what Duron's performance is closest too, not Celeron. Remember Duron has a full speed 200MHz FSB, while celeron is held back to 66 MHz.
Duron is a unbelievable price/performance deal!!!
I think you were using the -ffast-math flag (or a buggy gcc), which breaks the ANSI/IEEE compliance (it is not turned on with any -O option).
Anyway, i'm quite surprised. I've done some FPU intensive computations with a K6-2/300 and a PII/350 (time domain finite differences), and the PII was 3 times faster than the K6-2.
There's a new standard "AGP Pro" for AGP slots that provide more power.
It's nothing to do with AMD/Intel or specific motherboards. Certain video cards need an AGP Pro slot to get adequate power - you need to check the specs before buying.
I want to get a dual AMD system just so I can say it's a "Duron Duron".
--
Compaq dropping MAILWorks?
Linux MAPI Server!
http://www.openone.com/software/MailOne/
(Exchange Migration HOWTO coming soon)
Right now your PC processor choices are PIII/Athlon or Celeron/Duron. In both cases the AMD processors are clock-for-clock faster as well as cheaper, in addition to being available in higher speeds. The latest Dell catalog I received still only has 866MHz PIII's, while Gateway has 900, 950 and 1GHz Athlons. Duron is available up to 700MHz (actually there are 750's on sale in Japan!), while Celeron is still 700 MHz.
There's always going to be those who hold out for tomorrow's Mustang (AMD) or Willamette (Intel), or whatever, but for anyone buying today AMD's choices provide bother better performance as well as value.
The only benchmarks I've seen where Intel beats AMD are one or two where Rambus RDRAM makes a difference, but with cheap DDR SDRAM a few months away, it's hard to justify paying today's crazy RDRAM prices. If you're buying an SDRAM PC and want top performace, then AMD is your only choice.
Certainly AMD has disappointed as an investment in the past, which is exactly why the stock is so ridiculously cheap (forward P/E of 15 or so, vs 40-50 for Intel).
Last quarter AMD earned $1.15/sh completely blowing away street estimates of less that 50c, and by the estimates I am hearing it will provide another MAJOR positive surprise when current quarter earnings are out on July 19th...
If you wait to invest in AMD until it has a history of positive results you will miss out on the big gains to be made by those who realize that profits will follow performance.... Note that when AMD announced it's 1GHz Athlon, that for the first time in history Intel had to price the 1GHz PIII UNDER AMD's part!!!
The word "Duro" in Italian means "Hard"... AMD's marketing should have done a bit more research on meanings... I formally give AMD an "X-Rating" :-)
--
The world is divided in two categories:
those with a loaded gun and those who dig. You dig.
--
The world is divided in two categories:
those with a loaded gun and those who dig. You dig.
Duron is not intended as a mobile/low power part. It's meant to compete with Celeron (and even PIII) on the desktop.
AMD's next mobile (low power) part will be a version of their next generation "Mustang" core due out later this year.
AMD are selling processors as fast as they can make them and ramping up production in their Dresden fab as fast as they can. The timing of the mobile Mustang should coincide with their ability to produce them in sufficient volume to satisty the market.
http://www.sharkyextre me.com/hardware/reviews/cpu/duron_700/
I only trust the smartey reviewars.
Small potatoes make the steak look bigger.
As for the speed, like I've said, the network could have had something to do with it.
For the compile options, I used the same set at home, with the same compiler. The fast-math flag is turned off in both cases. I checked, because I was surprised. I spent several hours going through docs and man pages, checking (and manually configuring) every option. I couldn't find any difference outside the chip.
- W. Blaine Dowler
http://www.bureau42.com
You can't really compare the prices of a Duron 700 to an Athlon 700, because all 700MHz and below Athlons are not Thunderbirds (the newer Athlons with 256KB of on-die L2). The non-Thunderbirds are being phased out. BTW, according to PriceWatch, a Celeron 700 and PIII 700 are about the same in price aswell.
Not a winner on performance; not a winner on price; not a winner, period. Pity.
I just looked at AnandTech's Duron review, and on their benchmarks the Duron was far faster than the Celeron; IIRC, a Duron 700 beat a Celeron 850 in Quake3.
And as for it not being a winner on price, on PriceWatch a Celeron 700 is about $50-70 more than a Duron 700. It sure looks like a winner to me.
+-+-+-+-+-+-+-+-+-+-+
S i g n a l 1 1 <-- note an extra space at the end
+-+-+-+-+-+-+-+-+-+-+
Will I retire or break 10K?
I just read the review of the Duron at Sharkyextreme, and I have to say, it was really poorly done! Here's a link to it:
p u/duron_700/
i cle_id=5000159
http://www.sharkyextreme.com/hardware/reviews/c
If you look at the system specs (on page 9 I believe), he used different video cards, different sound cards and different memory for the various systems! Hardly isolating the processor/motherboard! What's more, there's no mention of which chips are running in-spec and which are overclocked (the Celeron 550 is obviously an overclocked chip since no Celeron 550 exists, but what about the other ones?). And what's all the big deal about the power consumption? Sure, 41W is too high for the mobile market, but it's not as high as the power consumption of many of Intel's PII and PIII chips, it's certainly lower then nearly all of AMD's other Athlons, and it doesn't come close to touching the 100W+ power consumption of some Compaq Alphas! Sharky also has some kinda weird numbers in his benchmarks, ie the Quake ones, where the Celeron 500 (with the older core and no SSE) scores higher then the Celeron 600 (with the improved core and SSE instructions). There were even a few things that were just flat out incorrect (ie the cache associativity of the new Celerons is only 4-way, not 8-way as stated in the article).
Anyway, here are a few reviews that were somewhat better:
http://www.aceshardware.com/Spades/read.php?art
http://www.anandtech.com/showdoc.html?i=1261
Actually, it's Latin (durabilis)-->Old French-->Middle English-->English, but who's counting? And then there's always the indo-european root d-e-u-schwa....
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
I've seen benchmarks of Intel vs. AMD under Linux, and unfortunately Intel generally kicks AMD's ass. I suspect someone needs to take a look at gcc's optimizations...
Windows 2000 RC2 for one, it worked inconsistently on an AMD (Compaq Presario 400 MHz notebook), sometimes BSODing at boot, other times working ok. Sure, they fixed it later, and Release is supposed to work ok.
If the release version is supposed to work with AMD processors, then why are you complaining? Beta versions of software aren't necessarily going to run on all of the possible hardware platforms. I'm sure there are plenty of pissed Intel owners who can't get Linux 2.4.0-pre1-ac8 to work, but they're waiting patiently.
For more information, click here.
If I can get a genuine Athlon 700 for ~$150 now, why would I want to buy a Duron? Not a winner on performance; not a winner on price; not a winner, period. Pity.
One reason would be that they are switching to a new architechture (socket). If you buy Duron, you could end up saving the cost of a new motherboard on your next purchase. Your pricewatch link shows Athlon Slot arch chips, which allow an upgrade to the Slot Athlon 800, no more. If you buy a Duron and want to upgrade to an Athlon Thunderbird, you don't need a new mobo, since they use the same socket.
If I were choosing between the two, I'd go with the future upgrade potential of the socket systems (Duron/Athlon Tbird).
Of course, if you already have an Athlon slot m/b, it's a no brainer to just continue to the best Athlon slot chips.
I was using 'dorks' in the semi-affectionate sense, though it's not obvious. I honestly do appreciate having read your thread.
And I automatically (not through moderation) get a score of 2 because of my existing karma rating, which, with posts like this one, is getting trashed daily.
; -- the corruption of government starts with its secrets. a truly free people keep no secrets. --
Intel and AMD can never play on the same mobo now.
The Athlon (and the Duron now) use *completely* different underlying architetures. If you put an Athlon into, say, an Asus CUBX, it won't post because the CUBX (and any Intel chipset for that matter) isn't DDR.
So, while an open socket and complete interopability between Intel and AMD would be nice, it will be a chilly day in Hell before it happens.
Is this post not nifty? Sluggy Freelance. Worshi
The Duron's FPU is just as powerful as the Athlon, since it's a derivative.
> Another was the AGP incompatibility. Many people had random lockup problems using an AGP card with an AMD. Again, maybe it got fixed by now, but it proves the point.
Some motherboards for the AMD did not provide enough power to the AGP slot for some of the hot new video cards. For instance, I burned up a Voodoo 3 and a FIC PA-2013 trying to get them to work together. (Later revs of the FIC do support the V3.)
The only solution for this is to do a bit of research before you buy. (I'm learning...)
> So considering that the days of the significant price advantage of the AMD are pretty much gone
You don't spend much time hanging around pricewatch.com, do you?
To pick a random example, I just looked at the Athlon 800 and the PIII 800. Best prices currently listed were $312 and $536, respectively.
--
Sheesh, evil *and* a jerk. -- Jade
Being on a network has nothing to do with cpu load
Normally, I agree, but college networks are usually plagued with tons of broadcasts. In my experience, I showed 10 times the interrupts from the network interface than from the clock. This was a simple workstation, but Intel-based. Intel architectures aren't known for their efficient interrupt handling.
In the end, the extra interrupts didn't seem to affect my K6 from a usability standpoint, but it definitely bit my old 486. So, I'm not too quick to say that network could have nothing to do with CPU load, but I will say that it should account for less then 1%, probably less then 0.1% of the CPU.
It may be an informative site, but the number of words per page seems to be dropping all the time. And given that it's regularly linked to from slashdot, it's a pain in the arse; it takes upwards of ten minutes to read a two-minute review. And some of the picture links were broken.
I hereby suggest that a printable option be mandatory on all stories where more than one page is used. You hear me, w3c?
Offtopic? Probably. But pertinent to all our lives.
so Duron = Hardon?
eww!! ;-)
Go get your free Palm V (25 referrals needed only!)
Shine on, you crazy diamond.
AMD had compatibility issues through the K6 release, but I haven't heard of any starting with K6/2. I've been running countless applications and games on K6/2, K6/3 and Athlon processors for quite some time and have yet to encounter a single 'compatibility problem'. The AMDs have proven to be nothing but a better value than Intel's offerings time and time again.
Frankly, the only people I have encountered to bring up 'compatibility problems' are uneducated intel loyalists who have never put an AMD into one of their boxes.
Can you provide a URL to an article that discusses these 'compatibility problems'?
So if this is the celeron killer what are the implications facing Intel now? How is Intel going to respond to this latest attack by AMD? Anyone know?
Simple. With 1-Watt Celerons. Check the specs on the Duron: It's hot. REAL hot. 22 W for the 700 part!!! And that's at a 90 degree T-die!!! That's insane!!! This thing will never make it into mobile, let alone set-top boxes... With those two options out, it's left to compete with Athlon and Coppermine. I don't think its gonna do to well.
---
https://www.accountkiller.com/removal-requested
I'd suspect that the disk subsystem would be the main bottleneck for compiling, followed by the memory subsystem.
When you do a compile, you walk through many, many source files and nested levels of include files (take a look at your dependencies list for an idea of how nasty this is). A fast disk with scatter-gather (coughcoughSCSI) and a large disk cache in memory will work wonders for eliminating that grinding noise that tends to accompany compiles.
Now, the memory subsystem. Compilation with a modern optimizing compiler involves building monstrously huge structures in memory and shuffling them around (I spent much of the last two semesters writing such a compiler). There's enough data here to overflow the cache, so you run into the system memory access speed bottleneck.
100 MHz FSB is Your Friend. 133 is even better, if you have RAM with low enough latency (another poster correctly pointed out that 133 MHz does you no good if wait states eat the extra speed). A larger chip cache will also help, if that's an option.
Lastly, if neither of these seems to be the bottleneck, having a faster CPU or a second CPU may help. Just make sure that the CPU is the bottleneck before forking out the money for a faster one. You can also safely ignore floating-point performance - almost all of the computation performed by a compiler is integer math (building bit vectors to calculate dominators and so forth with).
As far as benchmarks go, the disk- and memory-limited ones should still be useful to you.
Most people are only concerned with speed in their benchmarks. I switched from Intel to AMD a few years ago, for cost reasons. I started running physical simulations of wave propogations for class and lo! what do I find? The P2-300 and K6-2 300 chips produced different results! They were close, but not the same. (For those who understand this: I was using a fourth order Runge-Kutta integration with initial conditions that used a lot of exponentials.) The AMD chips actually produced more stable results: roundoff errors were more accurate more of the time! The results also came through substantially faster, but I don't know how much of that was processor, and how much was two different levels of overhead. (The school's machines were on a network, mine weren't. Both were RedHat 5.1.)
Bottom line: AMD produced better numbers, and it produced them faster. I have yet to see an official benchmark that looked at accuracy, but maybe they should.
- W. Blaine Dowler
http://www.bureau42.com
From the AMD press release:
The 700MHz, 650MHz, and 600MHz AMD Duron processors are priced at $192, $154, and $112, respectively, each in 1,000-unit quantities.
If I can get a genuine Athlon 700 for ~$150 now, why would I want to buy a Duron? Not a winner on performance; not a winner on price; not a winner, period. Pity.
Rev Neh
... and there is no doubt, that one day he will be
where the eye of his telescope has already been
Nah, they should just hire you two complete dorks to tell 'em all about it.
Seriously though, very interesting thread. I appreciate the fact that you two are very literate. Wish there were more like this around... sad to say, I can't play, but I have sure enjoyed watching.
; -- the corruption of government starts with its secrets. a truly free people keep no secrets. --
We've put up a review of Duron at AMD Zone, the largest AMD site on the net. We also hit a few of your problems (less pages, more paragraphs) and use more benchmarks. For example, Currently, there are five standard SPECopc viewsets: ProCDRS viewset is intended to model the graphics performance of Parametric Technology Corporation's CDRS industrial design software. IBM's Data Explorer (DX), which has 10 different tests, is a visualization application. Intergraph's DesignReview (DRV), which has 10 different tests, is a 3D computer model review package. Alias/Wavefront's Advanced Visualizer, with 10 tests, is an animation application. And Discreet Logic's Lightscape Visualization System, with four tests, is a radiosity visualization application. All five viewsets represent relatively high-end applications. These type of applications typically render large data sets. They almost always include lighting, smooth shading, blending, line antialiasing, z-buffering, and some texture mapping. So we use Specview, Seti, Distributed, and a few other benchmarks made by a friend of ours who attends the University of Illnois. QMC is best described in the words of Tim QMC tests cpu power, cache quality to some extent depending on the file Input1D0, bus io and ram. It calculates the ground state properties of a particular lattice model using a Quantum Monte Carlo algorithm. Linpack is a very interesting benchmark for comparing the Tbird to the classic Athlon because you can watch the data get too big for the CPUs caches and watch performance fall. Here is Tim's description of the benchmark. For a matrix of size N you use 8 * N * N bytes to store that matrix. For 25 sites you use 5KB. Each time the matrix is increased by 25 the memory consumption rises by 4 fold. So you use 5KB, 20KB, 80KB, .... , 5.12 MB. At some point the L2 is overflowed. At that point performance will depend on how quick data can be accessed from ram. This is where you will see a noticable performance drop, or at least you should depending on the latency of getting data from ram. So for matrices that are larger than L1(data) + L2 you are really stressing the quality of the cache and the bus speed. Also you pay a tremendous performance hit for using CAS3 ram. I think it's a very good benchmark. Also to those who commented that Athlon is about the same as Duron's price. That is off Price Watch, which is not the MSRP which was released as always in press releases, and was OEM. Duron will cost much less at Price Watch when it is readily available than its MSRP just like Athlon does now. Hope that helped.
PLUS, you get a free Hungry-Like-A-Wolf Cluster.
--
Compaq dropping MAILWorks?
Linux MAPI Server!
http://www.openone.com/software/MailOne/
(Exchange Migration HOWTO coming soon)
Having read a couple of Spitfire (I can't bear to say Duron) and Thunderbird reviews I am struck by just how irrelevant the performance figures are to me. Like a lot of Slashdot readers I do a *lot* of software development and spend forever watching gcc, msvc and Delphi chew processor cycles. I'd give anything to find out the most effective platform for what I do.... how quickly Word loads or Q3A runs is not likely to be representative. The few benchmarks that these sites run that do involve a compiler (some of the Spec* tests use gcc and the ZD Content Creation bm uses msvc AFAIK) never quote the results separately from the whizzy 3D piccie software.
Wouldn't it be great if we could get hardware reviews of setups that software developers are likely to want. Is the Thunderbird really quicker than the Coppermine and Spitfire for building Mozilla on Linux(or Windows)? Is it worth paying for SCSI or is that PC266DDR SDRAM what we should hold out for?
Sound good, but the socket 462 vs socket 370 aspect is about as funny as the "Duron, Duron" jokes.
It's expensive to develop chips of this size and I can understand that both Intel and AMD want to protect their IP, but we really need an OpenSocket (or maybe a slot variety, but I'm not going to write that).
That would be for the good of all, more fair competition and I would be able to use the newest CPU with my one year old motherboard.
TC - My Photos..
I've said this before and I'll say it again -- for you investors, AMD is a surefire thing. I bought it right when the k7 roadmap was announced, and they've done nothing but good things since. AMD is taking over their market -- might as well join in for the ride.
Mike Roberto (roberto@soul.apk.net) -GAIM: MicroBerto
Berto