Slashdot Mirror
AMD's Roadmap revealed
NoPants writes "It looks like the aces at Anandtech were able to get their hands on some of AMD's internal roadmaps. Anand has some interesting information including the new upcoming Socket 939 CPU standard as well as AMD's predicted release dates for Athlon 64 4000+ processors. Hopefully this will shed some light on what AMD is trying to do with all the different socket types..."
Maybe all of this preliminary information will help Intel markitecture their way to the Pentium 6!
but my amd 700mhz proc is still quite fast enough to give me debian, fluxbox, openoffice etc.. ah well there must be enough people out there who cant live without a fast proc
http://www.virtualconcepts.nl/
The Slashdot staff created this story with a timestamp from a Mars watch. It's thirty nine minutes off.
Ya, it bothers me too. But if Intel hadn't messed it all up by designing a "clock speed only" processor, then we wouldn't be in this situation. I definately agree that the Performance Ratings are a very poor way to describe a processor...
There's no question Athlon 64 and AthlonFX are great products. That being said:
*Do they really need to be different products? Opteron is your product for server/high-end workstations, Duron (and now Athlon XP) is low-end... you want Athlon64 to be mainstream, right?
*Is it really a good idea to have the memory controller on the CPU? OK, I buy that it increases performance, but it hasn't lowered mainboard costs and all I've seen it doing is causing a rift between the A64 and AFX product lines, since Athlon64 doesn't have a dual-channel memory controller.
*Why in the world introduce an AthlonFX based on Socket 940, especially at the outrageous price, when you're moving to socket 939 imminently?
I think it would have been more of a slam-dunk as a platform and a "brand" to release Athlon64 as all dual-channel, all Socket 939 (or some standard), and left Opteron as the high-end platform. Any other takers?
AMD's new stuff has been pretty impressive, but it really bothers me when they pull this type of stuff: AMD Athlon 64 3700+ 2.4GHz 1MB Q2 '04 AMD Athlon 64 3400+ 2.4GHz 512KB Q2 '04
,or how much you're gaining by going with the bigger (performance benchmark inflation not withstanding).
What's the problem. They're saying that having the smaller cache gives you less performance. Are you upset that they happen to have the same clock speed? I assume you'd prefer nomenclature more on the order of "AMD Athlon 64 2.4/512 and 2.4/1024"? In many ways they way they are currently doing it is more descriptive to the average buyer. No guessing as to how much performance you're giving up by going with the smaller cache
Wow if these new chips radiate at a proportionate level to AMDs current offering we could all be wearing shorts in Antartica by new year 05.
Seriously tho. I think AMD ought to work some better thermal performance into its cpu range. A low cost, low temperature, high performance CPU is what is required in the market.
Not to mention the fact that it lowers costs to them, and thus to the consumer, to do this. From what I've heard, they are able to take cpu's with some bad cache, which isn't uncommon, disable that non-functioning section, and then sell the cpu as a 512k cache cpu rather than wasting the entire chip. Lower performance, lower cost, but less waste. This is a far cry from the world of the Intel 486 sx vs dx with the math co-processor fiasco.
Was that night on the marge of Lake LaBarge I cremated Sam McGee...
The only reason I'm even considering one of these gems is so I can cram more memory into a system for video work. All the boards I've seen for Athlon 64 max at 3Gb. The SK8* boards for the Athlon FX will take, IIRC 8Gb. Where's the boards I can cram 32 or more into?
A feeling of having made the same mistake before: Deja Foobar
Hopefully this will shed some light on what AMD is trying to do with all the different socket types..
Making us buy more motherboards, of course!
need a free COBOL editor for Windows?
What the article does not cover, is when we will be able to purchase non-Opteron Dual processors. Since they are inherently capable, it would be nice to know when we'll be able to build a performance (non-ECC) dual desktop.
Not a prob- the 3700 and 3400 are for the average Joe. As long as they are a fair indication of performance there's nothing wrong with it.
The rest of us who cared about the details (which software, how fast) can and would find them out fairly easily.
The lower clocked, bigger cache part does run a bit cooler though, so if a quiet fan is what pops your cork, go for it!
When I am king, you will be first against the wall.
By the look of the figures the Athlon 64 is margionally faster than the G5 clock for clock... (the 2.2 Ghz beating the 2 Ghz G5 convinvingly and the 2Ghz ones locked in a tight battle). It looks a lot like AMD are gonna have to ramp up faster though, because IBM are gonna have 3Ghz G5s by Q3 this year, and AMD are only saying 2.6Ghz by Q4. Bob
Nice response, anonymous coward. First of all, I wasn't karma-whoring, I was making three (in my opinion) logical points about what I dislike in AMD's processor architecture choices.
If you don't like reading informative remarks, well-reasoned opinons, or getting factual link, I guess you can go somewhere else.
Run Doom 3.
Compute Pi to 5.497558e+11 bit precision during your lunchbreak.
Store this value of Pi in RAM.
Install Kazaa and not notice the spyware slowdown.
Use the faster page loading times to get FP more often.
Beep beep.
So 940 gets moved to opteron only
:)
939 encompasses both Athlon64/FX chips, starting in Q2.
754 is relegated to the next gen AthlonXPs (with the on die memory controller, but only 32 bit)
462 dies a slow death.
Why can't every CPU made just fit on Socket 7...
Faster processors are all very well and good, but what is the point of going to the 64 bit processor if there aren't any OS's or applications that can take advantage of it? We get really cool, expensive equivilents to what we already have. Maybe AMD and Intel should start a push in the app department at Microsuck, otherwise I'm content to hold onto my Athlon 2Ghz. It seems to be get the job done...
With all the talk of unexpected 90nm delays and scaling problems over at Intel, one hopes AMDs problems are just typical delays with a new process. At any rate, we hope AMD will push ahead of Intel with the K8 architecture. IFF this happens, how on earth are they going to market them using their "false" speed ratings. Their rating system is flawed in that it uses Intel chips as the gold standard to measure performance against. You can't market a 4000+ if Intel has no 4GHz processor. If you do, you risk having the rating not match when Intel catches up, which makes the numbers completely meaningless. Today, they at least help to compare apples and oranges.
The ratings have been mostly accurate when comparing AMD cpus, and they're useful when talking to non-computer people rather than explaining things like cache size, locality, etc. With the Athlon64, it should be even more accurate since they now control the memory controller. If the performance rating can make it easier for the average person to compare systems accurately and pick the better value for them, then I'd say it's a good thing.
Also, normal users don't care about 10% performance differences. I've found a good way to relate to this is to think about old computers. Do you really think of the difference between a P133 and a P166? No, they're both from the p5 generation, with a L2 caching up to 64MB or 128MB depending on chipset and the cpu performance is basically equal (slow). If a p166 will do the job you need nowdays, a p133 probably will too, and it's doubtful that you'll notice much difference.
False advertising? It's called marketing. They never said "Our 3200+ is equivalent to Intel's P4 3.2C". Find that on the AMD site and you are making sense. Also, compare a duron chip to a celeron chip that are "rated the same". There is no comparison, but then, neither company specifically said "these two chips are the same speed", so you can't really complain. You can't blindly trust a number to tell you how well a processor will perform, there's a lot more to it than MHz and GHz. As for Tom's Hardware, I would look at who pays their bills before counting on their "benchmarks" too closely.
the Performance Ratings are a very poor way to describe a processor...
...
I agree it's all marketing, but
As a data-point, I recently upgraded from an Athlon 2400 to the 3200 -- a 30% improvement in "PR," and saw my benchmarks go up almost exactly 30%. By benchmarks here I'm talking about a CPU bound computational mechanics code I wrote for my thesis. About as useful to everyone else as SPEC, but very relevant to me.
As usual, YMMV.
Running at full CPU load, an Athlon 64 3200+ uses less power than an Intel P4 3.2GHz. Furthermore, with AMD's Cool and Quiet power management enabled the Athlon 64 CPU slows down to 800MHz and drops to 1.275v when you don't need much CPU performence, ie, while I'm typing this message. ASUS has a nifty little program that displays the current CPU speed and core voltage on my desktop.
.
AMD CPU power requirements are expected to drop substantially when they switch to 90nm in the second half of this year. OTOH, Intel's prototype 90nm Tejas CPU burns up 150 watts
AMD chips haven't used more electricity than Intel chips for years. Pay attention.
BTW, Athlon 64 notebooks are out. $1,550 for a widescreen 64-bit notebook! I'm going to stick with my Athlon 64 desktop, at least until I come up with an excuse to buy a portable. Really, I am...
I'd guess it would be in the 2.5-2.6 GHz range, the fastest 64 bit stuff is in the low 2 Ghz range now.
Degaussing scares the bad magnetism out of the monitor and fills it with good karma.
there is big difference between p133 and p166mmx.
If you don't like reading informative remarks, well-reasoned opinons, or getting factual link, I guess you can go somewhere else. and where would we find these again?
Don't like to respond to a troll but:
The cpu speed is not on par with the ram speed. This fact cause that from time to time the cpu has nothing to do because the new datas from the RAM are not already there. A cpu doing nothing has a worst PR than a cpu doing something. Adding the support for the dual-channel double the speed to the ram and thus the cpu have more work to do and less idle time then a better rating.
*Why in the world introduce an AthlonFX based on Socket 940, especially at the outrageous price, when you're moving to socket 939 imminently?
Because AMD were making Opterons. And then some bright head said "Why not rebrand the 1x Opteron to AMD FX and sell it as the worlds fastest desktop system, 64 bit desktop, and whatever else? That way we can ship it NOW."
And Intel's Emergency Edition was almost the same thing. Basicly a rebranded Xeon, if I understood it right. Lots of cache, also hideously expensive. But it matched the FX.
I'm sure they were (and are) working on a dual-channel socket 939 board. But in the meanwhile, the FX was a perfectly good PR move. As for upgrades - I don't know about you, but my replacement cycle is getting so slow I don't care.
By the time I replace this computer, I expect the next to be with PCI-X, DDR2, different socket, 64 bit (if AMD), SATA native, preferably a smaller form factor (Shuttle-like mini-PC?) and more which simply doesn't exist yet. There was a time when CPUs were improving so fast, with compatible sockets that there was a point to upgrade, I did so from a Duron 700 to Athlon 1200. But now? By the time I don't like my XP2000+ anymore, I don't think I'll like the rest either.
Kjella
Live today, because you never know what tomorrow brings
Well you have compared an apple with an orange. A Athlon 2800 is not on par with a Pentium 2800. On some benchmark the Athlon could loose, but on other Win. If you want to try to compare the 2, compare them on the exact application you use and need power to use them. Not some synthetic benchmark, or ISPEC or FSPEC and then you could make your choice. But please don't troll about the PR rating. AMD NEVER say that the PR rating is something equal to the frequency of an Intel processor. And as a side note. If you compare your applications on the 2 architectures: Please recompile them with the exact architecture of your processor and the hadoc optimizer. This can change a lot of thing. I am not an AMD or Intel FAN. I just choose the best Proc/MB at the best price for the work I need to do. (And I have the facility to try them before buying :))
What I really want to see is dual CPU capability for the Socket 939. I guess the opteron is a viable alternative, but the price is just too high. Something like they did with the Athlon MP would be great.
more cache only matters when multitasking, not when running a single process. This is why Intel has always put larger caches on their Xeon line.
I don't read or respond to AC posts
Correct me if I'm wrong, but I was told the PR xxxx+ was in comparison of what a Athlon Thunderbird clock speed would be.
If you need better performance it may be another reason to switch to linux
Only Microsoft didn't catch up, but who cares ;-)
Regards
You can get a 4-way 846 Opteron with 16gigs already stuffed IN IT for that price, and a few 143GB SCSI drives.
Unless you meant to be funny. But really, they're not THAT expensive. Look to spend about $600-$1000 on a board that can support 8+ GB ram.
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
I think you misspelled "Pentium III", "Pentium 4", "Celeron 4", and most most especially "Centrino/Pentium-M".
... Oh, I forgot, AMD doesn't make money.
Or by successful do you mean with nerds who think IPC is somehow a better way to achieve performance than ramping clockspeed?
In the _real_ world, a successful processor is one that makes the company money.
Like the AMD
Read it again. Your XP 2800+ is a SOCKET A cpu. It won't work in a socket 754 mb. The new chip is a socket 754, same socket as the Athlon64. Get it? This new chip is a 32 bit Athlon-xp that plugs into an Athlon-64 mb! Why? Because AMD probably has plans to kill socket A and ALL desktop Athlons will move to the socket 754.
What a load of crap. I have bought only AMD processors and I swear by their performance. I jump at the chance to race a P4 touting fool through a compile of Mozilla. But this is getting ridiculous.
They've got to the point with all these different lines that it's no longer possible to talk AMD CPU's with anyone but the most avid AMD enthusiast. If you do try to talk the talk it ends up being a group memory excercise to see if together we can remember 50% of the difference in the varying jungle lines of processors.
Opteron is a good thing. Keep it simple. Give the FX a real name too. Don't call it an Opteron FX or 64 FX or whatever the hell it might be. Give it a damned name. How about an AMD Jargon? That would be a good name for a processor. If they all had names, you could associate the capabilities of the lines to the names and people could pick a favorite and learn the product.
As it is there must be extremely few people that can rattle off all the cache sizes, 398043+++ ratings, what core it is, blah blah blah. Not only do the different lines have different specs, but there are different specs within a line. There several instances of the same + rating with different specs in the same line. "I got a 2600+" "Which one?"
Not that I won't do all the research before I buy the next one, but I envy the Intel enthusiast that can just look and say Bigger is Better, and buy what they can afford.
I can't imagine my the other members of my family buying an AMD. They'd have to take a 3 week course, 2 hours a day, before they'd know which of the AMD's to buy. "This one costs more, but is it better?" "I have no idea." "This one has a bigger rating." "Yeah, but I heard this one is more advanced."
Is AMD hoping nobody will know what they're buying? Is that the ultimate goal? Why not just put a random number on each chip and put a MSRP on it and call it good.
The reason why they keep introducing new sockets is precisely because the processors are changing design so rapidly (power requirements, location of the memory controller, address/data line widths, # of hypertransport hookups).
The AthlonXP is radically different from AMD64 (using hypertransport) which is configured differently than the Opterons.
If the nextgen XPs get a builtin memory controller, OF COURSE THE SOCKETS GOING TO CHANGE.
The only design decision I'm wondering why they couldn't just use the 939 connector for the nextgen Athlons and just ground the unused pins.
Probably the answer is that the extra unit cost made the difference: they wouldn't be able to sell the XPs at a decent price point because the 939 sockets are not cheap.
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
You'd be a fool not to spend a paltry $90 on a 2500 and then a trivial overclock to 3200 (2.4GHz, IIRC).
RIGHT NOW.
You're lucky if you can get an PIV 2.4GHz for $90. And it won't be using the 400MHz FSB. (266, probably).
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
The plain Athlon 64, sure. I see why someone would buy that. If you want decent performance but also want to keep things dirt cheap, that's a nice chip. I think a low-power "mobile" version of that processor will also be a winner.
But if you want to spend a little extra money and build a "hotrod" machine, the Athlon 64 FX is a dumb move. Most CPU-bound stuff that people do, is parallelizable. (The only major exception I can think of, is that today's apps for multimedia encoding, tend to not take advantage. But they could (e.g. the portion after every key-frame could be handled by a different thread).)
So just spend a little more (it's really not much) and get multiple Opterons. If you're really hurting for money, get "obsolete" 240 models, and two of them will still run rings around any Athlon64FX or single-P4EE system that money can buy.
The class of problems that can't take advantage of multiple CPUs but still needs lots of speed, is small. Maybe I'm just being dumb, but I just don't see a market for a socket-939 or single-socket-940 board. Why would AMD, and motherboard manufacturers, bother to spend money development something that hardly anyone needs?
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
"As for Tom's Hardware,"
Anyone waving Tom's Hardware in your face is part of the "Knows Just Enough to be Dangerous" crowd anyway. These are the folks who think that reading a dumbed-down article or review makes them knowledgable enough to render a viable opinion.
They're also the people who burn up their mainboard and CPU trying some voltage mod they read about, then want to return everything for a refund because it no longer works.
-- "Government is the great fiction through which everybody endeavors to live at the expense of everybody else."
And since Microsoft is once again behind the curve, the various freeware Unix platforms could benefit a great deal by trumpeting their inherent advantage over Windows in these key areas.
...-.-
The "2800+" IS the speed. No, it's not the clock speed that the chip runs at, but that really doesn't make it any more or less relevant to the speed.
Intel decided to market their processors using one totally meaningless measure of performance, the clock speed.
AMD decided to market their processors using a different totally meaningless measure of performance, the model numbers.
The problem with the chips is not so much sleezy marketing as with are ridiculous focus on clock speed. That's like buying a car solely based on the displacement of the engine. Sure, a 3.0L engine might produce more power than a 2.5L engine, but that certainly isn't always the case, and it sure doesn't tell you much else about the car. But if the public bought cars based ONLY on the displacement of the engine like they buy PC's based ONLY on the clock speed of the processor, you better believe that companies would bring out "3.0+" engines with 2.5L of displacement.
In any case, AMD's marketing numbers have been extremely successful for the company. They almost instantly increased AMD's profit when they first brought them out years ago. Argue all you like, but the fact is that the computer buying public is, by and large, generally uniformed and buys systems based on two numbers alone, clock speed/model number of the processor and the price.
"They're also the people who burn up their mainboard and CPU trying some voltage mod they read about, then want to return everything for a refund because it no longer works."
You say that like it's a bad thingThe P166MMX wasn't a rip off, it overclocked well.
My P166MMX will happily do 266mhz @ 75mhz bus stable. Exactly the same as my P200MMX does in that board stably. It's still performing it's function happily running W2K server same as it has been for the past several years.
The real problem back then was the OS it was running on. It's not easy to judge stability when you are running Windows 95/98. But put them with a decent OS & decent drive controller and they happily clock to the max your board allows in most cases.
Freedom is merely privilege extended unless enjoyed by one and all.
more cache only matters when multitasking, not when running a single process. This is why Intel has always put larger caches on their Xeon line. This is not TRUE !!! Use SETI as example. More cache is needed only when you do work on big data chunk. In SETI example 16 MB. So you need LESS requests to SLOW main memory. IT DOES NOT depend on MULTITASKING !!! Boy, 95% of posts and replays here are from ppl. who doesn't know Sh.t about computer technology and how CPU realy WORKS.
Q: Which processor should I buy?
A: The one that makes the biggest profit for it's manufacturer.
Very strange logic.
And if we're talking about sucessful processors, lets not drag the Celeron 4 and Centrino into it. Those things are both complete dogs. Read some reviews if you don't believe me. But of course by your unquestionable powers of deduction, they have been sucessful product lines as far as intel is concerned, so they must be good for consumers!
OK, you want roadmaps, read the Inq. The roadmaps come in 2 flavors, Opteron and A64. The one Anand gets might have similar numbers to the A64, but it seems way way off. Some things are just not right there.
Before you jump down my throat, I write for the Inq, and I can say I have seen both roadmaps, and the Anand one seems, well, lets just say a tad off. Well, lets not be as polite, flat #&%^# wrong. The anand crew is under NDA to the eyeballs, so if they had a real roadmap, they couldn't publish it.
If they did, they get sued and cut off. If not, it means it is a deliberate leak from AMD, or they made stuff. I believe it is a little of both, and personally expect more from AMD this year. Hell, I would put money on it after the chats I had at CES with them.
-Charlie
Since motherboards and CPUs upgrade in lock-step anyway, there's no point in leaving them as separate components. What we need is a simple passive backplane design.
ATX was a huge step backwards in that it integrated the serial, parallel, PS/2, and other ports onto the motherboard, making all these connectors and their associated controllers "disposable" with every motherboard upgrade.
Put the CPU, caches, memory controller, and RAM sockets on a simple card that plugs into the backplane. This will be the only piece you upgrade regularly.
Then put the drive controllers, USB controllers, system health monitoring, and other miscellany on a "super i/o" card, covered with connectors. This piece won't get thrown out when you change processors.
What do we end up with? Getting a new CPU will result in replacing only those components that need replacing, keeping a pound of other components out of the landfill.
Manufacturers could still charge the regular amount for a CPU/carrierboard, because that's what people are used to paying. It would involve less design and debugging, because all the perhipheral crap is on a separate board. It would mean more choice for consumers, because you could get whatever CPU you want, coupled with whatever I/O board you want.
What's wrong with this?
Anyone besides me drooling at the prospect of a 2.6 ghz athlon64 fx? (course the price will likely make you piss your pants)
Url to testp ?id=as mpl&cookie%5Ftest=1
http://www.gamepc.com/labs/print_content.as
justified by a simpler motherboard; the 940 uses a six layer motherboard, while the 939 only uses a 4 layer one.
Since it is optimized for more compression in favor of speed (using Huffman coding instead of Lempel-Ziv-Compression). For compression speed, you might want to have a look at lzop.
sig is my sith nature.
I would have guessed that they'd take a page from Boeing (whose newest jet is the 7E7), and call it the i80E86. Since the mundanes don't consider "E" a numeral, Intel can go ahead and trademark it.
My other car is a 1984 Nark Avenger.
Aren't Riceboys a valid subset of consumers? Somebody should cater to them...
Seriously, why did companies start locking SMP in the first place? Did Intel really think that some Fortune 500 would decide to use overclocked Celerons instead of Xeons in their next mission critical server?
Doing the CPU interfacing up through a riser card isn't without it's hassles. Timing's critical. RF noise is a problem. Heat dissipation as well. Not to mention the added costs of socketing and various mounting issues. If anyhing that 'super I/O' card would be the thing on a riser. But then there'd be the hassles of interop.
In the face of how cheap it is to add the support chips and connectors it's not worth the bother.
Oh yeah, the entire cache is flushed every time you switch processes.
What was I thinking?
The secondary cache is intended to prevent misses to the primary cache, that is going to be more likely the more processes you are running concurrently. SETI may be a great way for you to feel good about yourself, but that is NOT the kind of work 99% of servers out there are doing.
Further, SETI is an example of something that WON'T at all benefit from a cache. Why? The data is unique! There is nothing to cache! Now, of course some of the analysis routines will be cached, but not the SETI data chunks themselves.
You are right about the purpose of cache, but you are wrong about how it goes about decreasing the number of requests to main memory.
Anyway, why I am responding to some fucking idiotic child with a warez site in his profile. He can't even write a sentence without ridiculous amounts of emphasis.
I don't read or respond to AC posts
The point of my statement was that things they're doing, like memory on the processor and such is in preperation for where computers will eventaully go, which is multi proc system in the home.
Anonymous Cowards - Oh God, How I hate you