What is Digg?142 Comments
- geminitojanus, on 10/12/2007, -8/+32I'll have you know that pretty much every desktop processor designed since the Pentium 4 has been a RISC processor.
Shocked and amazed? Don't be. Truth be told, CISC fought the good fight, but in the end, it was too easy to atomize operations, and multiplex them together to make CISC operations. So how is it that a modern RISC processor can run CISC operations? Simple; a CISC frontend decoder. AMD trumpets the fact that they have 3 single issue x86 instruction decoders in the front of their Athlon 64 lines, while the Intel processors only have 2 (though there's no real Apples to Apples way to tell if it is actually making a huge difference). What's even newer to the Pentium M and Core processors is something called Micro-op fusion, which actually takes the RISC instructions that the front end decoder has generated, and re-combines them into a much more effecient mono-instruction, or instruction packet, whichever the case may be. ([just an example, technically inaccurate, but for those not versed in microprocessor design this will work] if something adds then subtracts a number from a register, the processor can combine them into one add operation which helps the ALU out by only having to do one operation).
So yeah, RISC processors are great. That's why Apple moved directly from one RISC platform to another. And it's also why they dissed the older CISC processors like the Pentium (P5) and Pentium Pro (P6) (and not so much Netburst aka P7). - victory, on 10/12/2007, -6/+29i guess the 80% is from retail desktop pc's such as those sold at best buy/compusa/etc. so it doesn't include online or mail order such as dell.
- hunchback, on 10/12/2007, -7/+23the last time I checked, AMD chips are far lower priced than Intel's "premium products" and with better performance too
- Obsydian, on 10/12/2007, -6/+19Thsi couldn't have come at a more ironic time since just last night I convinced a die-hard Intel fan to get an AMD for his computer. We built it today. Gotta love Tom's Hardware comparo's.
- Obsydian, on 10/12/2007, -7/+18Intel's business tactics are monopolistic. Look at the deal they wanted with skype, to give "Intel-only" features. That's *****. Engaging in these tactics doesn't mean market share dominance, and AMD's been coming on strong for a few years now.
- inactive, on 10/12/2007, -5/+15I tell my clients who have never heard of AMD, Cheaper-Faster-Cooler.
Thats all they need to hear :) - geminitojanus, on 10/12/2007, -3/+12@happyscrappy
"These new processors may have been designed with what we learned from RISC, but they are not RISC processors. The Intel processor doesn't have a modern instruction set, which means it cannot be classified as a RISC processor."
Well, most of what's left of the CISC chip inside modern implementations are in the far early steps of the pipeline; nearer to the end of the pipeline, everything's micro-ops, and if the industry were ready for a shift in ISA (which they've been resistant to), Intel could uproot the units from these modern chips, take the new ISA and build a chip relatively quickly. A lot of people theorized this was exactly what was happening when Apple moved to Intel, and that the new Core chips were actually going to be PowerPC chips that used the Pentium 4 elements, and that wouldn't be all that far fetched (that is, if Intel and IBM weren't so resistant to working together on processor technologies; Intel would have to license PPC and IBM would make a mockery of them for it).
"One of the big principles of RISC is that the architecture exposes some of the hazards of running code and the compiler is supposed to schedule the instructions to work around them. That basically means to be RISC, you need to be a load/store architecture, and x86 isn't. It includes inherent loads and writebacks in with other instructions and doesn't have enough registers to even allow the compiler to separate these out if it wanted to."
I knew someone would call me out on this, but trust me, the internals of the Pentium 4 and beyond are so RISC-like that it would be nearly trivial to remove the archaic nature of the x86 archetecure from the chip and build a new archetecure (or move to PPC?) around what's left. A lot of the operations that would need to be required to be updated would actually have to be grafted on from scratch. But then again the LOAD/STORE units are present and integrated on the Netburst/Core archetecures, but could be dis-integrated and mapped to address their own instructions if Intel needed it to. Without access to the Micro-op spec I can't tell you for sure, but I can say it's a rather safe bet that MOV has been implemented through LOAD/STORE inside of these chips.
You'll especially see this on the Core chips if you were to pop the hood; more and more instructions from the old x86 instruction set are being atomized and "fused" (aka re-ordered and packetized) to make the processor faster without deepening the pipeline and driving up cycles (I think Intel learned their lesson).
Of course, going into any more detail would probably violate someone's NDA, so I won't go there. But we're a farcry from where we were with early P6. There's a lot of really detailed discussion of it in the Wikipedia (http://en.wikipedia.org/wiki/RISCvsCISC ) and older newsgroup archives (and the Wikipedia article highlights the best of it), back when the CISC/RISC debate was a real issue in the late 80s/early 90s (now it really isn't; RISC lives on inside CISC and in stand alone implementations like POWER). And it would be an interesting debate to get into whether or not you could classify a chip that used a legacy CISC ISA with modern RISC units and implementations a RISC chip; I would call it one simply because it's more RISC than it is CISC, but that debate could go on for years on a newsgroup, so I'll let it rest for now. But, more or less you're right. The chip is a CISC chip that's very very RISC-like in design. - manfesto, on 10/12/2007, -1/+9Actually a good point.
I work at a big box retailer, and of all of our display machines, there's about a two to one ratio of Athlon 64 or Sempron boxes to Pentium or Celeron boxes when it comes to Desktops. I wouldn't go so far as to say that the AMD desktops are marketed as budget boxes, though - but I do tend to build them up to customers because of their value over a similarly priced Pentium 4 box (all you have to say is "cheaper, faster, cooler").
I should also point out that the 2-1 ratio is reversed when it comes to notebooks - Pentium Ms vastly outnumber Turion 64s on my store's shelves. - leonbev, on 10/12/2007, -3/+11Exactly. That 80% number covers retail purchased PC's like those bought at Sam's Club and Walmart. Those are mostly lower price PC's, where the buyers are more interested in the total cost than what processor is in the computer.
Intel is still the major player in Mail Order PC's for home and business (thanks to Dell), and is still leading in laptops. - HappyScrappy, on 10/12/2007, -3/+10gemintojanus:
These new processors may have been designed with what we learned from RISC, but they are not RISC processors. The Intel processor doesn't have a modern instruction set, which means it cannot be classified as a RISC processor.
One of the big principles of RISC is that the architecture exposes some of the hazards of running code and the compiler is supposed to schedule the instructions to work around them. That basically means to be RISC, you need to be a load/store architecture, and x86 isn't. It includes inherent loads and writebacks in with other instructions and doesn't have enough registers to even allow the compiler to separate these out if it wanted to. - geminitojanus, on 10/12/2007, -4/+11" (ways Intel "*****" up)
a) doing anything else besides producing & developing microprocessors.
b) switching sockets or memory systems every other generation, ensuring incompatibility and a general sense of confusion.
c) not wanting to give up market share to AMD, and using MS-like tactics in order to maintain their hold."
Point #1. Any successful Microprocessor company has to think about the products that Microprocessor will be used in. Intel isn't that diverse of a company; they make Microprocessors for a very, very wide range of applications, but that's generally all they make. However, they also develop C/C++/Fortran compilers, which makes sense as the designer of those compilers would be allowed access to otherwise classified/trade secret documents on how the processor internally dealt with things. They've also recently stepped into making demo enclosures and enclosure specifications, which again makes sense seeing as they're developing the processors which will have to live inside of those enclosures; they'd know better than anyone the precise thermal band and thusly the materials to use and the amount of breatheing room and airflow to design for. They've also got top notch engineers that can do fluid dynamics and the computers to simulate the fluid dynamics of the internals of a computer case in order to engineer them to be as heat-conscious as possible. Oh, they also make NAND RAM, but again, expected, they're a microprocessor company.
To your second point; You're holding Intel guilty of something that every single computer/microprocessor company is guilty of. AMD has been known to switch sockets WITHIN the same generation, and now with Intel's P4, they've made the same mistake. Personally, I find buying microprocessors so confusing today that I just go to my vendors and say "Hey, I need this processor, find a motherboard that works for it with these features and I'll buy it". But you can expect that to happen with the ever-changing nature of these processors; they've all got different bussing considerations and such, so things will change a lot. (and to be honest, I think AMDs got it a lot worse now that they've chosen to integrate the memory controller. Now every time they update the chip, they HAVE to replace the socket to make sure that the memory controller has enough channels to address the RAM, meaning that with every new core, you should expect to replace your motherboard). (Note, I'm really biased against the memory controller for this reason. Otherwise I'd love it because of the low-latency, but I'm the same guy who owned a socket 7 board until the Pentium 3s were released...)
As to scare tactics, I've seen a few, and I do scorn Intel for it, but AMD has released FUD of their own, and personally, I just hate FUD flingers on either side of the spectrum. If you've got the better product, flaunt it, stand tall and don't sink down to other businesses level of taunting. Just market it in its best light, and let your customers judge the products for themselves. And that's how business should work. But we know America, and that business can't be business unless it's trying to rip out someone's throat in the process. - geminitojanus, on 10/12/2007, -4/+11Why should they fire anyone? 64-bit processors are only truly useful to business that need to saturate their larger servers with multiple GBs of RAM, and to those clients Intel has been trying to sell the Itanium (laugh if you will).
As desktop processors, the 64-bit revolution simply hasn't happened yet, and still has a long way to go. Once users find that they can't stuff any more RAM into their 32-bit machines (4GB cutoff, more if you use segment registers in the way they were intended, as the older Pentiums can actually address 36-bit memory addresses this way), they'll start moving to platforms where 64-bit is existant. And as shown here on Digg itself, 1GB is currently enough, with 2GB not revealing any positive reason to have it.
So that's really not a selling point for x64. What is a selling point is the general effeciency of the platform; it's damned fast, and it runs at lower heat than your other option, the Pentium 4. And as we all know, Intel was secretly developing their own implementation of x64 called Yonah, and funny enough, when they unveiled the chips using this core (the Core chips), they didn't have the x64 instructions and registers enabled; the desktop/laptop market simply hasn't asked for them yet (though you can bet your ass Intel's server chip based on the Core core will enable the instructions, and Intel will market the HELL out of them as being more power-effecient than the Pentium 4s and AMD64 chips).
I love the fact AMD is growing, but attribute it more to other good things on the chip (the memory controller and it's extremely low latency, the more effecient design overall, etc). - geminitojanus, on 10/12/2007, -2/+8Manifesto (yeah, I really hate the fact we don't have threaded comments too, but at least the quality of civilized conversation has increased threefold since this new system has been in place, and I'd like to thank you for a sensible reply; I think every digger with a good reply deserves that much courtesy).
I'm not recommending against buying a 64-bit processor at all. In fact, to everyone I recommend buying AMD processors if you're buying a desktop/mid-range server as of current (though I still have to recommend Intel for laptops.. dear AMD please hurry with a true laptop competitor). The problem though is that applications can't be selectively compiled against x64/x86; with system's like Apple's Universal Binary, you can actually have two binaries inside of the same "application", and it'll run fine, but on systems like Microsoft Windows, you'll actually have to buy the version that works for x64, and that's not something our computer market likes to do (and that was shown with the Itanium). It also is a strain on businesses because they'll have to compile their apps twice and make sure that the binary works on AMD64 and EM64T (as there is a slight difference between the AMD and the Intel instruction sets, which is why I try to call the technology x64).
So, whereas geeks will find their apps compiled for both, and Microsoft will ship an x64 enabled OS, I don't see it being generally preferred by consumers until everyone's on board and we can drop the x86 archetecture into history (FINALLY). And it's really a shame too; all of that computational horsepower getting held back by a measely compiler and a few memory alignment issues... - waynejkruse10, on 10/12/2007, -8/+14good stuff AMD. I would like to see a nice 60/40 market share going on. Mmm competitiveness!
- Silencer7, on 10/12/2007, -1/+7I'm actually still using my 1 GHz Athlon from summer of 2000, and it handles everything I need just fine. Even more surprising that it's a freakin' Gateway...when it comes time for a new one, though, I'm definitely building my own system with a 64 X2.
- Saintlink, on 10/12/2007, -3/+9I've been very happy with the price vs performance from my Barton 2500. You can OC this chip to 2ghz without too much trouble and it still is very cheap. Great overall CPU's of late coming from AMD, the X2 series really rocks. Intel on the other hand, they really don't seem to have an answer (at least for this round).
- heatus, on 10/12/2007, -7/+13How are they dropping AMD support?
- ACalcutt, on 10/12/2007, -2/+8I wont build a system thats not AMD. I've built a few people AMD systems and they are happy with their computers
AMD deserves the top. I am going to keep pushing AMD because they are the best. - Osjpr, on 10/12/2007, -11/+16Quite right! AMD has been giving consumers more bang for their buck for a long, long time now. I hate how Apple make fun of Intel in their stage presentations and public relations campaigns. Yet Intel allowed Apple to stomp all over the Intel brand. You can't help but feel Intel deserves some losses because of their meek conduct.
- malliemcg, on 10/12/2007, -5/+10I find it interesting that it is only for the first 6-8 weeks of '06. It will be interesting to see how the rest of the year pans out. I believe that the competition in the x86 marketspace can only be good for consumers (of all types) as it forces the companies involved (presently Intel and AMD) to innovate and continually improve their products.
- chadseld, on 10/12/2007, -1/+6I just built an AMD64 desktop system, here is what I learned:
- Core Duo: nice chip, but the mother board offerings are not there yet. Expensive.
- Pentium 4: A chip at the end of it's life. If I build one now, will I be able to upgrade the CPU a year from now?
- Athlon 64: A single motherboard will support Athlon64/AthlonFX/AthlonX2. So I can get a single core Athlon 64 today and upgrade to a dual core next year. - StatusQuoRules, on 10/12/2007, -8/+13Well, when you look at their desktop offerings.. AMD64 vs P4, it becomes quite clear AMD's chips are just currently better than those prescott thingies (more efficient, can cool easier and quieter, faster, cheaper most of the time etc)
I use AMD64 for desktop PCs and Core duo for notebooks - DCstewieG, on 10/12/2007, -0/+5Then just choose 'sort by date (show all)'
- chicken101, on 10/12/2007, -2/+7I still have my 2 year old Athlon 64 3000+ (socket 754), and it runs like a charm.
- sinner0423, on 10/12/2007, -7/+12Intel ***** themselves with the following :
a) doing anything else besides producing & developing microprocessors.
b) switching sockets or memory systems every other generation, ensuring incompatibility and a general sense of confusion.
c) not wanting to give up market share to AMD, and using MS-like tactics in order to maintain their hold.
AMD has what it has today because it's stayed true to form. Intel can kiss my ass, Apple deal or not, their time is over. This has nothing to do with AMD fanboyism, it's simply economics. - inactive, on 10/12/2007, -2/+7Hey I just broke down and put together a AMD 4200 x2 system for about $1000 , that's a 939 motherboard 1 gig ram and case plus a AMD 4200 X2. and a video card. I am one happy camper this rocks my world. I can capture video do a virus scan, and surf the net all at the same time no slow down at all!
- manfesto, on 10/12/2007, -3/+8@geminitojanus
(man, I thought I wouldn't have to do the @username thing again - we could use multithreaded comments)
While I agree that the 64-bit desktop is not coming any time soon, I still tend to recommend people purchase 64-bit processors. Windows Vista will have 64-bit support (Linux/UNIX is already there, and I don't know about OS X86), and there are more advantages than just being able to address more memory. Yes, its advantages will still cater to a niche audience (64-bit MS Office wouldn't be all that much faster) but graphic, database, and many applications that make use of repetitive algorithms will see a great performance increase if they're written as 64-bit applications. Games, video encoding/decoding, databases, math-heavy programs, etc. will probably be the first to benefit, and these benefits, while not coming tomorrow, are coming soon. I think it'd be nice if a few more people were ready (as an aside, I think that, for all of its great design, the Core Duo's lack of 64-bit extensions was a bit of a waste).
Not to say that AMD shouldn't be touted for everything else they did right with this chip as far as efficiency and the memory controller, but I bought an Athlon64 for its 64-bit support. - pacho, on 10/12/2007, -3/+8Intel performs better where you live?
INTELLAND? - DCstewieG, on 10/12/2007, -1/+5It already has. Wonder if we'd be seeing dual cores if it wasn't for AMD helping push along Intel...
- fahadmp, on 10/12/2007, -1/+5Competition is good for the free market..hopefully this will benefit us consumers
- geminitojanus, on 10/12/2007, -1/+5"I do agree with all of your points, actually, as far as market acceptance of 64-bit applications go, but people out there are writing 64-bit applications - lots of them. I say again that they'll cater to niche audiences first (gamers, video encoders, programmers, etc.), but there are measurable performance increases, and because of this, I don't think that the 64-bit desktop is as far off as you think (though, for the record, I'm known to overestimate things)."
I'm known to be a bit more conservative on the issue, simply because I'm so close to a lot of the engineers inside of these companies (I went to school with the lot of the newer guys, I know a lot of the older guys from when they would come to our school and answer questions about their future [as long as it didn't break their NDA, and trust me, we had day long discussions that consisted of {Question: "Can't answer that, protected under NDA"}.]).
"I'd prefer the Windows world adopt a fat-binary system (perhaps it will happen with the dawn of 64-bit appllications hitting the market) - that would very much ease the transition - but that's out of my own hands (and may increase file sizes - I've always been curious to see if OS X unibins are significantly bigger, but I've never gotten around to checking)."
Ohh they are significantly bigger. But it doesn't matter as much after first run; the system has a tool already built into it that can strip "fat" binaries (if you want to, when I go Mactel I'll probably have to leave both in simply because of the Apple hardware I already own).
"And I do believe that AMD is working on a dual-core Turion 64, but I haven't heard anything about that recently (and I tend to call Athlon 64 instruction sets x86_64, mostly because it's 64-bit extensions on an x86 chip, and because that's what I'm used to looking for when looking for Linux packages)."
I know AMD is working on a dual-core Turion, but the problem is the Turion still isn't as effecient as the Core archetecure. There are a number of things AMD could do to make it work better (and sadly, one of those things is to move the memory controller back off chip; memory controllers are in almost constant full tilt, and that would keep the processor as a whole in a higher energy state than otherwise would be neccesary), but they haven't gotten around to it because they're trying to punish Intel's marketshare where it hurts them most (Mid-range servers). I just hope they haven't forgotten the fastest growing market sector of computers is portables (notebooks/tablets/laptops). - manfesto, on 10/12/2007, -2/+6@geminitojanus
Yes, I agree that civilized discussion on Digg has jumped greatly (something I always claimed a more /. like comment system would foster).
I do agree with all of your points, actually, as far as market acceptance of 64-bit applications go, but people out there are writing 64-bit applications - lots of them. I say again that they'll cater to niche audiences first (gamers, video encoders, programmers, etc.), but there are measurable performance increases, and because of this, I don't think that the 64-bit desktop is as far off as you think (though, for the record, I'm known to overestimate things).
I'd prefer the Windows world adopt a fat-binary system (perhaps it will happen with the dawn of 64-bit appllications hitting the market) - that would very much ease the transition - but that's out of my own hands (and may increase file sizes - I've always been curious to see if OS X unibins are significantly bigger, but I've never gotten around to checking).
And I do believe that AMD is working on a dual-core Turion 64, but I haven't heard anything about that recently (and I tend to call Athlon 64 instruction sets x86_64, mostly because it's 64-bit extensions on an x86 chip, and because that's what I'm used to looking for when looking for Linux packages). - dstz, on 10/12/2007, -6/+10« It doesn't matter, Digg's core audience are AMD fanboys. Their main concern is gaming and f*ck productivity uses. »
All professional uses are not made of 3D rendering only. Intel leaded that area, at the cost of higher cpu temperature than their AMD counterparts, and now they don't even lead that small area anymore.
~~ We have to say that the situation was balanced at the time of mono-core processors. The Pentium 4 used to provide better performances than the Athlon 64 for video encoding or even 3d rendering but the release of dual core processors changed this situation. The Athlon 64 X2 now provide better performances in these domains with very good results in other areas as well. ~~
http://www.behardware.com/articles/607-1/a-survey-of-40-intel-and-amd-processors.html - dmron, on 10/12/2007, -0/+4Maybe if Intel processors didnt cost an arm and a leg, people would actually buy them.
- dissident, on 10/12/2007, -3/+7I don't hold a grudge against either company nor am I a big fanboi of one or the other, I merely buy the best value for the money, and right now, that's an AMD Opteron 165 dual core processor. This is what will be powering my new rig.
I happen to get a good deal on one from monarch computer with a 60 dollar off coupon code too boot. :) - cranium, on 10/12/2007, -1/+5I've been using AMD for ~12 years. They're awesome.
- muikano, on 10/12/2007, -3/+7"AMD's surge can be seen most strongly in the U.S. retail market, which accounts for about 9 percent of global PC shipments. In the first seven weeks of 2006, AMD's share in desktops in that area climbed to 81.5 percent, while Intel's has slid to 18.5 percent, Baker said. That's almost a complete reversal of their typical relative positions."
Intel din't come out with 64 bit and their dual cores fast enough. If you wanted Dual Core 64 bits, you had to get AMD. Someone should get fired in Intel. - geminitojanus, on 10/12/2007, -7/+11"Arent PS3's powerpc?"
Yes. (and so are the Revolution and the X360). The CELL processor has a single, in-order PPC core and 8 (though only 7 enabled) SPE SIMD units on its core. The X360's processor has 3, in-order PPC cores much like the one on the CELL, and the Revolution's processor only has 2, though it's unknown to whether they're in or out of order (speculative) chips. - OregonTrail, on 10/12/2007, -0/+3Yeah!
Now they have the better dual core architecture(X2), great mobile chips (turion), the gaming side (FX60), the single user (A64), and a new socket on the way
There's no reason why anyone should buy intel again! - cebbs, on 10/12/2007, -0/+3uhm, the frying processor thing is more dependent on the motherboard you are using than the cpu. I have an athlon 2800+, and when I moved the heat sink fell off but my mobo shut off instantly whenever i tried to turn it on. It still works now, although a little outdated.
- TheHim, on 10/12/2007, -5/+8I've really become an AMD fanboy as of recently.
The A64 X2 is an amazing product for the average customer and it was about time AMD got rewarded for the recent years of impressive x86-64 CPU lineups.
My gaming machine features an overlocked Dothan though, can't beat that price/performance for this specific task. - Demagogue, on 10/12/2007, -1/+4I would say it might be awhile, but 10 years? It'll be MUCH sooner than that. 10 years is a long time. Just look how far everything has come in the past 10 years.
- lane.montgomery, on 10/12/2007, -1/+4@7of7
You can fix that by browsing at a lower threshold.
Complaining about it doesn't do anything to fix it. - foxhoundadmin, on 10/12/2007, -8/+11i agree with ya, malliemcg.
amds are great. fast, cheap, and well-designed in most cases. they have a better naming system, too. i don't know what the heck a p4 90210 is. although, the current pr system is a bit obsolete. they're basing their performance ratings off of technology that's almost a decade old (go look it up). the 1ghz p3 and 1ghz athlon are dinosaurs. now, we have 64-bit code, dual cores, and vastly larger caches. it's time to find a name that fits. - UbuntuAtlantis, on 10/12/2007, -9/+12Intel just havent innovated in years......
AMD have had more than enough time to not only catch up, but get ahead, just as the ratings show!
Hopefully Intel will bounce back, 2 neck-and-neck processors (market wise) is far better for the consumer, and SHOULD help kick start innovation again - inactive, on 10/12/2007, -1/+4Be careful what you wish for, it might come true, and did :) hahahaha. But 80% of desktops?? Amazing... and understandable. Surely no-one can complain about AMD not having production capacity to deliver (ehem, Dell?).
- asdfer, on 10/12/2007, -8/+11"Blame" it on the manufacturers. They are marketting AMD-based computers as "budget" computers. When I go to electronic stores, I see more of AMD machines than Intel ones. Gateway sells more AMD machines than intel @ retail stores. More AMD machines from HP/Compaq are on sale than their Intel counterparts.
More important than all, on average AMD machines performs better on the same price level. - HappyScrappy, on 10/12/2007, -1/+4Lane: every chip has a load/store unit. And MOV of course uses it.
You need to understand chip architecture better though. A load/store architecture has separate load/store instructions, it doesn't bundle them in with other operations.
To add 1 to a location in memory on a Load/Store architecture, you do 3 instructions:
load register, memoryloc
add register, register, immed
store memoryloc, register
The thing is that a load has a long latency, you issue the load, then it takes several cycles before the data is available. You can't add 1 to the data until you have it.
On a non load/store architecture, you do it all in one instruction.
This bundles 3 operations into a single instruction, and one of them is a high latency operation.
The principles of RISC were to get the data into registers, then operate on them, then put them back into memory. Ideally, the data would stay in the registers a while so you could do many operations on the data before incurring an expensive memory-accessing operation.
Intel cannot do this because they bundle combination operations into a single instruction and because they don't have enough architectural registers to store the data in locally to avoid having to send it out to memory and back all the time.
x86 is not RISC. The chips use principles of RISC (this started with the 486 actually), but they are definitely CISC.
This doesn't mean you can't get good performance from x86 chips. It does mean you couldn't make a low-transistor count, in-order chip that implements x86 and does so quickly. An in-order chip isn't useful when you need a family of processors, like in PCs. But it is useful when you make a system where you design it once and build many that are identical. Which is why you see all the consoles using RISC now. - jasdev, on 10/12/2007, -4/+7listen, this is good news, allthough for most of us weve been telling the saying "you can build an AMD system more powerful than an intel's but cheaper too" for quite some time already now ;b.
Thing is, its worrying to subscribe to this cheering for AMD(or firefox/mac/ or anything for this matter). Its great if AMD can rechannel the profits for more efficient production etc etc so the end result benefits consumers. The way history has been, if they become monopolies and they dictate terms to consumers - we have all this fanboy crap to ponder on. - CopyNinja, on 10/12/2007, -3/+5ha, i know someone who OCed his 71 dollar sempy 2500+ and compared it against his Pentium "Extreme Edition" which was a $1000+ chip, and it blew the pentium outta the water.
-
Show 51 - 100 of 142 discussions
The Digg Toolbar for Firefox lets you Digg, submit content, and keep track of Digg even when you're not on the Digg site. Download the official 
© Digg Inc. 2009
— Content created and posted by Digg users is