Slashdot Mirror
AMD Quad-Core Opteron (Barcelona) Tech Report
crazyeyes writes "AMD has been very tardy with Barcelona. Countless AMD fans have eagerly awaited a new processor. As the day draws closer, TechARP takes a look at the upcoming quad-core AMD Opteron. Is there more to it than just its four processing cores? Will it be the Intel-killer that AMD promised long ago? From the article: 'AMD is in the same boat as ATI. Delays after delays of their long-awaited Barcelona core not only ensured the dominance of their rival, Intel, in the desktop processor market, it also ensured that Intel would be the only choice for those who want a quad-core processor. Although that wait will end in August, 2007 when the Barcelona is finally launched, it remains to be seen if AMD's new processor will be able to inflict serious damage to Intel's dominance.'"
Actually, ATI is now in AMD's boat, which is literally the same boat.
Intel's sever / workstation chip sets suck next to the ones for Opterons / AMD FX cpus.
FB-DIMMS cost a lot and need alot more power to run then DDR ECC ram and the Intel chipsets have very few pci-e lanes. The nforce pro chipsets have the lanes for 2 full x16 slots with 2 x4 slots and pci-e lanes for on board sata / sas raid with x4 lanes left over that are some times used for pci-x slots.
Also the amd chips have better cpu to cpu link.
I recall Intel was up in a fuss when AMD released the 64 bit chips. The market 'ooh'd and 'aahd' in delight of the new architecture, supposing that it would herald in a new era of computing in a similar way that the jump from 16 to 32 did.
The reality of the situation became that the great majority of Athlon64 users were running 32 bit apps, and continue to do so.
There has yet to be a dire 'need' for 64 bit processing, much to the similar way that there isn't a dire need for more than 4 GB of ram in a desktop machine.
At work, I'm the Sysadmin for a dedicated hosting company (Linux, mostly Gentoo), and even in that market I don't know of any of my users running 64bit. any performance advantages are outweighed by incompatibilities and plain old PITA to get things working.
That said, the delay in developing these quad core procs shouldn't put that big a dent in the pocket / market share of AMD simply because it's a niche market that has yet to be widely adopted.
This is not the greatest
... it NEVER MADE _true_ QC CPU...
All existing Q6xx0 solutions are dual-dual core ie two dualcores sharing same FSB - and that is _NOT_ the same as true QC as Barcelona is claimed to be.
That difference is enough to make Barcelona the main choice for many core servers even if it were made with old K8 and not the new K10 cores.
Intel should have true QC chips in a year or so...
Someone want to buy me one? The only thing I have to trade for it is nothing... but I got a whole lot of it.
"When life gives you lemons, don't make lemonade. Make life take the lemons back!" -- Cave Johnson
My gf is going to school in late September and will need a new machine. Shes currently looking at the Intel Core 2 Quad since it just had a great price drop (its $300 at newegg) and may be doing some vmware stuff. I'd love to tell her to wait for AMD bench marks but none are out, there is no release date and even if we had all of that she doesn't want to spend that much on her system so knowing the price would be nice(it would suck to wait till September only for Barcelona to be $800).
Intel needs more then just true quad core. They also need memory controllers build in the cpus and a cpu to cpu link that does need to use the NB to talk the other cpu and also have so you can have more then one NB like chip like you can on a amd system. With a 2 way amd system you can have 2 chip set links and up to 2 HTX slots.
Intel basicly took a big hammer to any AMD claims of "more affordable quad-core" with their cut today from $530 to $266 for the cheapest quad core, which I doubt AMD can do much better than. I also don't expect them to top the QX6850 for performance right off the bat, since they clearly fail to do so in dual core. AMD is bleeding a lot of money right now and Intel knows to push when it hurts. Right now AMD is staying competitive but with the massive cuts to margins it can't be good for neither profits nor R&D. Intel is not going off on a huge strategic blunder like the PIV or Itanium again, this time they're on the ball and overclocking results suggest they have a lot of headroom.
The latest batch of ATI cards have failed to compete with the 8800GTX and instead compete against lower clocked cards, presumably again with cut margins. Right now AMD and ATI to me look like two second place companies, and if they try to integrate closer they'll drag each other down. I'm certainly not inclined to buy those two as a package...
Live today, because you never know what tomorrow brings
Finally I can run four things at once on Windows without a block.
Why UNIX?
This is close to press BS as you have to put this into perspective.
- Intel is the 800-pound gorilla, and AMD (compared to Intel) is just a mosquito.
- Three years ago AMD use to have the best top-of-the-line processor for, and kept the performance crown for over a year, that did not changed the roles Intel and AMD have.
- Getting an alternative to Intel is good thing per se, and the press should not push a development to crush and burn even before it is out to the market.
We have yet to really harness the power of two cores, so why all the fuss over four? How many multi-threaded programs are there anyway?
622677120
Guess what? The market doesn't give a shit, they just want multiples of 4 in one socket, period. Even AMD admits it was a mistake not to go MCM; Intel got the drop on them, and has deepened their lead quite substantially, leaving AMD sitting on their hands with no competitor for far, far too long (and their upcoming competition will quite frankly devastate them in the short run, however in the long run...).
Intel had the option to rest on its laurels; they don't like to work any harder than necessary to remain on top, and the Core marchitecture gave them a huge.. well I'll say it.. "Leap Ahead" of the competition. Unfortunately, Intel's more of a bunny; hop a few times then get tired and sit around, whereas AMD is more of the turtle (slow to market, but rather constant). Well all know who wins the race.
"Victory means exit strategy, and it's important for the President to explain to us what the exit strategy is." G.W.Bush
"True" QC, "fake" QC, what does it really matter? The only things that really matter in the end are performance and price (and possibly power dissipation). From the standpoint of a consumer, the internal technology has no importance at all.
Now, if you said that "true" quad core was going to make the chips be twice as fast as Intel's, at half the price, then that would be interesting. Of course, you could say that the chips would twice as fast at half the price, and that would be just as interesting - the technology has nothing to do with it.
Ewige Blumenkraft.
Geeks love to work themselves in to a lather over technical difference but to the end user, quad core is just 4 processors in a single socket, doesn't matter how it is delivered. Now if the Intel solutions performs poorly because of the 2x2 design then it could be a problem. However, thus far, it doesn't seem to. On the kind of apps that can use the power (like a 3D renderer for example) they just shine.
In the end it doesn't matter how it is delivered, it matters who can deliver the good performance per $$$. Intel's quad core chips go a long way to doing that in the markets that can use them. The reason is it gets expensive to add physical processors to a board. A single socket board might be $100, but the same thing in a dual socket variety can be $400-600 and you don't even want to see the prices on quad sockets. Thus being able to drop 4 cores in to a standard desktop board, even if they aren't a monolithic 4 core package, is a good deal for many.
Technical arguments and contrived benchmarks mean nothing. The only things that matters is how fast it runs the things you actually, really do, and how much it costs.
If amd where to go MCM there better cpu to cpu and cpu to chip link the hypertransport bus will make things a lot better they the way intel does it.
Never the choice?
:)
Intel has sold 1 million quad-cores (4 cores in a socket) already.
AMD has sold approximately 0 quad-cores to date
I guess nobody told those people that they should have waited...
... it NEVER MADE _true_ QC CPU...
All existing Q6xx0 solutions are dual-dual core ie two dualcores sharing same FSB - and that is _NOT_ the same as true QC as Barcelona is claimed to be.
That difference is enough to make Barcelona the main choice for many core servers even if it were made with old K8 and not the new K10 cores.
Intel should have true QC chips in a year or so...
You're very convinced the difference will be drastic, that's very funny thing to be when you never saw a single benchmark.
According to Intel's engineers there's no identifiable bottleneckeck in a design that uses two 2-core dies on a single chip. And after all, if two dies was so terribly horrible, people would use multi-processor systems long before we saw multi-core systems.
Intel is moving to single die 4-core processors, sure, but their technology, design and process is better in all other respects, which COULD easily be enough to leave AMD in the dust.
People won't buy a 4-core processor to make an art statement, they'll get it to do work, so they can't care less if it's a "true" 4-core cpu or just.. well.. is a cpu that has 4-cores.
1. it all depends on task at hand. With tasks with intensive intercore communications true Barcelona could be say 10x faster than anything Q6xx0 on Intel's side even without superior core.
2. If your general understanding of the problem is poor than any explanation that could be "interesting" to you is likely to be marketing bull**it, optimized for technical morons.
You can't make universally valid "X-times faster/slower than" comparisons between these kind of machines.
Results tend to be program-and-load specific...
Not true.
If AMD wanted to, they could have hads Intel's style "quad core" long ago.
Hell, even two "x2 4800" dies on one substrate, connected through HT link would be an equivalent, and they could do it in a few weeks even if they would decide to go for it _today_. There is not much to it.
Opteron/AMD64 was _made_ so it could be connected it LEGO-like fashion...
I am very "convinced that performance difference will be drastic" since:
...
1. This is mainly what currently holds Opterons over Xeons on servers, despite superior C2D core and heap of cache.
2. I have exchanged dual Opteron boards for single socket DC 6000+.
Despite HT link and dual RAM bank of existing Opterons being superior for most uses to Intel's shared FSB, there is tangible speedup just due to having really fast intercore communication path.
3. AMD has onboard memory controller even now with K8 and K10 will bring further optimisations there.
4. There are further optimisations for VT, which will bring ie. speeds of WinXP, running under VmWare on my Gentoo Linux (64 bit) much closer to native speeds...
5
"According to Intel's engineers"
Hahahahahah, please don't try to kill me, like you just did with your credibility.
AMD have indicated they may do MCM in the future (Multi-Chip Module), like Intel. But since they are releasing a true quad-core CPU before Intel, it is going to give them an advantage: to make an 8-core CPU, they will just need 2 quad-core chips, whereas Intel will need 4 dual-core chips.
I wonder how significant this technical advantage really is on various levels: performance, power consumption, reliability, yield, simpler to manufacture, cost, etc. Could this also mean AMD will be first to market with 8-core CPUs ?
Or you can't use all your memory sockets because the memory controller for half of them is in the second potentially non-extant CPU.
Balancing the amount of memory on each side is a good idea too.
Help stamp out iliturcy.
She's unlikely to tax either one.
She'll need a lot of RAM for VMWare to work well. That will have a huge impact on your cost calculations. Use RAID for performance, she'll need that too.
Processors are important but they're not the whole answer in the performance equation. At this time the bottlenecks tend to be more in the RAM and HD I/O.
Help stamp out iliturcy.
I sincerely doubt that anyone still in school - any school - is going to overtax the current Core 2 Duo/AMD-64 X2 offerings available today. Short of running simulations of the Universe in real-time, or high resolution Maya renderings (remember when Photoshop was once the app that justified the most powerful machines, and AutoCad before that?) before class is over, that Quad Core is going to be performance, and wallet, overkill.
She may want it. She may actually get it. But I truly doubt she really needs it. Feel free to point out what heavy duty computing she is getting into to justify this. And VMware isn't the answer. It actually runs quite well on current dual processor rigs with enough memory.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
Actually, intel has sold 0.000 QUad cores (= 4 cores on a die in a single unit)...
And if you count Q6xx0 as quad core system, then by that standard qualifies every dual socket dual core AMD system sold, since architecturally they are about equivalent, if not better ( 2x memory channels, HT links)...
You have reason!.
1. Windows x64 is useless. This OS runs slower in AMD64 or EM64T than Windows XP.
2. Windows Vista is slower and buggy than Windows XP.
3. AMD64 64-bit mode only is used to run OSes Linux, FreeBSD, NetBSD and OpenBSD.
4. The 64-bit and 32-bit applications run faster in Linux or FreeBSD than in Windows XP.
5. The Linux and FreeBSD community can develop applications for different targets of CPU.
Dear President of AMD's CEO,
Can you read my proposal?
1. Today, is more important to follow the x86 compatibility of your AMD's CPU for the Linux/*BSD people?
1.a) Why don't release a new CPU of 64-bit with less archaic instructions and more new 64-bit instructions of complex modes of addressability?
1.b) Why don't use 8 KiB pages instead 4 KiB?
1.c) Why don't remove x86 32-bit instructions? It eliminates ugly problems of 3264 bits system calls.
1.d) Why don't double the pins to double bandwidth?
1.e) Why don't double the number of DDR modules to 8 instead 4 to permit 8x1 GiB or 8x2 GiB?
1.f) Why don't follow the K.I.S.S. principle to don't lose time making complex tasks!!!
2. It's not important to improve the performance of the poor Instructions System Architecture (ISA) of AMD64. It's more important to release new ISA to improve gainly its performance maintaing the same principles of the previous design.
3. You must to change the timelines. The time is precious in this hard race that Intel is winning.
Change the plan in this time to don't lose the lost market and to re-born a new era of CPU 64-bit race.
I'm currently doing research with MRI imaging on multi-core systems and she may be joining in. Plus we're Gentoo users ;)
The reality of the market is that Intel killed the 4 socket market by making a 2-socket system behave like a cheap 4 socket system. This trend will continue.
AMD lost the opportunity to grow the 4 socket market by failing to drop prices sufficiently. That is where their greatest advantage is (system architecture - HT, etc...).
Now Barcelona is late and underperforming (2GHz at launch with scaling to 2.4 by Q1 2008)
This is not enough to fight off Intel's 45nm offerings, especially the newer server platforms on deck such as caneland and tigerton.
By the time AMD finally scales Barcelona to a respectable frequency, they will have to deal with Nehalem which IS a native quad core with on die memory controller and direct-connect links (CSI). Furthermore, Nehalem will have 8 cores AND SMT.
The future is looking pretty grim for AMD in server-land.
They do however appear to be making headway in the low-end of the laptop space which is a fast growing market. But...they need the high end to pay the bills. Gaining marketshare at the expense of profits is a recipe for disaster.
AMD could get 99% marketshare if they priced all of their processors at $0 for example.
If AMD wanted to, they could have hads Intel's style "quad core" long ago.
And yet they don't, and they just posted a $600 MILLION loss in one quarter. The difference between what AMD lost and Intel made last quarter is almost 2 billion dollars. Maybe you should take your market genius over there and help them turn it around.
Education is a better safeguard of liberty than a standing army.
Edward Everett (1794 - 1865)
Umm...a 2 socket Q6xx0 system is 8 cores.
A 2 socket AMD Opteron system is 4 cores.
Intel killed the 4 socket market for AMD. This used to be a huge cash cow.
Why would you pay for 4 sockets and get 8 cores, when you can get a MUCH cheaper 2 socket systems.
The 2 socket (8 core total) systems are more dense and give you greater throughput per dollar - even with an FSB - thank Intel for their gigantic caches which help to reduce FSB traffic more than an opeteron might.
"Barcelona could be say 10x faster than anything Q6xx0 on Intel's"
BS: At what tasks?
Because Intel has already demonstrated near 4x performance with it's "untrue" quad core. You are not going to get more than 4x single core performance. There is only a tiny margin of efficiency of multi-cores that AMD could improve upon.
AMDs only real Barcelona hope is if it increases basic core performance significantly, there is no magical "real 4 core" performance leap to be had certainly not a 10X performance increases. That is impossible.
It used to be that ATI had better hardware, and Nvidia had better backroom deals.
Things have changed. The only thing ATI/AMD offers open source is the middle finger. AMD/ATI has done virtually nothing except add user-hostile features and low-level DRM on the framebuffer level.
I won't be buying from them anytime soon. They need to get on the ball and lead, not just throw slipshod proeducts out just to try to keep up. AMD needs to take a charge off the quarterly earnings, and do some R&D in order to survive.
IMHO, even VIA is starting to outshine AMD/ATI.
1) What Linux/BSD people need is irrelevant, the market share is too low. (I run WinXP, Debian and FreeBSD, for the record)
1.a) Complex addressing modes are only good for people writing software in assembler - few are needed by C-compilers, and those are provided already anyways. They're also more silicon space that could be better used for wider ALUs/FPUs or more cache.
1.b) 4kB/4MB pages are already available. 8kB, while marginally nice, isn't really necessary.
1.c) 32-bit instructions are just being translated to micro-ops, just like 64-bit instructions during instruction fetch, the benefit would be marginal, while losing 32-bit compatibility.
1.d) Doubling pins means more layers in motherboards, skyrocketing the manufacturing costs. It's possible if you're ready to pay 300-500$ per motherboard, to produce those 10 layer monsters. It would also mean another new socket type.
1.e) Because the memory controller is on the CPU (on the die), this would also mean more traces on the PCB -> radically higher costs.
1.f) AMD-64 keeps things simple, at least compared to some unnamed competitor. *cough* Itanic *cough*.
2. New ISA is often a suicide. Ever heard about Itanium?
3. Biggest blunder AMD has done recently is dropping Socket-939 prematurely.
What know you about the EPIC technology after the unsucessful VLIW and sucessful multi-core speculative-superscalar?
* The problems of VLIW were the excesive NOPs and its stalls. So, the VLIW technology is not the solution of today.
* We're 7 years later of the start of 3rd Millenium, it's a poor race of an American. Because of the stupidity of mea culpa.
Perdón por mis pecados, tenía que haberle dicho a AMD que tendría que haber hecho una nueva CPU de 64 bit diferente claro que puede reinventarla! de tal forma que puede llevar 8 IPs (Instruction Pointers) integrados en el scheduler x N cores para multithreading and/or multiprocessing en vez de complicados y lentos 8xN cores x1 IP para multiprocessing.
Las ventajas son el permitir la ejecución paralela del mayor número de unidades enteras y de coma flotante aprovechables por cualesquiera de las IPs que necesiten. Y además permitir la cooperación de información entre los registros de las IPs integrada en el complejo scheduler.
Los sistemas multicores tienen el problema de la consistencia de memoria y coherencia de las caches. MOESI? Dragon?
Pero también tienen otro problema, no están aptos para sistemas multithreading porque sus IPs no están en la misma caché sino en diferentes cachés, y las hebras comparten el mismo espacio de memoria y por lo tanto de la caché.
Of course, Intel is also rushing out a similar solution, in the form of their V8 programme. So, it is a race to see which company will be the first to release an 8-core platform. They talk of putting two quad cores on existing dual core systems to get 8 core platform. This was done over a year ago on Intel quad cpus. That race was over long before the article.
running VISTA + anti virus + background apps with rendering / cad or video work can tax a system even more so with some of carp that some schools make you run on your system can tax it.
Games are a other thing that can push systems also it not just the cpu that gets taxed in is also the ram, IO , hard disks , network, video and so on.
They're called errata. The most recent bunch are more plentiful than usual but it's not unheard of. Get your microcode updates, whichever vendor you get your chip from. AMD calls them BIOS updates which partly makes sense since you usually patch the BIOS at the same time. You get them from the OEM of your motherboard or system usually but as you see from those links operating system vendors can put them out too. The errata that have been in the press lately are unlikely to affect chips you buy right now because new chips and systems will almost certainly have their BIOS and microcode updated from those issues before they ship.
No computer is future proof. You can get some extra months on one by buying above average, but the best desktop you can get today will still look sad in three years. Pay extra for bleeding edge if you want to but the best value is middle of the road and frequent upgrades.
Help stamp out iliturcy.
Vista 64-bit and Xp 64-bit work just fine, thanks. I'm running Vista-64 right now. We make a fair bit of use of both at work, for precisely the reason that we have apps that run out of 4GB of RAM.
As for your drivers comment, well let's see here: Intel has 64-bit XP and Vista drivers for their motherboards (and by extension graphics) and NICs as far back as their 865 series (anything older doesn't support 64-bit CPUs). Vista-64 has native support for older nVidia chips (GeForce 2 is the oldest I've tried) and nVidia provides downloadable drivers for their 5 (FX) series and newer. ATi likewise has support in the OS for some older chips, and downloadable drivers for the 9500 and newer for XP-64 and Vista-64. Broadcom has XP-64/Vista-64 drivers out for all their NICs (both 44XX and 57XX series). LSI has 64-bit drivers for, well, all their products that I can see for XP and Vista (and Linux and Solaris). Colorvision has 64-bit drivers and is Vista compatible. Logitech, Microsoft, and Saitek all have 64-bit drivers and support apps out for their input devices.
I could go on but basically any modern hardware seems to have no problems at all with 64-bit drivers. In fact, on all the 64-bit Windows systems I've set up, I've never encountered a component we didn't have a driver for. I'm not saying there aren't some oddballs out there, I'm saying that the vast majority of stuff DOES have a driver and thus it is a non-issue.
When you are countering some FUD, please don't spread your own. You may to like MS OSes, that's fine, but it is a lie to say that finding drivers for 64-bit Windows systems is hard. The vast majority of devices, including specialty devices (I've got 64-bit Vista drivers for my colorimeter and StudioCanvas for example) have 64-bit drivers. It is just a non-issue. Far more rare is 64-bit software, but thankfully 32-bit software runs without problems on the 64-bit OS.
Oh god these AMD guys just won't drop this.
WE KNOW ALREADY. the intel quad core still performs very well in benchmarks - faster than the dual core in multi-threaded apps, one can only conclude it damn well works.
As for AMD's 'true quad' performing better, no one knows for sure, there's no real benchmarks yet at this point.
AMD could release it and say 'wow, our system is 25% faster than the Q6600" and intel could say 'err so what' and release the Q6600 clocked 25% faster because there's so much damn headroom on their chips at the moment.
I'm a fanboy of neither company but right now AMD fanboys claiming "K10 will rule all!" and "Barcelona 'rapes'!' etc are totally foolish, NO ONE KNOWS YET.
Finally there's the efficiency of each core itself.
Look at the Intel C2D @ 2ghz vs the X2 3800+ at 2ghz - both dual core but one of them is substantially faster than the other - efficiency on each cpu will make a difference.
A couple clarifications: I was referring to the DOS extender technology, and the x86-64 article does indeed refer to it being "possible" to enter x86-64 "long mode" using such a technique.
Is anyone actually doing this for a non-turnkey application?
Any sufficiently advanced technology is indistinguishable from a rigged demo. -- James Klass
... but alas it's the Q6600 for me as a) Intel just drop kicked its price, and 2) AMD are still nowhere near releasing quad cores for the consumer market. "Later this year" just doesn't cut it, so here's yet another person switching from AMD (Newcastle 3200) in their desktop to Intel. :P
ISO certified == THX certified
The Doctor: Rose Tyler, I was gonna take you to so many places. Barcelona. Not the city Barcelona, the planet Barcelona. You'll love it, fantastic place, they've got dogs with no noses!
[laughs]
The Doctor: Imagine how many times a day you end up telling that joke, and it's still funny!
Well, that aside but because it had to be done. But to business. I remember when 64bit cpus came out Sun and a few others had them (12 years ago) the programming took forever to catch up. They are still selling 32bit processors as new. I think I will leave that hanging in the air.
Sorry about the writing. Robot fingers, you know? Cliff Steele in DOOM PATROL #23
A fable is only a fable, and what's worse, this race is supposed to run forever. Surely AMD can show their persistence and diligence despite their past ill judgments and mistakes, Intel for now has the upper edge.
Slashdot me with L$s!
I have an AMD X2 3800+ (s939), 2GB of ram and an nVidia GeForce 8800GTS all in an nForce 4 motherboard, and drivers for all of it are available for both Windows XP x64 and Ubuntu x86-64.
My 250GB Windows disk is overflowing with major games, and all run wonderfully. Except for the occasional one I have to crack, either because it's too old and tries to load a 32-bit driver (understandable) or the company went for a shit copy-protection system that tries to load a 32-bit driver (eg Overlord). That's not such a big problem as I crack most games anyway.
My Linux install is reasonably new (Ubuntu 7), and already has launchers for 8 commercial games on the desktop. Admittedly a couple are running in Wine, but they still run well. Neverwinter Nights and UT2004 are examples of native linux games, EVE online and steam are running in Wine.
So far I haven't found anything that doesn't work well on a 64-bit OS. Well, except copy-protection.
Go here:
http://www.fftw.org/
Download and compile it (and get your free compiler here, because Sun Studio 12 is 5% or more faster than GCC even on Linux...). Do so on both AMD and Intel chips.
Now, benchmark both architectures.
FWIW, my 2.4GHz Opteron runs FFTW about 30 to 40% faster than my 3.2GHz Xeons. Even when the Opteron is running the 32-bit binaries compiled for the Xeon.
PS - when you compile FFTW, make damn sure you disable ALL inlining...
I think it's important to remember that Intel inadvertently delivered the high-end server market into AMD's lap.
Intel had done so much heavy marketing, pushing claims that the Itanium was going to blow away all the proprietary CPU architectures, that damn near EVERYONE EOL'd their Unix servers... Alpha, MIPS, PA-RISC, etc., etc. Intel and Itanium made them announce the end was coming, and then when Itanium turned out to be the biggest flop in history, companies around the world had nowhere to go, but perhaps POWER and SPARC.
Then, into this vacuum, appeared AMD's x86-64, which had great performance, and everything needed for the high-end server market. AMD ate everyone's lunch (and got a bit more comfortable than they should have).
And also interestingly... You can't run your old proprietary Unix OS on AMD64 hardware, so what do you do? You switch to Linux. So it got a boost out of Intel's stupid marketing blunder as well.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
but think your information is a little dated. For many years what I said was true. Then for about 18 months it wasn't. I believe the race is back on now.
Today's quad core, vm supporting 64bit machine will be quite useful in 8+ years, especially if you get the ULV processor. But compared to what's on the shelves at WalMart on that day it will still look dated.
What I wonder is what sort of hideous application would require the type of computing horsepower that should be available 8 years hence. By then what's that process supposed to be? 21nm? 24? I trust the rocket scientists at the chip companies will continue to come up with technology to astound us.
Help stamp out iliturcy.
$300 is not an unreasonable amount for someone to spend on the CPU in a system that they think will need horsepower. It's when you go beyond the $250-$320 price range that prices escalate rapidly compared to the performance increase. I would dissuade most users from spending $600 on a CPU, but wouldn't be as negative towards someone in the $250-$300 range.
Of course, for the budget users, I'd be pointing them at the lowest cost dual-core CPU. Or possibly a few steps up from the bottom. There are some pretty decent dual-core CPUs in the $100-$120 range.
Wolde you bothe eate your cake, and have your cake?
I've always been a fan of AMD, but I agree with your point.
It doesn't matter if it's two dual cores in one package. It doesn't matter if it's 4 single cores with a highly trained monkey dividing up the instructions between the cores. It's the end results that matter.
However, right now you can't get the 300$ quad core from Newegg, it's sold out. You *can* get a 61$ dual core from AMD though. And unless you require a lot of processing power, it's more than enough for most people. Especially college students on a budget.
I'd go as far to say that if you're on a budget, it's a no-brainer. Even if you could use the processing power the Quad core chip brings you, a 239$ difference is huge.
Democrats or Republicans. They are both taking us to the same place and they are not afraid of us anymore.
Intel will make a 9 core chip. 8 cores for the OS to use and the last one for the memory controller.
**Intel if your reading this, and do this I want a percentage.
**** laugh, it makes you feel better
AMD has had an 8-core platform since the 8x Opterons came out. They had a 16-core platform with the introduction of dual core CPUs a couple of years ago, and they'll have a 32-core system when Barcelona comes out.
But why not say Intel was "first"? They were the first in the x86 space to slap 4 cores into a single socket package. Let's wait and see how AMD's solution actually performs when it comes out. Either way, it's going to be interesting and prices should drop, so we should all like both Intel and AMD for their continuing competition.
I still think AMD has the better server side solution by far, but for the consumer side Intel is the better bang for the buck since C2D came out.
The cesspool just got a check and balance.
The big x86 problems are:
- Inadequate register set. Mostly addressed by AMD. This has required more L1 cache ports and fancier stack hardware, but okay, not too big a performance penalty.
- Complex instruction decoding. This is the biggie that I could see a redesigned instruction set solving. You could rip a couple of pipeline stages out of the front end, which would reduce branch misprediction penalties and improve cold-cache performance.
It's the decoding latency that I see as the major remaining x86 liability. The entire P4 trace cache mess was a (failed) attempt to reduce the branch misprediction penalty, but it sucked as soon as you escaped the L1 cache.It's probably not worth it, but even without major VLIW rearchitecting, there are ways to improve on the x86.
What sort of problems are you encountering? I'm not having any.
Only a fraction of the 64-bit capable desktops and workstations are running 64-bit applications. What's more, there are very few mainstream 64-bit applications out there, despite the fact that for gaming, audio processing, image processing/photoshop apps, and video there would still be performance advantages (memory bandwidth and operation throughput), even if you don't yet have > 4GB of RAM.
It's not really compelling - plus or minus a few percent. And you need to test two binaries which is expensive. So unless you're absolutely forced to use more than 4GB per process, I think people won't bother.
Looks pretty compelling from here! Take a look at the 32-bit vs. 64-bit benchmarks shown here for instance! If I was a game developer, I'd gladly take 40FPS over 30FPS if it only meant a recompile targeting a 64-bit platform! You only need to test two binaries if you also choose to support 32-bit as well - a suitably advanced app/game could just make a 64-bit capable AMD/Intel chip a prerequisite these days (DOOM required a 386 or better during a similiar 16/32-bit transition period)
"Sherman set the wayback machine to the early 90's, and the great 16-32 bit transition" You might recollect the introduction of the mighty 386 processor, extended memory modes, the Win32s API, but probably most importantly, killer apps like DOOM loading up their own 32-bit memory managers to sidestep the OS, which really wasn't ready to provide good 32-bit native support. Apps that did this completely took over the system, putting the host OS in stasis until the app was exited.
Sounds much like the same situation - the majority of users are running an OS that can't tap the full potential their CPU has to offer. So - why aren't we seeing similar application tricks, like those that enabled 32-bit protected mode now? The proposition of writing an application which would sidestep Windows XP 32-bit and set up a mini 64-bit host environment (not really a full OS) is not that radically different, right?
If you really want 64 bit, I don't see why you can't use Windows x64. Sure you'll need to be careful that you have hardware which has x64 drivers, but that's life.
*I* can (and regularly use XP-64, W2K3 64-bit, and even Vista 64-bit these days) - but if I'm writing a high-end app/game for a reasonably wide audience, you have to realize there's a lot of Windows XP 32-bit boxes out there running on CPUs capable of running in 64-bit "long" mode. It would be mighty sweet to tap into that power for high-end gaming, audio/video processing/transcoding, ray-tracing, etc.
32 bit Windows already has PAE which is the moral equivalent of a Dos extender. I think Outlook and MS SQL server can use it. So there isn't really a hole for a 64 bit Windows extender.
PAE only gives you access to more memory, it doesn't enable the CPU 64-bit processing, so it's not interesting to me. What's more - programming to these "bank switching" style memory extensions really is quite cumbersome.
I've often wondered what would happen if you could make bootable games - e.g. Linux+ATI and NVidia drivers+a game binary on a LiveCD. But to make it work you'd need to be able to offer much better graphics performance than regular Windows, just like Doom's extended Dos had better performance than regular Dos.
And given the amount of effort NVidia and ATI spend on Linux drivers compared to Windows ones, I'm not sure that's the case. DirectX is thinner layer over the driver than OpenGL too.
Yes - using a stripped down Unix core could achieve the same goal, I suppose. Again the DOS4GW dos extender did some mojo regarding drivers such that it leveraged the existing 16-bit DOS drivers for a number of s
Any sufficiently advanced technology is indistinguishable from a rigged demo. -- James Klass