AMD Breaks 1GHz GPU Barrier With Radeon HD 4890

MojoKid writes "AMD announced today that they can lay claim to the world's first 1GHz graphics processor with their ATI Radeon HD 4890 GPU. There's been no formal announcement made about what partners will be selling the 1GHz variant, but AMD does note that Asus, Club 3D, Diamond Multimedia, Force3D, GECUBE, Gigabyte, HIS, MSI, Palit Multimedia, PowerColor, SAPPHIRE, XFX and others are all aligning to release higher performance cards." The new card, says AMD, delivers 1.6 TeraFLOPs of compute power.

Comments

It Was Epic

[eldavojohn]

AMD Breaks 1GHz GPU Barrier

I was diligently working at XYZ Corp a few buildings down when Incident One happened in their lab. At first, I was just sitting in my cubicle when suddenly we felt a severe shuddering of space & time around us. Then a few seconds later everyone heard a loud "Ka-BOOM" and everyone stood up to see what was going on outside. The buildings directly adjacent to the AMD lab had all their windows blown out and every car alarm within a square mile was going off. Some scientists with their hair blown straight back and carbon scoring randomly on their faces and white lab coats were seen to climb out of the rubble of AMD's R&D building. They immediately began dusting themselves off, high-fiving each other and patting each other on the back laughing and ecstatic. Then they headed towards the liqueur store down the street to pick up some champagne. Shortly after it was discovered that 1Ghz is the frequency at which æther vibrates when it is at rest so once you pass it, you leave a wake of æther behind your time cone. Roger Penrose and Stephen Hawking are due to give a speech at "GPU Ground Zero" this week, I hope to make it.

If I were working marketing for AMD, I would be pointing out how switching from base ten to base eleven, twelve, thirteen, etc provides a theoretically unlimited amount of newsworthy advertisements in broken barriers. "We just need to make it to 2,357,947,691 hertz and we'll be the first to claim we've broken the 1 Ghz (base11) barrier! Where the hell was the report that we broke base9 last year?!"

Re:It Was Epic

[FredFredrickson]

base9 wasn't really that much of a feat. Not to mention, the class action law suit on differing bases really put a damper on that party.

Re:It Was Epic

[ArcherB]

Shortly after it was discovered that 1Ghz is the frequency at which æther vibrates when it is at rest so once you pass it, you leave a wake of æther behind your time cone.

Wow! And here I thought it was 1.21Ghz at 88 MPH.

Re:It Was Epic

[Anonymous Coward]

oh god, this was seriously the best post i have ever read on slashdot. thanks for the great laughs! :D

Re:It Was Epic

[Cornelius the Great]

Great Scott! Don't you mean 1.21 JHz (jigahertz)?

Re:It Was Epic

[yabos]

jigawatts

Re:It Was Epic

[Anarchduke]

AMD Broke the 1 GHz barrier on their CPU, and now they break the 1GHz barrier on their GPU.

It doesn't matter what base you use, AMD owns that achievement.

According to AMD top researchers, whether it was base-9, base-10, or base-11 doesn't matter. According to AMD,

"All your base are belong to us."

Re:It Was Epic

[binarylarry]

jiga what?

Re:It Was Epic

[tlhIngan]

Great Scott! Don't you mean 1.21 JHz (jigahertz)?

"Giga" in some countries is actually pronounced "jiga". (History says that is how "Giga" is pronounced everywhere except the US, but that's debatable). Thus, 1.21GHz would be an accurate figure in this article.

Re:It Was Epic

[Anonymous Coward]

boo!

Re:It Was Epic

[fbjon]

I would think history says it's pronounced with a hard 'g', specifically greek history.

Re:It Was Epic

[Anonymous Coward]

Giga comes from the same root as "giant" and "gigantic" so properly should be using the j sound. Unless we're all pronouncing giant and gigantic wrong, as we do with Caesar (The Germans have it right by pronouncing it Kaiser)

Re:It Was Epic

[fbjon]

That's precisely the point: English pronunciation is so screwed up, there's no point in getting riled up in trying to harmonize it anymore. I pronounce it with a 'g' in English because that's how it's pronouced in all other languages I know.

AMD CPU too

[Devistater]

Didn't AMD break the 1ghz desktop CPU "barrier" too? ;)

Re:AMD CPU too

[LoRdTAW]

Digital Broke that with the DEC Alpha (Was it DEC at that time?). Wasn't popular but it was a desktop CPU for high end workstations.

Re:AMD CPU too

[LoRdTAW]

Sorry. It was Compaq who owned the Alpha at that time. It was still DEC who designed it though.

Re:AMD CPU too

[Ceseuron]

I think it was IBM that broke the 1Ghz CPU barrier.

It's not really a huge feat to break the 1Ghz "barrier" for a GPU anyway. And since it's AMD product, it'll run hot as hell and require a massive heatsink. You'll be able to barbecue a steak on your CrossFire enabled rig with two of these installed. Since it's also ATi, the card itself will be really awesome but the drivers released will be buggy and unstable, turning the card into little more than a giant red paperweight.

Re:AMD CPU too

[Anonymous Coward]

How is this a troll?

ATi's drivers are buggy, unstable and forget about Linux support. Where nVidia supported all their cards when Kubuntu 9.04 came out ATI did not and as far as I know they still do not.

The OP is 100% correct, ATI could make a 12 Ghz chip but it is still worthless because of their drivers and lack of Linux support.

They need to quit giving Fanbois modpoints

Re:AMD CPU too

[hairyfeet]

Actually as a PC repairman I can tell you the "trick" with AMD, and it is this- always buy a generation or two behind. I have sold many ATI and Nvidia cards as well as AMD PCs with ATI chipsets, and as long as you stay a generation or two behind you're good to go. My dual core Kuma with 780V chipset is solid as a rock

So what I tell my customers is this: If you want to spend top dollar and be on the bleeding edge, go Nvidia. Their drivers will be rock solid even for the card they just released. With AMD/ATI always get a generation or two behind and NEVER upgrade the drivers! Unlike Nvidia whose drivers are pretty painless to upgrade, upgrading to the latest Catalyst drivers usually end up bring nothing but instability and headaches. Now I don't know if this "trick" work with Linux, as I'm a Windows only shop. But I have found in Windows if you follow this rule you'll be good to go and save a few bucks as well. The "bang for the buck" ratio is very good on AMD/ATI which is why I just built my first AMD PC since the old Barton Core. You just have to be careful not to get too close to the bleeding edge with ATI, as out of the box their new drivers always suck.

Re:AMD CPU too

[Dyinobal]

I've got to say I don't share your experience at all. Beyond the little cpu bug fiasco I've never had an issue with AMD and I wasn't even caught up with that bug as I adopted slightly later when they released revised phenoms. Honestly AMD don't have a bad track record. I see you're recommending from personal experience though, I'd just have to say I'd go the other way.

Re:AMD CPU too

[TexNA55]

Nope, it was later publicly released that it was PR-rating1GHz barrier they broke.... >:p

Re:AMD CPU too

[Kneo24]

I would say that your customers need to go elsewhere since you aren't staying up to date (and it appears that's what you should be doing for your line of business!). What you're describing is no longer true. Not only is NVidia having a hard time with their drivers, their cards that are a generation or two behind are dying left and right. If you had bothered to take notice, which you clearly haven't, ATI's drivers have greatly improved since AMD became their new overlords.

The only thing you have correct is that if you go AMD, you'll get the "bang for your buck".

Re:AMD CPU too

[Anonymous Coward]

Rubbish, updating drivers is important with ATI products.

I run an ATI XPress 200M chip in my laptop and I can tell you now that when I bought the thing it had next to no openGL support. After AMD took over they released a new catalyst legacy driver for my chipset and now it runs openGL games, granted not well, but essentially a driver update solved my problem.

If there drivers are rubbish out of the box as you claim, then why would you not update them?

Re:AMD CPU too

[hairyfeet]

Uuuhhh....you DO realize you read my post EXACTLY ass backwards, right? I said get NVIDIA new, and AMD/ATI a generation or two behind. It usually takes about 6 months for ATI drivers to give you the full bang for your buck with any stability, and by then that card is one generation behind (and about half price). NONE of my customers got burnt in the Nvidia solder fiasco, because I DO keep up to date. I always troll the ATI forums and with the release of a new card you get nothing but complaints about drivers.

So take my word as you will. I don't discriminate on brand and am not a "fanboy" of any company. I just go with what works. I have been through the "fun" of trying to figure out which Rage driver will make a customers PC work without a BSOD, I've dealt with the "leaf blower" Nvidia cards and had to tell customers that a MX4000 is NOT the same as a Geforce 4ti. I have seen the good, the bad, and the ugly from both Intel and AMD, from both Nvidia and ATI. And that little "trick" I listed is what has worked for me for the past 3 years. Maybe ATI got their heads out of their asses and wrote good drivers for their bleeding edge cards now, I don't know. But considering the trouble that I have had with ATI drivers in the past and the fact that my rep is on the line, I don't think I'll chance it. Hell even my oldest boy, who I picked up an ATI card for because it had better "bang for the buck" had XP BSOD while trying to install the drivers. That is the first time I have seen a BSOD in XP in years. So no thanks, I'll stick with my system for my customers.

Re:AMD CPU too

[Xest]

No, AMD don't have a bad track record you're right, however ATI do, I can't comment on them now though.

I was in tech. support for 7 years from 2000 to 2007 and the fact is that ATI drivers were consistently crap. Our team supported over 10,000 systems, ranging from low end desktops, to high end workstations and big high end laptops to small netbook type systems.

Time and time again the only cards that provided driver issues were ATI, it was absolutely horrendous how bad ATI's drivers were, they've simply never been up to scratch historically and if they couldn't get their act together in releasing stable drivers in over 12 years between releasing their first 3D accellerator cards and me finishing dealing with them in tech. support I'd be surprised if they've made any strides since then, even if AMD has acquired them, judging by comments from the tech. support guys where I work now they really haven't, but I wouldn't like to say that for sure without first hand knowledge.

It's easy to see things differently one way or the other if you're only basing your view on your home systems, but if you deal with thousands of systems like I have, and like this guy you're responding to probably has in his job then you'll find ATI cards have always been consistently far worse in terms of drivers over the years.

Re:AMD CPU too

[Kneo24]

Uuuhhh....you DO realize you read my post EXACTLY ass backwards, right?

Nope, just merely pointing out that you aren't up to date on what's what.

Getting NVidia brand new is fine, if you want to replace your card in two years, which a lot of people do not want to do. Look at the failure rates now on the 8 series and 9 series. If they're using a certain chip, they have a high failure rate. I see people on gaming forums complaining about it all the time, running going, "this game killed my graphics card!"

And to go one step back, in Vista's early life time, NVidia drivers took up a large portion of the pie when it came to BSOD's. Talk about "quality" driver support there, huh?

These are things you're clearly neglecting.

I would also say that if you're getting a BSOD on a matured OS like XP while installing drivers, you're probably doing something wrong. No one ever wants to admit they're doing something wrong, but you usually are. There's a note telling you to shut off all programs for a reason. There's also the chance that someone forgot to make sure the previous driver was completely uninstalled.

If you're that worried about your reputation, fine, stay in the dark ages. You will lose relevance.

Re:AMD CPU too

[somersault]

I always troll the ATI forums

Heh heh.. I assume you meant trawl but it kinda works with either word ;)

Re:AMD CPU too

[Dyinobal]

I only recently addopted ATI cards over Nvidia. The only reason I didn't use ATI cards prior was due to poor drivers yes. How ever I decided to give them a try since AMD took them over, so I waited a few development cycles and got myself a 4850 card. Works great.

Re:AMD CPU too

[skintigh2]

I have bought several ATI cards over the years, and not once could I get the drivers that came with the card to install. I always had to download newer ones and install parts of those, reboot, then try to install the ones that came with it and get a little more to install, reboot, then try the new ones again, and they would bitch that I needed to install Y to be able to install X, but when I tried to install Y it required X, until I manually installed the card as a generic VGA card and then manually installed drivers for a much older card and then started over with alternating between different versions, etc. etc. etc.

My once many-hundreds-of-dollars top-of-the-line 850 XT PE AGP is still suiting me just fine (perhaps the fasted AGP card ever), but I haven't dared try to upgrade the drivers since I bought it years ago. But I might have to, CLI.exe or something crashes at bootup now.

Every nvidia card I've owned installed flawlessly the first time, without 3-30 reboots that ATIs require.

Re:AMD CPU too

[hairyfeet]

The only thing I do with Vista is wipe it and put on XP, so I honestly wouldn't know about that. My customers hate Vista with a hatred I hadn't seen since Bonzi Buddy had users coming into my shop foaming "If you don't kill THIS DAMNED MONKEY right now, I swear to God I'm gonna put a bullet through this damned box!"

As far as the Nvidia "dead chips" issue, haven't run into it. of course I also don't "pull a Compaq" and stuff 4 dozen parts into a mini tower with nothing but a single 80mm fan to cool the damned thing either. If they are spending the cash to get them a really fast card I recommend water cooling or at the very least a 120mm fan at the front AND back of the PC to get maximum airflow. Most of the boxes I build for SMBs end up in some hot dank office in a construction zone so I learned a long time ago the value of going a little overboard on the cooling. Most of my boxes will be lucky to reach 105F and that is running full bore for hours.

As for I was "doing something wrong" installing the drivers? Oh please! The drivers that came with the card sucked ass. I mean you know you're in trouble when Catalyst requires .NET just to fricking run! I had to download 4 separate drivers and try them one at a time (after using driver cleaner Pro each time) before I found a set that worked stable. Go to the ATI forums and you'll see tons having the SAME issue! Are they ALL "doing it wrong"? Like it or not, there is a REASON why ATI has a bad rep on drivers. It is because often their driver really DO suck the big wet titty. So for my customers, where money and my rep is on the line, I will stick with my little system.

Considering I'm staring at a 4 foot high wall of new boxes waiting to be built with customers having paid up front in this crappy economy, and customers are raving to their relatives to go plunk down $500+ to me instead of going to Dell because "His systems run so nice and are so easy and solid", I think my system is working pretty damned well, wouldn't you say?

So this means...

[smooth wombat]

one will finally have a graphics card capable of playing Duke Nukem Forever.

Oh wait...

Re:So this means...

[Anonymous Coward]

This makes me so sad on the inside. Admit it, you are hurt too when you make jokes like this.

Re:So this means...

[DigiShaman]

Don't wait for "never".

Re:So this means...

[cootiesratsemen]

Too soon.

Why is it harder on GPUs than CPUs?

[G3ckoG33k]

Why is it harder to raise the clock frequenceies on GPUs than CPUs? Is more code in use at the same time per unit area, or?

Re:Why is it harder on GPUs than CPUs?

[jgtg32a]

They've never needed to get the clock speed up that high before, remember Ghz != Performance

Re:Why is it harder on GPUs than CPUs?

[KillerBob]

Heat. Because of the form factor, you can't put a massive heatsink on a graphics card, certainly not the kind that you see on high end desktop CPUs.

GPUs are also generally a completely different architecture than a CPU... they're usually massively parallel and optimized for working with enormous matrices, whereas a CPU is significantly more linear in its operation, and generally prefers single variables.

Re:Why is it harder on GPUs than CPUs?

[caerwyn]

GPUs are a little more CISCy.. Since the cycle time is constrained to be as slow as the slowest operation that must complete in one cycle, it means that it's a bit harder to cut down on cycle time.

Re:Why is it harder on GPUs than CPUs?

[djupedal]

> "Why is it harder to raise the clock frequenceies on GPUs than CPUs?

Speed costs money...how fast 'ya want to go?

Re:Why is it harder on GPUs than CPUs?

[mdm-adph]

GPU's have recently become massively parallel -- not as much need to go too fast in overall clock speed.

Re:Why is it harder on GPUs than CPUs?

[zolf13]

Wide vector processing with "800 stream processing units" (or "pipes" or "cores") - it is hard to put 800 cores in one chip and not to boil the silicon.

Re:Why is it harder on GPUs than CPUs?

[Firethorn]

I think it has to due with the massively parallel operations. You can't pipeline stuff as far. Of course, I'm just guessing.

Basically, due to the parallelization it's more efficient to add more streams/'processors' than to ramp up the overall speed of the system - for example, the referenced 4890 has 800.

In order to have all the stream processors work, you might have to be a bit more conservative in your timing.

Re:Why is it harder on GPUs than CPUs?

[mikael]

You have so much data being churned around. The high end GPU's have 240+ stream processors, compared to a handful for a mobile phone. Then there is the constant punting of video data from the VRAM chips to the LCD screens (width x depth x RGB x bits/channel Hertz. VRAM is like standard RAM memory except there is a special read channel to allow whole rows of memory to be read by the video decoder simultaneously as it is being read/written by the GPU. It would be possible to
raise the clock frequency, but they would need a larger heatsink. If you visit the overclocking websites, you will see some of the custom water cooling systems that they have. Early supercomputers like Cray used Fluorinert.

Re:Why is it harder on GPUs than CPUs?

[91degrees]

I think it has to due with the massively parallel operations. You can't pipeline stuff as far. Of course, I'm just guessing.

That can't be it. Graphics cards can have vast pipelines. Pipelines' main problems are with branches, and graphics cards don't need to be able to branch.

Re:Why is it harder on GPUs than CPUs?

[dhanson865]

Yeah you can't put the exact same heatsink on them but take a look at the Accelero S1 Rev. 2 at http://www.arctic-cooling.com/catalog/product_info.php?cPath=2_&mID=105&language=en

You even putting a 120mm fan on it doesn't cover the entire fin area. http://www.silentpcreview.com/article793-page5.html

Yeah with fan it'll be a 3 slot solution and yeah it only weighs half the weight of a high end CPU heatsink but then again that is not their biggest GPU heatsink.

The heaviest solution on AC's site is the Accelero XTREME 4870X2 at 680g which is getting up there for weight on a graphics heatsink. http://www.arctic-cooling.com/catalog/product_info.php?cPath=2_0&mID=244&page=spec

I'd say its more of an issue that pure clock speed only covers some GPU problems. Memory bandwidth/latency, number of GPU cores, design of the cores, programming issues are all more difficult to balance than just ramping up the clock. They could cool these chips better but would it really be worth the cost/effort if the rest of the design and supporting software can't take advantage of it?

Re:Why is it harder on GPUs than CPUs?

[Pulzar]

They've never needed to get the clock speed up that high before, remember Ghz != Performance

Err... It's not that black and white, you can't just say that GHz != performance. If you take a card and raise its clock, you'll usually get more performance. If you raise memory speed you'll usually get more performance. The only time you won't is when the one is bottlenecking the other.

All we're learned from CPU wars is that within two different architectures, the faster one isn't necessarily the one with more GHz. But, between two identical designs, more GHz means more performance.

Re:Why is it harder on GPUs than CPUs?

[Cornelius the Great]

Newer (SM 3.0+) shaders allow flow control, so branching is supported in more recent architectures.

Re:Why is it harder on GPUs than CPUs?

[powerlord]

Pity there isn't a GPU socket on the motherboard the same as the CPU socket. Then we COULD use those big honking CPU cooling solutions (or some derivative of them), provided the case were designed to accommodate the board. You could also get high speed runs between memory (perhaps it could have its own bank), and the CPU.

Pity some CPU maker couldn't come along, buy a GPU maker, and make something like this.

(of course existing GPU solutions in slots are MUCH easier to upgrade, which is something against this sort of solution, unless they come out with a form factor that combines Chip+Cooling solution (similar to the old Slot1/A)

Re:Why is it harder on GPUs than CPUs?

[Anonymous Coward]

would you have separate slots for the video memory as well? that actually sounds kind of cool.

Re:Why is it harder on GPUs than CPUs?

[feepness]

Pity some CPU maker couldn't come along, buy a GPU maker, and make something like this.

ATI is owned by AMD.

Re:Why is it harder on GPUs than CPUs?

[Fulcrum of Evil]

Are they? Looking at CUDA, I'd say that this is debatable. More likely, the massively paralellizable problem space means they scale out instead of going for high clockrates, which also means less fancy crap with caches, as the speed differential is lower and memory access more predictable.

Re:Why is it harder on GPUs than CPUs?

[caerwyn]

Right- it's a design choice. Rather than incredibly simple micro-ops that the real instructions get translated to, or instructions spanning multiple clock cycles, they've chosen to keep the per-core implementation much simpler. That lets them pack more of the simple cores on the chip, getting additional parallel processing at the cost of per-core optimal performance- which is fine, because these are things running on massively parallel problems.

CISCy was perhaps the wrong choice, but it's valid in a sense- each cycle on a gpu is more complicated than a cycle on a modern CPU, but that avoids inter-cycle complexities that occupy a huge amount of die space on the CPUs to speed up single-threaded performance.

Re:Why is it harder on GPUs than CPUs?

[Firethorn]

This site suggests a couple possibilities.

A: A GPU had, until fairly recently, only a 1 high slot. Even with 2 slots, it has less room for cooling than the CPU, where weight actually matters more than size.
B: Transistors. The site dates from 2006, but mentions that my core 2 duo has ~291 million transistors. A G800GTX has 680M, and my research shows that the 4890 this review is about has 959 Million. Even a Core 2 Quad is 582M, and we know they cost a bit more for a given speed rating. A GT200 is listed as 1.4 Million.

That's quite a difference.

Re:Why is it harder on GPUs than CPUs?

[R.Morton]

Either that or make a Video card like they do mother boards.

Have an upgrade slot so you can drop on a new GPU like you do with a CPU on a Mobo, and then have a Notebook style Ram slot so you can upgrade the Ram on the card itself.

Provide a driver that lets you select if you want to supplement the Video Ram using Extended ram, or in cases where the card would have more ram than the PC then let the PC borrow some of the Video card Ram.

I know this is pure science fiction on my part but I figure if PC's could one day use the same High speed Ram like Video cards do but in a Universal sense then it would possible.

But I am not a Tech guy so I do not know.

R.Morton

Re:Why is it harder on GPUs than CPUs?

[imboboage0]

of course existing GPU solutions in slots are MUCH easier to upgrade, which is something against this sort of solution, unless they come out with a form factor that combines Chip+Cooling solution (similar to the old Slot1/A)

You're not gonna believe this dude, but someone beat you to the idea of a slotted GPU. Sorry. =[

http://www.legitreviews.com/images/reviews/378/ati_radeon_x1950.jpg

They put all the pins on the bottom in such a way that it fits into a modular slot on the motherboard and even comes with built in cooling. =]

Re:Why is it harder on GPUs than CPUs?

[Anonymous Coward]

Isn't this exactly what the HTX slot and HyperTransport is supposed to enable? I hear they're popping FPGAs into HPC systems already like this. It'd be nice if stuff like this was more available and usable in the mainstream market.

Re:Why is it harder on GPUs than CPUs?

[Chris Burke]

Pity there isn't a GPU socket on the motherboard the same as the CPU socket. Then we COULD use those big honking CPU cooling solutions (or some derivative of them), provided the case were designed to accommodate the board. You could also get high speed runs between memory (perhaps it could have its own bank), and the CPU.

Not a bad idea, though discrete cards today have dedicated memory for the GPU, with a bus custom designed for that card. Not expandable, but high performance. The connection to main memory would be faster if the new socket was connected directly to the cpu socket, though that mostly helps UMA (i.e. cheap) solutions and having the GPU on die with the CPU would be cheaper. The extra cooling potential would be helpful for sure, either for performance or maybe to provide the same cooling quieter since my GPU is already my loudest fan!

Pity some CPU maker couldn't come along, buy a GPU maker, and make something like this.

Yeah, unless that company was also a motherboard maker, that might be a hard sell. An extra ZIF socket with heat sink mounts and extra memory channel, in mainstream desktop mobos? Yeah, foisting that kind of cost onto the mobo makers doesn't go over well unless you're the 800lb gorilla. Which of course isn't the company you're talking about.

Re:Why is it harder on GPUs than CPUs?

[Ecuador]

Something that I don't see other posters mention is that the design of many parts of the CPU are hand-tweaked down to the transistor level exactly for this purpose - low heat, high frequency. GPU's are designed in a larger scale, which is logical if you remember that if you exclude the cache, the CPU is a much simpler (in transistor count) than a GPU, when GPU generations occur much more often and differ more from one to the other. So, you have a fraction of the CPU design cycle to incorporate a radically different feature set, there is no time for the same kind of tweaks - if you end up with very low clock speeds you can still parallelize some more.

Re:Why is it harder on GPUs than CPUs?

[Xest]

I've actually had larger heatsinks on my GPUs than CPUs in recent years with those double height graphics cards taking up two back plate slots and that was with a 2.83ghz quad core when it was high end! That said, the heatsinks often seem to be on the wrong side of the card in most motherboards/cases, that is, they're on the bottom, and of course, heat rises, so presumably it's more that than the actual size of the heat sinks? They seem to get round this by creating those heat tunnels that try to lead the hot air out the back of the system presumably drawing heat away from the components using convection by creating a cooler space underneath the components for heat to dissipate to.

Re:Why is it harder on GPUs than CPUs?

[somersault]

(that was his point)

Re:Why is it harder on GPUs than CPUs?

[Anonymous Coward]

They do have this...GPU on the mobo that is. It's called IGP (Integrated Graphics Package), almost all budget motherboards have them.

Graphics cards already have their own bank of memory...but you still need a bus between the CPU and the memory, it's called PCI Express.

What exactly are you asking for? A graphics card that you can replace just the GPU??

apples to apples

[alta]

I have a intel quad core 2 duo, a Q6600 I think.

How many TeraFLOPS is that?

Re:apples to apples

[pshuke]

According to intel it's about 0.04.

Re:apples to apples

[Anonymous Coward]

I doubt it's even 20% of a teraFLOP. The Cell Broadband Engine is more comparable.

Re:apples to apples

[wjh31]

last time i checked, a graphics card will get about 100x more Flops than a similarly priced CPU, give or take an order of magnitude (hey, im an astrophycist, order of magnitude is good enough)

Re:apples to apples

[Randle_Revar]

Even quad-core x86 CPUs are in the 10s of GigaFLOPS.

CPUs have to do a lot of integer ops, and have to be good at everything. GPUs simply have to crunch a lot of Floating Point numbers,

Re:apples to apples

[Anonymous Coward]

Modern GPUs including every single Nvidia GPU since the G80 series has had a full integer instruction set capable of doing integer arithmetic and bit operations.

CPUs aren't designed to be good at everything, they're designed to be exceedingly good at executing bad code, which is the vast majority of code written by poor programmers or in high level languages.

You can write code for a CPU without worrying specifically about the cache line size, cache coherency, register usage, memory access address patterns and alignment or memory latency on branches or pipeline stalls and the difference in performance compared to optimized code will significant but not unbearable.

GPUs devote significantly less (or in some cases no) die space to things like branch prediction and automatically managed caches. Poorly written GPU code is sometimes almost 2 orders of magnitude slower than well written GPU code, but well written GPU code has much higher potential than what is achievable on modern CPUs. See: CUDA.

Ummmm.....

[Anonymous Coward]

Didn't we learn long ago that clock speed by itself means nothing? Even those stupid enough to buy the hype learned this around the time the Core came out at significantly slower clock speeds than the P4.

If this is the best thing AMD has to hype, the card is likely a piece of junk.

Re:Ummmm.....

[conteXXt]

I may not completely understand graphic cards but,

I think in this case clock cycles actually *DO* mean something.

Re:Ummmm.....

[Anonymous Coward]

Except 90% of people still think a 3.2 GHz dual core P4 will outperform a 2.8 GHz Core 2 Duo. Hate to tell you, but people are still buying that crap.

But you have to realize that clock speed does mean SOMETHING. This card is already in production, but at a lower clock speed. So, all else being equal, increasing the clock speed really DOES increase performance. It's when you're talking about different architectures that clock speed loses its value.

Re:Ummmm.....

[Randle_Revar]

A 3GHz P4 is faster than a 2.6GHz P4
A 3GHz core 2 is faster than a 2.6GHz core 2
A 1GHz R700 is faster than 800MHz R700

Anyway, the R700 (Radeon 4xxx) series has been very good, mostly equaling or beating Nvidia's current lineup at similar prices.

Re:Ummmm.....

[mr_mischief]

How about the fact that it runs each instruction on 800 pieces of data at once? This isn't a 1 GHz one, two, four, or even 16-way chip. It's processing up to 800 pieces of data at once, and its clock for doing that ticks every billionth of a second. You're absolutely right, the clock speed by itself means nothing. The clock speed times the amount of work done per clock does mean something. If you raise either without lowering the other, you raise the overall amount of work the chip can do.

Re:Ummmm.....

[Anonymous Coward]

How about the fact that it runs each instruction on 800 pieces of data at once?

Small correction: The R770 chip used in the Radeon 4870 and 4890 is capable of executing up to 10 simultaneous instructions, each of which is sent to sixteen shader units, each of which consists of 4 FP ALUs and one transcendental ALU.

http://en.wikipedia.org/wiki/Radeon_R700#Execution_units

1 TF/s is so 1996

[Anonymous Coward]

The original ASCI Red is notable for being the first computer on Earth to bench above 1 TeraFLOPS on the Linpack benchmark (1996), as noted in Top500 Supercomputer sites. After being upgraded with Pentium II Overdrive processors, the computer has demonstrated Linpack performance above 2 TeraFLOPS.

from top500.org

Re:1 TF/s is so 1996

[Penguinoflight]

There's nothing stopping you from putting 6 of these in a single system. Even better, go with 4 4870x2s for around 12TFlops

Re:1 TF/s is so 1996

[wjh31]

if you're imagining insane numbers of cards for silly TFlops, the 4TFlop nvidea Tesla has a 1U rack form, so you can shove as many of them as you like in a rack

Re:1 TF/s is so 1996

[Randle_Revar]

Great for ASCI Red. Now, in 1996, can I buy ASCI Red in a size that fits on a single PCI-e card and costs less than $300?

Re:1 TF/s is so 1996

[mr_mischief]

Well, up to 42 anyway...

Re:1 TF/s is so 1996

[Ilgaz]

Well, if you got money, you can have 180 GigaFlop (32bit) or 90 GigaFlop (64bit) right now, on a PCI-e card.

http://us.fixstars.com/products/gigaaccel/

It is Cell powered as you may guess. There is also mention of "720 GF computing power" which I can't even dare to think about it. I guess it is when you combine 4 of them. Oh, just $6100 per one :)

Dear AMD, intel, nVidia, etc

[Yvan256]

As you may have seen from the sales of netbooks and low-power computers, the future is... wait for it... low-power devices!

Where are the 5W GPUs? Does the nVidia 9400M require more than 5W?

Re:Dear AMD, intel, nVidia, etc

[Lonewolf666]

Even for desktops, I'd like to see more of those. Lets say below 20 W, so a not-too-massive passive heat sink will do.
I'm quite happy with the performance of my NVidia 6800 GT, and it needs about 50W at full usage. With the latest chip technology (40 nm anyone?), the same performance should be possible with much less power consumption.

Re:Dear AMD, intel, nVidia, etc

[drinkypoo]

Does the nVidia 9400M require more than 5W?

Google is your friend

The GeForce 9400M claims a TDP of only 12 W.

Re:Dear AMD, intel, nVidia, etc

[Randle_Revar]

>Where are the 5W GPUs?

Intel integrated graphics

Re:Dear AMD, intel, nVidia, etc

[Anonymous Coward]

Speak for yourself, some of us enjoy being able to play crysis/project origin at high res with detail quality maxed out. If the only thing you use a computer for is email and beating off to pictures of drunken sluts on facebook, go get yourself a mac. I hear you can get SHINY ones now.

Re:Dear AMD, intel, nVidia, etc

[Randle_Revar]

Radeon HD 4350 (55nm) is ~20 W and I think should be somewhat better than a 6800.

Re:Dear AMD, intel, nVidia, etc

[Anonymous Coward]

Have a look at the ATI Radeon HD 4350 and Radeon HD 4550. Both are said to use a maximum of 25W under full load and test show an average of about 9W in 2D.

Re:Dear AMD, intel, nVidia, etc

[Cornelius the Great]

Nvidia has the mobile graphics line, which is designed for cellphones. AMD used to have a mobile graphics division, but I believe that they sold it to Qualcomm.

So in a the next few months, we'll be seeing mobile chipsets from both companies (Nvidia's Tegra and Qualcomm's Snapdragon) that will have scaled-down tech capable of handling HD video and impressive 3D graphics on embedded devices.

Re:Dear AMD, intel, nVidia, etc

[Totenglocke]

No one said that this video card was going to be shoved into every computer. This video card is for people who use a computer for more than reading slashdot and checking email.

Re:Dear AMD, intel, nVidia, etc

[hguorbray]

as you can see from the pictures of the massive heatsink (covers the entire board) this is NOT a low power device

and until there is a market for laptop gamers wanting 60fps and millions of polygons specialized cards/chips like this will be found only on render farms, gamer desktop rigs and graphics workstations -which is their intended market anyway

you generally do not get high performance with an economical product, so, for my car analogy I will say that a Pontiac Vibe that gets 35 miles to the gallon is not going to beat a Mustang V8 off the line....

-I'm just sayin'

Re:Dear AMD, intel, nVidia, etc

[Anonymous Coward]

Cheaper / lower power cards are derived directly from their flagship brethren: cut out half the pipelines, underclock the core, use slower memory et voila. (In fact sometimes they are literally the same card just speedbinned with a different BIOS.)

Admittedly however, to some extent a really low power GPU would have to be designed that way from the ground up.

Re:Dear AMD, intel, nVidia, etc

[StikyPad]

So "only" about as much power as a hard drive.

Re:Dear AMD, intel, nVidia, etc

[Ilgaz]

Yes, when you offload the entire thing to CPU and even ignore hardware t&l feature from GeForce 2 ages, it goes down to 5 watts.

Even Apple couldn't stand to their junk and switched back to real GPUs, down to "non pro" laptops.

Re:Dear AMD, intel, nVidia, etc

[drinkypoo]

It's a big step in the right direction. I had been hoping to answer the other question but it looked like it was going to be too hard to find information on an embedded GPU core (like for cellphones and stuff.) I wonder what's in the GP2x Wiz

Re:Dear AMD, intel, nVidia, etc

[Ilgaz]

Soon, not just gamers but ordinary users may need way higher "FPS" than today. 3d stuff (200hz), artificial 3d, massive amounts of transcoding, 12bit per channel video, 2K (or even 4K) are all making their way to average home user. Slowly but sure. These things were all pro high end studio stuff just some years ago.

For example, Apple is still testing a technology which scales desktop to infinite levels of DPI. It is there, embedded to core of OS but not stable or complete yet. To display such a desktop on a hardware accelerated manner, you really need some GPU power. You probably know what we see as "2d" on modern, accelerated desktops is part opengl/direct3d.

Re:Dear AMD, intel, nVidia, etc

[averner]

They've been around for a long time - they're called integrated graphics.

Re:Dear AMD, intel, nVidia, etc

[ZosX]

uhhh....wikipedia?

Specifications

        * Chipset: MagicEyes Pollux System-on-a-Chip
        * CPU: 533MHz ARM9 3D Accelerator
        * NAND Flash Memory: 1 GB
        * RAM: SDRAM 64 MB
        * Operating System: GNU/Linux-based OS

Re:Dear AMD, intel, nVidia, etc

[Fulcrum of Evil]

get a quadro 240 (I think) passive heatsink, 25W power dissipation.

Re:Dear AMD, intel, nVidia, etc

[Randle_Revar]

I wouldn't call it junk. My X3100 plays ioquake3 just fine.

Re:Dear AMD, intel, nVidia, etc

[blahrvat]

and ioquake3 is how old now? Tried anything with graphics demands from this decade? Even older games like Quake4 and Doom3 at their lowest settings are unplayable on even Intel's fastest, the X4500HD.

Generation for generation Intel's fastest IGP graphics are at least half as fast as the bottom end IGPs from AMD/ATI and Nvidia.

And with AMD/ATI you can still have your open source drivers. http://www.x.org/wiki/RadeonFeature

Power consumption?

[LoRdTAW]

No mention of power consumption or heat dissipation. My PC is already a radiator and in the summer fights with my AC.

I am interested in the computing power, 1.6 terraflops is no small number even if it is single precision.

Re:Power consumption?

[wjh31]

wikipedia (http://en.wikipedia.org/wiki/Comparison_of_ATI_graphics_processing_units#Radeon_R700_.28HD_4xxx.29_series) suggests the 4890 comes in at 190w, go to s little under double that if they make an x2 version. entry level 4000 series comes in at 25W.

if you want TFlops, try the 4870x2 at 2.4TFlops, or NVideas tesla (http://en.wikipedia.org/wiki/NVIDIA_Tesla) series, made just for GPGPU which reach over 4TFlops

Re:Power consumption?

[F34nor]

This is why I am going to literally make my next PC a hot water heater.

Re:Power consumption?

[Pope]

Personally, I'd recommend you make it a cold water heater, and get more bang for your buck!

Re:Power consumption?

[HTH NE1]

That's a waste of power. Do a luke-warm water heater and half the work is done for you.

Re:Power consumption?

[Anonymous Coward]

You own an AC? I thought slavery was illegal!

-AC

how's ATI driver quality and performance on linux?

[yanyan]

I'm a long-time Nvidia user because of good driver support on Windoze and Linux. I would love to give ATI a try but i've read a lot of negative things about driver quality in Linux. Granted, that was some time ago and things may have changed today. I'd be interested to hear about other slashdotters' experiences using today's ATI hardware + drivers under Linux/X.

Not first ??

[Anonymous Coward]

The GTX 260 I have reports 1242 MHz using CUDA. The clock rates are different for graphics and computation (not really sure why) but this article does seem to be about computation.

So perhaps claiming to be first to 1 GHz is a bit spurious?

Re:Not first ??

[Warlord88]

The 1242 MHz speed is the frequency of vertex shaders, not the core speed. Also, 1 GHz is the core speed without overclocking.

Re:Not first ??

[JackARot]

Also, 1 GHz is the core speed without overclocking.

False. It's overclocked alright, it just doesn't have to be overclocked by users or the third party manufacturers to run at 1 ghz. From their press release:

Nine years after launching the world's first 1 GHz CPU, AMD is again first to break the gigahertz barrier with the factory overclocked, air-cooled ATI Radeon(TM) HD 4890 GPU -

Re:Not first ??

[tannsi]

Just exactly what do you call overclocking?

GPUs (or CPUs) haven't got the clock speed set from the beginning of the manufacturing process, it's actually based on a number of quality checks done by the manufacturer.

Therefore I find the term "factory overclock" just a bit misleading, since the clock speed is set at the factory anyway.

Re:Not first ??

[Warlord88]

False. It's overclocked alright, it just doesn't have to be overclocked by users or the third party manufacturers to run at 1 ghz. From their press release:

Oh. So is this the same as their HD 4890 OC edition? TFA says

At any rate, AMD has today proudly announced the planet's first 1GHz graphics processor (without third-party / user overclocking, of course) with the ATI Radeon HD 4890 GPU.

I had seen and read reviews on OC edition quite some time back. So I thought this was a newer version.

*Punches fist in air*

[martin_henry]

I can finally get a 5.0 on the Vista Experience Index!

Re:*Punches fist in air*

[Warlord88]

The highest score you can get on Vista is 5.9.

Re:*Punches fist in air*

[socrplayr813]

I think that was the part that was meant to be funny, but my 8800 has gotten a 5.9 on that test for over a year now. Isn't it time we moved past the 'Vista is slow' thing?

Re:*Punches fist in air*

[dogmatixpsych]

Just because your graphics card is fast does not mean Vista isn't slow (yes, the double negative was on purpose).

Re:*Punches fist in air*

[Anonymous Coward]

Actually, the scale goes to 10.0.

But you lose 4.1 for running Vista.

FLOPs/Hz

[dolphinling]

1600 FLOPs per Hz? That's actually rather impressive.

And....

[MasseKid]

And it's still slower than a GTX 285 OC edition. Ghz != Preformance. And Nvidia, stop renaming your cards damn it!

Re:And....

[PitaBred]

The 4890 actually DX 10.1, and probably has support for almost all the features in 11. Does the Nvidia card? Didn't think so.

I'm also interested in your "slower than a GTX 285" assertion. I just looked at some benchmarks, and Xbit labs has an overclocked 4890@1GHz beating the tar out of the 285.

Re:And....

[MasseKid]

Yeah, sure just as long as you ignore Quake Wars, FarCry 2, Left 4 Dead, Devil May Cry, Prince of Persua, Mass Effect, H.A.W.X., World in Conflict, and many more and compare a stock card to the max overclock of a card. :rollseyes:

uhhh

[nomadic]

AMD does note that Asus, Club 3D, Diamond Multimedia, Force3D, GECUBE, Gigabyte, HIS, MSI, Palit Multimedia, PowerColor, SAPPHIRE, XFX and others are all aligning to release higher performance cards."

Wait, let me get this straight. Graphics card manufacturers are actually attempting to make their graphics cards perform better? Why was I not informed of this before???

"Barrier"

[Burning1]

AMD Breaks 1GHz GPU Barrier

You keep using that word. I do not think it means what you think it means.

Re:how's ATI driver quality and performance on lin

[Jeek Elemental]

Driver is fine (finally).

Re:how's ATI driver quality and performance on lin

[Randle_Revar]

ATI drivers are great in Linux

"factory" "overclocked"?

[chrispitude]

*slaps forehead*

Only marketing weenies would play up such an oxymoron.

Re:"factory" "overclocked"?

[chrysrobyn]

"Factory" "overclocked"? *slaps forehead* Only marketing weenies would play up such an oxymoron.

I'm pretty sure that word doesn't mean what you think it means. "Overclocked" means "our reliability people don't think this is smart, but it might work for you." In this case, you get a part that may or may not die before you expect it to, it might not last much beyond the warranty, it might have non-standard cooling to enable an operating window that Reliability can't assume (say they model frequency shifting at 85C and they have a heat sink that puts it at 55C; feel free to substitute any other numbers).

Any conditions where the company's Reliability department didn't endorse frequency over the lifetime of the product for 3 sigma worth of sellable parts would be "overclocked".

Re:"factory" "overclocked"?

[windwalkr]

I'd have to agree with GP - this really is a marketing term. As you correctly note, overclocking implies that the part is unreliable. I doubt that any large business is going to deliberately risk the high rate of returns that would be associated with shipping an unreliable part. IANAL, but I doubt that "it's sold overclocked!" is going to get you out of honoring a warranty of merchantability or fitness.
While they may indeed be up-clocking a standard part, and then improving cooling and so on, you can be sure that they will have also performed an extensive reliability analysis on the new hardware. If you're saying this is being done at the "Factory" then the normal meaning of "overclocked" doesn't apply.

My 2c.

Re:"factory" "overclocked"?

[chrysrobyn]

There's a difference in clientele and expectations. Most people want a part to last for 5 years and be qualified for 85C. That's one speed sort. Some may provide better cooling and keep it under 55C and only expect 3 years. Depending on the technology, that's a different sort entirely. I'm not disagreeing that it's a marketing thing, but this is not the same QA standards.

It Will Still Be Epic

[Anonymous Coward]

I "can't wait" until they reach 9GHz, and everyone goes "OVER NINE THOUSAAAAAAAAAAND"

Re:It Will Still Be Epic

[Mozk]

Actually that would be over nine million thousand. Don't worry, though—you're only six orders of magnitude off. You could work at NASA with that kind of cunning.

What does this mean for us Linux users?

[Anonymous Coward]

Can I finally play Doom 3 on ultra with the slow ATI Linux drivers? How long until they pull support for this new GPU out of the proprietary drivers?

taking bets

[stenchcow]

I'm taking bets on how many days it will take Nvidia to one up them with a faster card. I'm guessing 3 days. Any other guesses?

Disclaimer ....

[karvind]

Maybe you want to check the disclaimer too ...

Note: Damage caused by overclocking AMDâ(TM)s GPUs above factory-set overclocking is not covered by AMDâ(TM)s product warranty, even when such overclocking is enabled via AMD software.

Still an ATI

[wiedzmin]

Still, it's an ATI... I'll just wait a couple days for nVidia to come out with a better card.

What about a real revolution?

[Ilgaz]

Offer the card, in same price down to cents along with a goodly written driver for Mac Pros and even more miraculously to last generation G5s (Quad/Dual Core).

Open Firmware, Endianness, Altivec, non standard interface (???), all excuses gone. If anyone wonders what I talk about, just watch this card's price when (if!) it ships to Macs. You will understand the comedy going on. In PowerPC times, we had some sort of excuse as "Firmware is hard to code", "drivers man, they can't code for PowerPC" etc. Now all excuses are gone and we sometimes get up to 3x price by this duopoly named NVidia and ATI.

Re:What about a real revolution?

[blahrvat]

As I remember it, the PPC models required a larger firmware then the PC model's BIOS chip could fit normally, so when a card came out for Macs you first had to find a way to get the open firmware cut down to work, this often meant you lost something like the TV out or the DVI out capability.

I've not done it for the new intelmacs but I suspect something similar has to be done in a bios flash of the PC version cards to support Macs. You see more options these days since most option cards are from nvidia, since nvidia has been rebranding their 8800 series for a while now without really making improvements. See why AMD/ATI have DX10.1 support and has had it for over a year, while Nvidia doesn't. Also note that nvidia paid to have DX10.1 support removed from the game Assassin's Creed as it caused the game to run faster with higher detail settings then the direct competition cards from Nvidia.

Asymptotic, my ass

[Cajun Hell]

Everyone thought it would be 999MHz this year, 999.9 MHz the next year, 999.99999 MHz a few years later. It looked uncrossable! Well done, AMD!

Re:how's ATI driver quality and performance on lin

[Anonymous Coward]

i find my Nvidia card actually has better opengl performance under linux than windows, i think that because the windows drivers for the geforce restrict GL performance, so you buy their exoticly priced quadro crap, which by the way have identical hardware to the gaming cards. I might buy a radeon next time, although its probably the same deal with the fireGL cards right?

should be in a HTX slot not a pci-e onee

[Joe The Dragon]

should be in a HTX slot not a pci-e onee

Bull Fucking Shit. Alpha ev5 hit 1GHz

[Anonymous Coward]

I had a memo across my desk that with a stepping glitch, a revision of 21164b crossed the 1GHz barrier with a TEC/Peltier device and at normal operating conditions (unlike all the cryogenic shit AMDum and LintHell likes to do). Of course, the 21164a was on 666Mhz back when Intel was at a classic non-MMX Pentium 180MHz processor (50MHz bus, by the way).

Alpha was bought, suppressed, stocked, chopped, quartered, and sent to the far gates of heaven as Hell's reminder that it intends to suppress anything that violate Moor's Law of Strategic Economic Regulation at the CEO/Marketroid level. In other words, Moor's Law was created to stretch technology out long periods of revision so it can sell more crap. A. B. B2. B2+. B2+SGRAM. A1000DDR. A1001BBQ. A2000X. A2000X+Booger. A2000XXtreme. B3UNDERDOG! B3+A2K. B4FTW! B4FTW1337.

Fuck you Compaq. Fuck You Intel. Fuck you Tenacious D!

Sincerily,
Alpha Troll

NEWS: Duke Nukem Forever CANCELED as of May11

[Anonymous Coward]

Look at TimeDoctor.Org on the main page, and you'll find this story.

Why isn't this news on Slashdot? We are talking about the most anticipated title for the past 7 years, is complete 99% yet will not be released because it didn't polish, even has exceptionaly great footage that would shame all ID Software titles in terms of Fun-factor and engine function, yet it has tailspinned into a massive faceplant for which Duke himself has been spanked. We're talkin' 'bout programmers DIEING in the course of production in such an environment, and it is cancelled!

I think he meant to say the opposite.

[Anonymous Coward]

All motherboards should be the core function of visual/graphics performance, and with an old school "unix" type of design back on DEC where we have an expansion card containing the CPU's and Memory Modules but plugged into the Motherboard. Maybe a Fiber Optic backplane would be better, but what is the superior Bus?

I'm. I'm confused. Graphics technologies are updated quicker than CPU and RAM performance. Why don't we just pick a technology to build the Bus legacy upon? Let's build everything into a Keyboard because we throw those away like never. I puke on my keyboard and just cary it onto my next tower that I buy or put together. I have this old rag of a keyboard with a converter over to AT style, then a converter over to PS/2 mini-DIN, then a converter over to Firewire, then a converter over to USB.

You can't keep me down, biahtch!

lol?

[Anonymous Coward]

All I know is nVidia never failed me. I've had 2 ATI chipsets fail on me, roughly 9800 era. I have been nVidia for a few years now. I will say, ATI will have driver problems and heat problems.

I recall gaming with my 9800GT, the screen had literally "gone plaid". It looked like a plaid pattern. Moved to a GTO something or another, went blank screen. I was fed up with ATI, moved to nVidia, never looked back.

Re:how's ATI driver quality and performance on lin

[blahrvat]

The drivers are pretty good now, and for the older hardware the open source drivers are getting pretty nice http://www.x.org/wiki/RadeonFeature