AMD's Next-Gen Steamroller CPU Could Deliver Where Bulldozer Fell Short

MojoKid writes "Today at the Hot Chips Symposium, AMD CTO Mark Papermaster is taking the wraps off the company's upcoming CPU core, codenamed Steamroller. Steamroller is the third iteration of AMD's Bulldozer architecture and an extremely important part for AMD. Bulldozer, which launched just over a year ago, was a disappointment. The company's second-generation Bulldozer implementation, codenamed Piledriver, offered a number of key changes and was incorporated into the Trinity APU family that debuted last spring. Steamroller is the first refresh of Bulldozer's underlying architecture and may finally deliver the sort of performance and efficiency AMD was aiming for when it built Bulldozer in the first place. Enhancements to Fetch and Decode architecture have been made, as well as increased scheduler efficiency and cache load latency, which combined could bring a claimed 15 percent performance-per-watt performance gain. AMD expects to ship Steamroller sometime in 2013 but wouldn't offer timing detail beyond that."

Comments

AMD has cool code names.

[Anonymous Coward]

They all sound like sexual positions.

Re:AMD has cool code names.

[Johann Lau]

And Intel ones don't? Who are you kidding?

Aladdin
Bad Axe
Bad Axe 2
Batman
Batman's Revenge
Big Laurel
Black Pine (a cute name for anal sex I guess)
Black Rapids (I don't want to know)
Bonetrail
Caneland
Cougar Canyon
Glidewell
Tanglewood (sounds bi to me)
Warm Springs

and last, but never least, the

Windmill (also known as "Helicopter Dick")

Re:AMD has cool code names.

[Anonymous Coward]

Just wait until they come out with the Killdozer.

(haha... the captcha for this post is "mighty")

Re:AMD has cool code names.

[Anonymous Coward]

To me they sound more like REALLY slow and clunky vehicles... Not sure what AMD is hoping people associate their products with.

Re:AMD has cool code names.

[Anonymous Coward]

Big powerful machines.

Doesn't necessarily work, but I can sort of see where they were going with it.

Re:AMD has cool code names.

[Anonymous Coward]

Interesting that this story is about how great Steamroller is going to be, when we're stlil waiting for desktop versions of Piledriver, the CPU that was supposed to fix the first-generation problems of the Bulldozer.

Of course the initial Piledriver cores that were included in the Trinity APU turned out to be just a 5-15 % improvement over Bulldozer, so it's not that bad, but AMD is still sadly lagging years behind Intel.

Re:AMD has cool code names.

[thsths]

> but AMD is still sadly lagging years behind Intel.

Exactly, so a promised 15% increase in efficiency next year is not going to cut it. Intel has an advantage of about 50%, and they will probably deliver improvements by next year, too.

So for me, this message says that AMD has lost the race.

Re:AMD has cool code names.

[dbIII]

In some races they are just about alone on the track. An AMD based server with 64 cores and 128GB of memory will set you back $9000. with Intel you can now get 80 cores for about ten times that, or 40 cores for about five times that.
For some tasks when you can get 640 slightly slower cores (the ten core Intel chips have a lower clock than the ones with less cores) for the same price as 80 it's pretty easy to see which way to go. If anything is massively parallel you can forget about Intel at this point.

Re:AMD has cool code names.

[gweihir]

Indeed. I think AMD is actually far ahead of Intel (again, think e.g. integrated memory controller, for quite a few server-loads Intel was vastly behind for a time due to that). The speed increases of CPUs have become slower and slower and mater less and less. The trick for AMD will be to survive intact until Intel gives up and gets a next-gen architecture of their own. By then AMD will have ironed out the kinks and they will be on an equal footing again. When looking at their relative sizes and cash-reserves, it is impressive that AMD can compete at all. But the bottom-line is that in almost all cases (exception: You need a small number of CPUs with high power because your software is stupid, and cost of the CPUs is not an issue) you get significantly better value for the money from AMD.

Re:AMD has cool code names.

[SJHillman]

Sounds to me like they're all machines that can get massive amounts of work done in a relatively short time (compared to any alternative).

Re:AMD has cool code names.

[michelcolman]

For me, the first association that comes to mind when I hear "Steamroller" is "slow". Sure, it can do one task really well, but come on...

Re:AMD has cool code names.

[Anonymous Coward]

I guess the idea was that it bulldozes down the competition.

Re:AMD has cool code names.

[rullywowr]

They should just drop the "driver" and name it "pile."

Re:AMD has cool code names.

[neokushan]

I'd rather make a getaway on top of a steamroller instead of running with a clawhammer between my legs.

Re:AMD has cool code names.

[Eponymous Hero]

http://www.clusterflock.org/wp-content/uploads/2010/09/gayroller.jpg

Re:AMD has cool code names.

[neokushan]

You could at least link to the original source:

http://theoatmeal.com/comics/literally

Re:AMD has cool code names.

[trilliwig]

This hasn't been true for about 4 years. Intel is now so crushingly ahead of AMD in the server space performance-wise and energy-efficiency-wise that they hold 94.5% of the server/workstation market. http://www.tomshardware.com/news/amd-intel-cpu-processor,15041.html

And since server/workstation hardware costs are dwarfed by server/workstation software license costs, which tend to be per core, that means that the total cost of ownership of fewer Intel cores is actually less than the TCO of more, slower AMD cores that deliver the same performance. It turns out that due to Amdahl's Law, single-threaded performance matters even more in the many-core space.

Re:AMD has cool code names.

[Eponymous Hero]

why? the image was all i needed for the joke.

Re:AMD has cool code names.

[trilliwig]

The cost difference is not nearly an order of magnitude when you take into account all the other components of the server besides CPUs you have to pay for anyway.

http://arstechnica.com/business/2011/11/bulldozer-server-benchmarks-are-here-and-theyre-a-catastrophe/

In desktop systems and low-end servers, the processor is often the most expensive component (or second most expensive, behind the video card), and can easily be 25-30 percent of the total system cost. But as the server becomes bigger and more expensive, with hundreds of gigabytes of RAM and terabytes of storage, the processor can become a smaller part of the total cost. For example, Dell's PowerEdge R910, a 4-socket Xeon server, lets you spend up to about $22,000 on processors, if you get four of the most expensive parts offered (the Xeon E7-4870). That's a lot, but it's nothing compared to the $185,000 that equipping the machine with 2TB RAM would cost.

Not to mention that Intel wins on performance/watt and performance/core and thus in total cost of ownership when power consumption and software licensing costs are taken into account. http://www.anandtech.com/show/5279/the-opteron-6276-a-closer-look/12

If you calculate the price of a Dell R710 with the Xeon E5649 and compare it with a Dell R715 with the Opteron 6276 with similar specs, you end up more or less the same acquisition cost. However, the E5649 is an 80W TDP and should thus consume a bit less power. That is why we argued that the Opteron 6276 should at least offer a price/performance bonus and perform like an X5650. The X5650 is roughly $220 more expensive, so you end up with the dual socket Xeon system costing about $440 more. On a fully speced server, that is about a 10% price difference.

When we look at the higher end OLTP and the non low end ERP market, the cost of buying server hardware is lost in the noise. The Westmere-EX with its higher thread count and performance will be the top choice in that case: higher thread count, better RAS, and a higher number of DIMM slots.

AMD also lost the low end OLAP market: the Xeon offers a (far) superior performance/watt ratio on mySQL. In the midrange and high end OLAP market, the software costs of for example SQL Server increase the importance of performance and performance/watt and make server hardware costs a minor issue. Especially the "performance first" OLAP market will be dominated by the Xeon, which can offer up to 3.06GHz SKUs without increasing the TDP.

The competitive picture has not improved for AMD since then, with the release of the Intel Xeon E5's that are Sandy-Bridge based 6 months ago, as opposed to the Westmere-based Xeons that were benchmarked in the above reviews.

Re:AMD has cool code names.

[Targon]

One thing that many need to consider is that an eight-core processor will be excellent for multi-threaded applications, but if the implementation isn't great or the OS does not talk to it well, it may not be as fast in older applications. Now, Bulldozer left a lot of people unsatisfied, mostly because the performance was NOT where it should have been. Piledriver itself should provide a 10-15 percent performance boost over Bulldozer, and will be available in another few months(FX 8350 from what I have read). Steamroller will hopefully give a nice boost beyond Piledriver in terms of performance.

One issue that I have is when people use "performance per watt". From that standpoint, a very slow chip that does not require much power to run has a very good performance per watt, yet will still be a very slow chip. Boosts in performance per watt can mean the chip is no faster than previous chips, but takes less power to operate, something that MOST people do not care all that much about, they want a FASTER chip. The Phenom 2 955 initially came out in a 125 watt version, and later was released with a 95 watt version. This means you get more performance per watt, but it isn't any faster, meaning most people will not care. If the electric bill goes up or down by $5/month, people just don't care when everything else is so expensive($5 means very little these days and you can't even buy lunch for that little).

We really are at the point where computers for the most part are "fast enough" for most people, and it is only the continual increase in what we are doing at the same time that drives the desire for faster machines. Moving to solid state drives would be a bigger improvement for most these days than going from a quad-core processor to an oct-core processor.

Re:AMD has cool code names.

[SJHillman]

Performance per watt is mostly important in mobile applications, especially laptops, where battery life is something that people really do care about. I've never heard of people caring about performance per watt for non-mobile systems with the exception of very low power builds. I went with a 25W Athlon II X2 for my router mostly because it does have very little power draw. It's not very relevant for the desktop/server market, but it's still very important to other markets - especially the very large mobile market.

Re:AMD has cool code names.

[dbIII]

It varies. I've got a few of those big AMD machines running independently and a mostly idle cluster as a direct consequence of per host licencing. Being able to share a lot of memory across threads is a bonus - which of course improves performance. Amdahl's Law isn't the only thing happening if you have a choice between fast cores on lots of hosts and slightly slower cores on multiple hosts, and the fast cores are only marginally faster anyway (the ten core Xeons are not as fast per core as your desktop CPU).
Price differences of an order of magnitude for marginal improvement are not likely to get obscured by software costs unless the software is incredibly expensive, in which case you may as well be using Cell based CPUs instead of Xeons.
I've got an entire cluster full of Xeons and got one two socket Xeon node just a couple of months ago for single threaded stuff, but here I'm describing a niche (single machine with a lot of cores) where AMD pulls ahead.

Re:AMD has cool code names.

[hazydave]

AMD's these days all about performance per Watt.. in that, they may be closer to Intel than if you're just looking at peak performance (where, yeah, Intel's 50% faster, give or take, at least on some work).

Re:AMD has cool code names.

[Anonymous Coward]

In some races they are just about alone on the track. An AMD based server with 64 cores and 128GB of memory will set you back $9000. with Intel you can now get 80 cores for about ten times that, or 40 cores for about five times that.
For some tasks when you can get 640 slightly slower cores (the ten core Intel chips have a lower clock than the ones with less cores) for the same price as 80 it's pretty easy to see which way to go. If anything is massively parallel you can forget about Intel at this point.

How does nonsense like this get modded +5 informative?

I just played around on Dell's website configuring PowerEdge M910 (4 socket Intel) and PowerEdge M915 (4 socket AMD) systems. As you might guess by the model numbers, these are very similar machines other than CPU brand. The cheapest systems I could make, after giving each one 128GB RAM and downgrading to the minimum cost disk config (yes, I kept it identical between the two):

M910 : $15,827 for 4x Xeon E7-4820 CPUs (2.00 GHz, 8 cores / 16 threads, 32c/64t total)
M915 : $12,076 for 4x Opteron 6272 CPUs (2.10 GHz, 8 modules / 16 cores, 32m/64c total)

Now, that's not the 40 cores you were talking about, but... Intel has such a huge per-core performance advantage, and also what AMD calls "cores" these days aren't quite the same thing as they used to be, thanks to Bulldozer's "module" architecture. One BD module contains two cores, which share hardware resources and are slower than truly individual cores would be. Likewise, Intel has hyperthreading, which turns one full featured core into two logical cores. In theory, the Bulldozer module was supposed to deliver ~1.8x the performance of a single full core of similar design. In practice, for a wide variety of workloads, I'd expect Intel to win at identical hardware thread counts unless AMD has a large clock speed advantage (and they don't, here).

But wait, there's more! Let's evaluate your 10x cost for 40 cores claim. Dell does support 10-core Xeons from the E7 family in the M910, but for whatever reason they don't let you configure the 4-socket E7-4xxx versions, only the E7-2xxx versions, which means you can only configure 20 cores. I'm not interested in spending even more time on this post tracking down another system which lets you select 4-socket 10 core E7, so let's just go by the difference in Intel list price between the E7-4820 used in the 32-core machine I priced out and the E7-4850 (10 cores):

http://ark.intel.com/compare/53675,53574

Which is $1391, so it would cost $5564 to upgrade the M910 to 40 cores (if Dell would let you). That's $21,391. Which is a far cry short of 10x, wouldn't you say?

P.S. The reason why AMD offers lots of cores cheap is that they have to in order to carve out a market niche. Their processor cores are very uncompetitive right now, so the only thing they can do is sell them at crazy bargain bin prices. This has limited market appeal -- there aren't too many applications where you want lots of slow cores over fewer faster ones, even if the system price is less -- but it lets them stay in the game. If AMD was truly "alone on the track" in a meaningful way, you'd see Intel adjusting prices to match.

Re:AMD has cool code names.

[Anonymous Coward]

AMD's these days all about performance per Watt.. in that, they may be closer to Intel than if you're just looking at peak performance (where, yeah, Intel's 50% faster, give or take, at least on some work).

Everybody is all about performance per watt. Both AMD and Intel target ~150W max TDP for server processors, and 130W or less for desktop, since those figures are about as much as you can do without requiring liquid cooling. The only way for Intel to deliver better peak and sustained performance inside those power limits is by delivering more performance per watt.

Consider this comparison of 32nm desktop processors, AMD FX-8150 (125W TDP) versus i7-2600K (95W TDP):

http://www.anandtech.com/bench/Product/434?vs=287

Despite using 30W less power, the 2600K beats the FX-8150 in almost all benchmarks (even most of the highly threaded ones where you might expect the 8150 to do a bit better). And if you step up to this year's Intel CPUs (22nm "Ivy Bridge") it gets even worse for AMD since Intel improved performance a bit more and dropped TDP to 77W at the same time.

Re:AMD has cool code names.

[Anonymous Coward]

Performance per watt is mostly important in mobile applications, especially laptops, where battery life is something that people really do care about. I've never heard of people caring about performance per watt for non-mobile systems with the exception of very low power builds. I went with a 25W Athlon II X2 for my router mostly because it does have very little power draw. It's not very relevant for the desktop/server market, but it's still very important to other markets - especially the very large mobile market.

Performance per watt is relevant everywhere.

Both AMD and Intel operate inside relatively firm TDP (thermal design power) limits. For example, neither wants to target more than about 130W for mainstream single socket desktop CPUs, because that's the practical limit for inexpensive and quiet air coolers.

The challenge for CPU designers is to deliver as much performance as possible inside a thermal design limit, which means perf/W is probably the most important metric to them. (That and power density -- the power needs to be spread as evenly over the die surface as possible to avoid creating hotspots.)

A couple years back, Intel gave some talks about what guided them while developing Sandy Bridge, the core used in i7-2600K and all related CPUs. At the beginning of the project, they had their architects generate a giant list of ideas for changes and new features, and evaluated every one by "coolness". If a feature reduced power or improved performance per watt compared to a baseline design, it was "cool". Almost all un-"cool" features were summarily thrown away, and only then did they triage the "cool" features by design effort vs. reward. When it was all said and done, Sandy Bridge delivered substantially better performance at any given TDP than the previous generation.

So you might not personally be aware of perf/W outside of mobile, but it's a very important thing. Intel's ability to offer higher performance desktop CPUs than AMD is built on top of superior perf/W.

They need to innovate

[Gothmolly]

Things like hitting the 1GHz mark first, and making a workable 64bit chip that also speaks x86 only get you so far. AMD needs to come up with something cool, else they're doomed to play catch-up.

Re:They need to innovate

[ArhcAngel]

Well they definitely need to step up with their current offerings but I will forever be grateful for their 64 bit x86 extensions. If not for that we'd be stuck with Itanium desktops...*SHUDDER*...

Re:They need to innovate

[rrhal]

There Heterogeneous System Architecture (HSA) is pretty innovative stuff. If AMD is successful this will change the way software is written and move us to a more parallel world.

Re:They need to innovate

[davester666]

"may finally deliver the sort of performance and efficiency AMD was aiming for when it built Bulldozer in the first place"

So, this will finally be able to compete with Intel had last summer? A year late is an entire product cycle, so what Intel has now should crush it, performance-wise. So AMD will still have to go for 'cheap, and good enough". I guess their saving grace is they can weld a real GPU to it, then beat the GPU benchmarks for Intel's welded on GPU.

Re:They need to innovate

[smash]

Unless they're planning to cut their margins drastically, "cheap and good enough" won't work. By the time this comes out, intel CPUs with the performance of even mid-range sandy bridge CPUs will be quite cheap in their celeron (or whatever their cheap equivalent is these days) line up.

Re:They need to innovate

[smash]

CPUs don't really drive software development that much. Or else we would have migrated off x86 years ago. If intel can get the same/similar performance without a paradigm shift in development methodology, developers won't bother.

Re:They need to innovate

[TheRealMindChild]

Really? I got a nice combo bundle (case, PSU, AMD "APU", ram, motherboard) for $120, shipped. It runs Diablo 3 and all of my steam games with no trouble. What can you get me with Intel offerings that can do the same, at that price?

Re:They need to innovate

[Anonymous Coward]

Amd surprised me with some of their more power efficient offerings. I was able to get an e-350 based motherboard for under $100 for an always on little home server. The thing draws less than 40 watts with a bunch of spinning hard drives. When doing research, it very solidly competed with Atom (Tom's hardware) in terms of price/ performance. I'd be interested to see what this refresh does for this line of chips.

Re:They need to innovate

[Pulzar]

What can you get me with Intel offerings that can do the same, at that price?

Probably nothing. The problem is that AMD hardly makes anything on selling you that whole setup, and there are too few of you who need something like that to make it up in selling huge volumes.

It's not that their stuff is awful. It's just that they can't sell the cheap stuff at enough of a profit, and they don't have expensive stuff to make up for it.

The business side of the company is failing.

Re:They need to innovate

[tlhIngan]

CPUs don't really drive software development that much. Or else we would have migrated off x86 years ago. If intel can get the same/similar performance without a paradigm shift in development methodology, developers won't bother.

The migration from x86 has already started, actually - the architecture they're moving to is ARM. (After all, there are more ARM-based SoCs shipped than x86 CPUs - every PC includes one or more ARM cores doing something).

But on a more user level - tablets, smartphones are becoming the computing platforms of the day, all running ARM processors. Regardless of whether they run iOS or Android. Developers have embraced it and cranking out tons of apps and games and other stuff for this. It's so scary that Intel's investing a lot of money bringing Android to x86 because the writing's on the wall (when more phones and tablets ship than PCs...)

But x86 won't die - it has a raw performance advantage that ARM has yet to reach, so for computation-heavy operations like databases, it'll be the heavy lifter. Perhaps serving an entire array of ARM frontend webservers.

Re:They need to innovate

[Anonymous Coward]

Intel completely dominates AMD in terms of process tech, but due to antitrust concerns, they tweak their prices so that AMD can stay barely alive in the "budget segment".

In the last 20 years, AMD had the best parts for only 2 years, and were in the running for maybe another 3-4 years. The game has always been rigged in Intel's favor.

Re:They need to innovate

[hairyfeet]

Not to mention there is a BIG rotting elephant in the room that nobody is talking about which is this: If you don't like Windows 8? You're screwed if you buy a new AMD chip. You see the braintrust at AMD didn't bother to keep MSFT in the loop about what they were cooking up with their "half core" design and now that BD and soon SR will be out there MSFT has made it clear that the scheduler bug is a WILL NOT FIX except in Windows 8.

So if you run XP, Vista, or Windows 7 with one of the new AMD chips you either have to disable half the cores or Windows will treat it as hyperthreading and tie a nice boat anchor to your new chip. This is why you can disable half the modules (thus making them like a traditional chip) and your benches will go UP and not down, because Windows doesn't know WTF to do with such a strange CPU, just as it didn't know what to do with the first hyperthreaded chips. This is also why the Thubans will win in most benches against the BD chips, even though BD has more cores, because Windows doesn't know about the half cores and how to schedule them properly.

I just hope the new chip designer they hired away from Apple can do some good and come up with another truly great design, because otherwise AMD is stuck with a chip that will only run correctly on an OS that looks to be the most hated Windows since MS Bob.

As someone who has put my money where my mouth is and not built a single Intel PC since the bribery and compiler rigging came out I've been sticking with AM3 but its getting harder and harder, and since my customers aren't going to Windows 8 I may have to start looking at Intel again. After all who is gonna want to buy a system that has to get stuck with Win 8 just to have it run correctly?

Re:They need to innovate

[hairyfeet]

Ya know what? Nothing wrong with cheap and "good enough" the problem has been their new designs are cheap and shitty thanks to that lame "half core" they went for.

You take a good 85%+ of the people out there and a MOR AMD Deneb quad will frankly be twiddling its thumbs because it will blow through any jobs that they have, even gaming, even more so for Thuban. And their Brazos chips were fricking great, an APU designed for mobile video and basic tasks that got great battery life while often being cheaper than an Atom+ION setup.

I've sold many an Athlon II and Phenom II and the people are damned happy with them, they just blast through everything they want to do with plenty of cycles left over. I even put my money where my mouth is with regards to my family, me and the oldest are gaming on Thubans while the youngest took my Deneb, and they blow through any game we throw at 'em.

I see from TFA they've partially dropped the "half core" design but I can only hope that with Piledriver they'll drive a stake through it, as most of the people I've talked to Win 8 is a DO NOT WANT yet the half core scheduler bug is only fixed in Win 8. Meh, hopefully I'll still be able to get enough Thuban, Deneb, and Liano chips to get me through the whole BD/SR phase and the new Apple chip designer they hired will give us another Athlon64. One can hope after all.

Re:They need to innovate

[afidel]

or Windows will treat it as hyperthreading and tie a nice boat anchor to your new chip.

Actually it's the opposite, the system SHOULD be treating the co-cores like HT units and not scheduling demanding jobs on adjacent cores (at least not ones that both need the FP unit or lots of decode operations). The problem is that AMD basically lied to the OS and told it that every core is the same and that it can go ahead and schedule anything wherever it wants. If they had just marked the second portion of each co-core as an HT unit the normal scheduler optimizations would have basically handled 99% of cases correctly. In reality BD's problem wasn't so much the gaff with the co-cores (though that certainly didn't help things), but that Global Foundry is more than a process node behind Intel (one node plus 3D transistors).

Re:They need to innovate

[iamhassi]

Ya know what? Nothing wrong with cheap and "good enough" the problem has been their new designs are cheap and shitty thanks to that lame "half core" they went for.

You take a good 85%+ of the people out there and a MOR AMD Deneb quad will frankly be twiddling its thumbs because it will blow through any jobs that they have, even gaming, even more so for Thuban. And their Brazos chips were fricking great, an APU designed for mobile video and basic tasks that got great battery life while often being cheaper than an Atom+ION setup.

I've sold many an Athlon II and Phenom II and the people are damned happy with them, they just blast through everything they want to do with plenty of cycles left over. I even put my money where my mouth is with regards to my family, me and the oldest are gaming on Thubans while the youngest took my Deneb, and they blow through any game we throw at 'em.

I see from TFA they've partially dropped the "half core" design but I can only hope that with Piledriver they'll drive a stake through it, as most of the people I've talked to Win 8 is a DO NOT WANT yet the half core scheduler bug is only fixed in Win 8. Meh, hopefully I'll still be able to get enough Thuban, Deneb, and Liano chips to get me through the whole BD/SR phase and the new Apple chip designer they hired will give us another Athlon64. One can hope after all.

This. I have a six-core 1055T. Bought it to overclock and it does hit 4ghz stable on air but guess what? I run it at stock 2.8ghz. Why? Because 99.9% of the time six cores at 2.8ghz is more than enough. Even games run perfectly. CPUs have finally reached the point where faster isn't better anymore, its power usage and heat output. Rather have it run cool using little power at stock then run it full blast all the time sucking watts and heating the room at 4ghz I'm not even using.

When I bought this intel didn't have anything close in price that performed as well. Sure I could have spent double and bought a faster intel chip, but why? What was the point of spending more on something I wouldn't use? Rather spend the $ on a ssd for real performance gains then extra ghz I'd never use. So I bought AMD and I'll probably do it again next year if the price is reasonable and the speed is "good enough"

Re:They need to innovate

[the_humeister]

This is just so weird. 20 years ago it was Alpha, MIPS, SPARC, PA-RISC, etc. that were the ones counted to do all the heavy lifting backend, HPC stuff. x86 was kind of a joke that everyone frowned upon but tolerated because it was cheap and did the job adequately for the price. Then x86 steamrolled through. Now no more Alpha or PA-RISC. MIPS is relagated to low-power applications (my router has one).

Re:They need to innovate

[JDG1980]

I guess their saving grace is they can weld a real GPU to it, then beat the GPU benchmarks for Intel's welded on GPU.

You may not have noticed, but Intel is fast closing the gap in integrated GPU performance. They are catching up to AMD on the integrated GPU front much faster than AMD is catching up to Intel on the CPU side.

Re:They need to innovate

[hairyfeet]

In ANY case, which your description does make sense, the problem is NOT with GloFlo but squarely with AMD.

If they would have let MSFT in on the idea from the start (and no point not telling them, it isn't like Intel was gonna gut their superior performing Core line to completely re-engineer their chips to steal an AMD idea) then they could have written the scheduler to recognize the AMD chips and treat them as modules instead of cores but they kept MSFT in the dark and now frankly everyone is getting punished because if you want a fully functional chip on XP, Vista, or 7 you should avoid AMD like the clap.

Personally I thought the whole idea was retarded except for the mobile chips like Brazos, on the desktop the idea was completely stupid and on the server even more so. For those that don't know the original plan was to go "Full APU" and have the GPU take the place of the FP on chip, which would be a much simpler and weaker design than in years past thus freeing up more TDP for more cores. Why is this dumb? Well what if you want to use the GPU AND do some floating point heavy task? Or what if you don't want the integrated GPU because you can't OC worth a crap with the GPU built in?

Frankly the ONLY place that would have been a really good idea is Brazos, because you could then use the die space to have a quad core APU that would use very little power when not doing FP heavy tasks yet would still have excellent multimedia capability thanks to the powerful GPU taking up the slack. Instead the braintrust at AMD killed the replacement for Brazos, instead going for the usual minor speed up thanks to die shrink. That is completely moronic as Brazos was selling as fast as they could crank them out and even today you still see laptops up to 15 inch using Brazos, so a quad Brazos II that used less than 17w under load and got great multimedia and even okay gaming would have been a slam dunk, and instead they bet it all on BD arch which unless they pull off a miracle may be their very own Netburst. Whereas Netburst was clock above all the BD design is cores above all and in both cases just bad design.

I'm sure AMD fanboys will have a shitfit but this is from someone who owns not a single Intel unit, my netbook is Brazos, my Office box is a Sempron nettop, and I'm typing this on my Thuban home machine. So I WANT AMD to succeed, I really really do. I remember what it was like when Intel had a monopoly and it wasn't good for anybody. After all if it wasn't for Athlon64 giving Intel a spanking we'd probably have 8GHz Netburst space heaters right now. But if AMD is gonna survive they are gonna have to do better, simple as that. I truly hope that the new chip designer they lured away from Apple can give them a kick ass design, because there really isn't a selling point for BD right now. It sucks more power, is more expensive, gets less performance than Thuban which can be had for MUCH cheaper, and it doesn't work well with anything but Windows "Hai I Iz A Smartphone LOL" 8, which looks to be as big a bomb as Vista or even Bob.

Re:They need to innovate

[Anonymous Coward]

They are doomed to play catch-up, at best.
Intel made sure of that way back when they strong-armed OEMs into not offering Athlon64 systems, later denying any wrong-doing long enough for AMD to end up in a dire financial position, so they would settle out of court for pocket change.
And yet Intel keeps selling a lot of their chips. Is it really that small of a minority that cares about a healthy competitive landscape?

Re:They need to innovate

[hairyfeet]

Hi fellow Thuban user! I have the 1035T and my oldest has the 1045T and like you I got some crazy OCing when I first got it (I ended up with 3.5GHz with 3.9GHz turbo) before going back to stock because even at 2.6Ghz it just mows through everything.

What I really love is the TurboCore, with my Asrock board I can tweak to my hearts content but even at stock settings with TC I'm getting a little over 3GHz when gaming thanks to most games using 3 cores or less. No muss, no fuss, it just kicks it in automatically when I need the single threaded boost. And with the N520 cooler, which I paid a grand total of $30 for, it stays around 8 degrees above room temp idle and never hits above 127F even when the cores are being pounded. When you can keep a chip that cool with just a $30 heatpipe cooler and arctic silver what's not to like?

What did it for me though was like you how much I could save while still having damned good performance. I have 2 teen boys that also game so I try to keep us pretty close to parity and when you can grab a complete 6 core kit for $345 and that's BEFORE the MIR gives you another $30 off? It was a no brainer. I got myself the Thuban, gave the youngest my X4 925, which considering he prefers MMOs is frankly overkill, while the oldest ended up with a kit I just like I linked to given to him by his grandpa as a back to school present before I could grab it for him.

All told for THREE systems, with the family pack of Win 7 HP X64 and 3 HD 4850 GPUs? $1400 before the MIRs, after I got those back all told it was around $390 a system. You just can't beat that and all the games we play run just beautiful at the 1600x900 res our screens run on. In a year to a year and a half I'll pick up some 6770s or 6850s when they drop to the $50 price range and make my money back on the 4850s off of CL. With two hexas and a quad we couldn't be happier, the kids gaming and movies, my gaming and transcoding along with multitrack audio editing? We have tons of cycles to spare.

So I have to agree, what's the point when these systems already can tear through anything we care to do at half the price of a similar Intel unit?

Re:They need to innovate

[kermidge]

"After all who is gonna want to buy a system that has to get stuck with Win 8 just to have it run correctly?"

Guess it depends on what one wants to do. I was under the impression that the patches for BD had been included in the last Linux kernel or two (not that it'll help AMD's bottom line viz. market percentage.) As for Win8, if nothing else a third-party dev will have a 'Metro' app with a "click/touch/punch/yell here to get to a real desktop" icon.

I like a lot of what Intel has been doing recently with performance/power/price but have found it daunting to plow through the huge list of SKUs and trying to match them up with sockets and all. After too much reading I think I settled on something like the i5-5430 for a theoretical build.

Meanwhile, my three-year old system has seen two Phenom x4s and a Phenom II x6 on the same mobo with one or two BIOS updates, so it's been simple enough for me to deal with. If I can do a build next year, I suspect it's gonna be a tough choice depending on how well Steamroller does.

Re:They need to innovate

[lexman098]

Rather have it run cool using little power at stock then run it full blast all the time sucking watts and heating the room at 4ghz I'm not even using.

Take a look at CPU-Z sometime. The CPU scales back its clock automatically when it doesn't need all the cycles.

Re:They need to innovate

[travbrad]

Why? Because 99.9% of the time six cores at 2.8ghz is more than enough. Even games run perfectly

There are virtually no games that actually use 6 cores though, so 2 of those cores just sit idle doing nothing. There are a lot of games that still only use 2 cores too, for example Skyrim (one of the most popular games in recent years):

http://www.tomshardware.com/reviews/skyrim-performance-benchmark,3074-9.html

That's one of the big problems with Bulldozer. In programs than can use all 8 cores it is somewhat competitive with 4 Intel cores, but in the programs that can't use 8 cores (quite a few) they get absolutely crushed by Intel.

Re:They need to innovate

[gman003]

Agreed. AMD give you 90% the power of Intel for 50% the price.

I built my parents a computer to replace the old Athlon XP. I could have gotten a Core 2 Duo, or a Phenom II X3. Guess which one I got?

I actually told them that if they want to upgrade (new Civilization game out, seems to run a bit slow), the Phenom X4 and X6 chips are dirt-cheap now. Haven't heard back from them yet.

Re:They need to innovate

[DuckDodgers]

That's why Tom's Hardware's latest "best gaming CPU for the money" article put the AMD FX-4170 at the top of the list on the AMD side - only four cores, but it runs at 4.2 GHz right out of the box. So in the rare event you're doing something massively parallel it will get creamed by one of the Bulldozer 6 or 8 core processors, but for games it's the best option AMD has. $120 at Newegg.

I'm building a new PC next year, if the Steamroller chips aren't really good both on performance and on price, I will probably just get the FX-4170.

Re:They need to innovate

[DuckDodgers]

There are a number of documented cases in the past few years when Intel paid major PC makers to only carry their own chips.

Regardless of the morality of that, the result was that it really hurt AMD sales, and in turn that prevented them from getting the investment capital they needed to keep improving their products.

Of course, the counter point is that AMD failed to make a compelling case to the major PC vendors that dropping AMD products was a serious mistake. That's competition at work. But I dislike the prospect of having only one major processor vendor in the market and I sincerely hope the world moves to ARM and other alternatives just so that there is more major players in the game.

Re:They need to innovate

[ZenDragon]

Same here. build a 1055T rig with 16gb of ram, on a vertex 3 SSD for under $1000 bucks about a year ago, with a Radeon 7950 it plows through every single game I have thrown at it in pretty much max resolution with all the goodies turned on. I have yet to see a reason to "upgrade" to an i7 yet. And probably wont even bother until this thing dies.

Re:They need to innovate

[Anonymous Coward]

CPUs have finally reached the point where faster isn't better anymore, its power usage and heat output.

I agree, and that's why AMD was what to buy prior to 2011.

I agree, and that's why as of January 2011 and later, you buy Intel. SB and IV rule on power usage and heat output, and they're fast too, whenever you need it.

I love Athlon IIs but I wouldn't buy one today, when I could get a Core i3 21xx instead.

Re:They need to innovate

[Cajun Hell]

otherwise AMD is stuck with a chip that will only run correctly on an OS that looks to be the most hated Windows since MS Bob.

WTF are you talking about? Nearly all OSes work just fine with Bulldozer modules. You just happened to cherry-pick three example that don't, and one that does but which you happen to not like.

Interesting that all 4 OSes you mentioned, just happen to be from one team/company.

You remind me of the kind of people who complain about Democrats and Republicans, and then go out and vote for Democrats and Republicans.

"I tried Tree Top apple juice, a local farm's apple juice, the supermarket's generic apple juice, and all 3 of them completely lack citrusy character! Why can't any make a citrusy juice, something that tastes like, oh I don't know, oranges? Should I try yet another brand of apple juice?"

Re:They need to innovate

[QuantumRiff]

That may be true in the consumer space, but is not anywhere close in the server space.

We specced recently a pair of Dell servers, (poweredge 810 and 815). Both with 256G of ram. Difference between the 810, with dual 6-core CPU's, and the dual AMD 16 cores? About $7500 per server. Both the CPU, and the RAM are much, much cheaper. The Ram might run slightly slower, but since were mainly using it for "buffer" space in Oracle, we don't care.. its still 1000X faster than disk. And our software doesn't need another few hundred Hz of speed per thread, we need lots and lots of threads.. Btw, that difference in price, is about the same as the cost of a Fusion-IO card for each server.. which will add much more of a performance boost to the DB and apps than a fancy processor. :)

Re:They need to innovate

[MrFlibbs]

Actually, the "best gaming CPU for the money" article to which you refer only gives the FX-4170 a rating of "Honorable Mention". These are Tom's Hardware recommendations for gaming CPUs at varying price points (from the July version of the article):

~$70: Pentium G630
$100: Pentium G870
$110: None (FX-4170 Honorable Mention)
$125: Core i3-2120
$180: Core i5-2310
$200: Core i5-3450
$230: Core i5-3570K
$590: Core i7-3930K

Sadly, the best desktop CPU AMD has to offer is bested by the lowly Core i3, and is crushed by any of the Core i5s.

Re:They need to innovate

[Courageous]

>I'm sure AMD fanboys will....

*perk*

There are still AMD fanboys? Where? ;-P

Re:They need to innovate

[Targon]

Only in select tests, in most situations, Intel graphics still SUCK, and their drivers just can't cut it. Intel keeps claiming PLANS for something much faster, but outside of artificial benchmarks, Intel still has a LONG way to go. AMD also does not have much to really worry about, because at ANY time, they can put in a better GPU in new APUs. There is also something called graphics quality, where the speed doesn't mean much if what you are looking at is ugly.

Re:They need to innovate

[Targon]

Piledriver you mean. Steamroller will be the next generation after Piledriver(which is due in October of 2012).

Re:They need to innovate

[hairyfeet]

Man, I just gotta ask....WTF are you doing on your desktop that 6 cores won't cut the mustard? Because i'm sitting here with a 1035T, which was strictly the MOR Thuban chip, and it just chews through anything I throw at it with plenty of cycles left over. No problem with some of the latest games like Saints Row 3 or Deus Ex:HR, just tears through video transcodes like nobody's business, I'm even running multitrack audio editing with effects and not a skip nor a hang one.

This is why me and mine are gonna sit it out for a couple of years, with me and the oldest having 6 cores and the youngest my 925 Phenom II frankly we all have cycles to spare, 8Gb of RAM a piece, All we'll have to do is change out the HD4850s for 6770s or 6850s in a year or so and we'll be good to go, Not to mention with any halfway decent board the clocks you can get are just nuts, my Asrock runs just a little over stock naturally and I get nearly 3.1Ghz on TC and between 2.7Ghz-2.8GHz running 6.

So what are you running that even a 6 core won't do the job?

Re:They need to innovate

[hairyfeet]

Dude, let go of the crack pipe. Would you like to see the data for yourself? By cutting out XP,Vista, and 7 you are losing 83.24% of the market and the rest of the market? Apple products which don't use AMD chips. Might want to look at where Linux is on that chart friend, its at....0.97%.

Now if you think a global chip company can survive by giving the finger to 83% of the market and catering to 0.97% of the market I have some magic beans you might be interested in, blessed by RMS himself I swear!

Re:They need to innovate

[kermidge]

Well, it depends on prices, but it'll take most of a year to save for the build. So I'll look at both PD and SR, then decide if it makes more sense for me to consider Intel.

Re:They need to innovate

[kermidge]

Oh, my 1090T at stock works just fine, especially considering I don't do much - a few games, the occasional virtual machine, some odds and ends. I wouldn't mind a somewhat better video card with more RAM, and doing another build would give me a backup system. (My power supply failed last month but luckily I had another on hand.)

But: World Community Grid - I wouldn't mind having a few more cores to contribute.

Re:They need to innovate

[DuckDodgers]

I knew that and I was not intentionally trying to mislead anyone. Sorry if I did.

But to be fair, from the same page the best desktop CPUs from AMD are on par with the first generation Core i5 and Core i7 chips, and that means it's still plenty for almost any conceivable consumer application. They lose to the current i3, but not the original i3. And again, value per dollar is impressive.

Re:They need to innovate

[hairyfeet]

Meh if you just want to throw a few more cores to some project throw you together a cheapie with an AMD low power from Starmicro. I've been buying from them for years, great guys and if you keep an eye on their site you can get some crazy deals. I was buying AMD Phenom I low power quads for $40 a pop and now they have the X3 unlockables for $48, cheap prices. That way you can just throw the box in the corner and let it do its thing 24/7/365 and not slow down your main machine.

As far as GPUs? The place you want to keep an eye out on is Geeks as their refurb GPUs are often crazy cheap. Again been buying from them for years, no hassles. If you ever have a problem with either site just give them a call and they fix it quick, great guys. With 2 teen boys that also game I have to try to keep us close to parity which triple everything adds up pretty quick. The HD4850s they are selling are dirt cheap, just blow through all the games we throw at 'em and the drivers are solid and mature. We'll run 'em for another year and then I'll get the 6770s or 6850s and make my money back on the 4850s off of CL.

Re:They need to innovate

[kermidge]

Thanks for the tips, I'll keep 'em in mind - have saved the info in a text file.

I've got a Phenom x4 95W, don't have the ready to put it to use right now, nor available outlets/amps to operate more than I have at the moment. My apartment is in an old house with too many common underpowered circuits.

Re:They need to innovate

[hairyfeet]

I know the feeling, my apt is from the late 20s so I had to have the super run me a separate line from the main box just to feed my babies LOL!

But you ought to look at some of the AMD forums about underclocking/volting, because you'd be surprised how long you can drop that X4 and still have it run well. depending on the chip you can drop them to as low as 60w and still have them perform decently.

Like I said though just keep an eye out on those two sites, with them its what they get in off the truck this week but if you can snatch one of the next load of low power triples and quads they get it'll be just perfect for what you are wanting to do. You undervolt one of those and you'll have a 1.5Ghz X4 that uses less than 30w under load, AM2+ boards are a dime a dozen and for specialized apps like grid (or BT or any other long term single function) you can slap in a dirt cheap 2Gb stick and be good to go.

I've found its a LOT better to use a low power system for long term programs and keep your bad boy gamer PC for gaming and other heavy tasks, saves a lot of power and heat with very little expense. I keep a low power Sempron in the shop for just that purpose, downloads, surfing, no point in using a high dollar chip sucking power when a smaller sub 40w will do the task and this frees up my Thuban for the heavy lifting like gaming, transcoding, what have you. Shop smart and you can build a decent low power box for less than $100, a KVM is like $10 for a 2 port, and my low power has been running for 8 years solid so the life you'll get out of a low power machine makes it easy to make your money back in cooling and power savings. And this way grid or anything else you want to run long term can go 24/7/365 without going nuts on your electric bill or raising your cooling costs.

But I've built several of those low power E series X4s for office units and they are just great little chips. You can power them with a bottom of the line 250w PSU no problem, no need for any fans other than the HSF, and 4 cores at 1.5-1.8GHz can still crank through a hell of a lot of work. For something like grid computing I bet it would be perfect, a silent system you can run all the time and just remote into once in awhile to see how things are progressing, easy peasy.

Re:They need to innovate

[kermidge]

Lots of good, useful stuff, I really appreciate the input. Thanks! As it happens, I've got a 9350e sitting around....

it's the boards!

[slashmydots]

I think 15% would put them around even with Intel. That means it's a toss up except, oh wait, their boards are ungodly expensive. I really don't know why, probably chipset manufacturing cost or something. A really nice MSI B75 board with all metal caps is $65 and my usual H77MA-G43 board is a mere $90. All the AMD ones i looked at that had the features I wanted (basically same ones as those last two I mentioned) are all $100 and usually well over a hundred. Just to get SATA III at all was terribly expensive on any AMD board with any chipset. So unless it has an onboard 1833MHz memory controller, that's the end of that. But who knows, boards might come out super cheap for the new chips.

Re:it's the boards!

[Dyinobal]

Huh to be honest I've not looked at Motherboards in a while, but I've never been able to get a board I wanted for less than 100 dollars and if I could I'd be pretty skeptical. Then again Elitegroup pretty much ruined my opinion of cheap motherboards.

Re:it's the boards!

[Ironhandx]

I don't buy MSI or ECS as a general rule for any chip... additionally theres pretty much feature parity for price in AMD vs Intel boards. Not sure where you're shopping

Re:it's the boards!

[Anonymous Coward]

No idea what this person is talking about. Maybe you are only shopping Newegg w/limited stock or comparing different classes of mobos. I'd be willing to bet that 90 bucks buys you a much better Asrock AM3+ or Intel board than that MSI.

Re:it's the boards!

[Osgeld]

I just got a gigabyte with dual PCI-e 4 ram slots (1833 and if you OC it a little 2000) with all the latest buzzwords for like 70 bucks ... you need to shop some more

Re:it's the boards!

[bhcompy]

I got my Asus SATA3/USB3/Firewire2 AM2+/AM3 board in 2010 for $85, so I have no fucking idea what you're talking about.

Re:it's the boards!

[Anonymous Coward]

You mean like this,

FM1 with latest SATA 6G, USB3, 1866 DDR3 support - $50 - http://www.newegg.com/Product/Product.aspx?Item=N82E16813138361

or perhaps for AM3+, SATA 6G, $50

http://www.newegg.com/Product/Product.aspx?Item=N82E16813138345

Sure, you can spend about a $100 for premium board, but you can also spend $200 for a premium Intel board too.

I got an AMD board with all solid caps $65 few months ago.

Re:it's the boards!

[Rockoon]

Every time Intel vs AMD comes up, some complete dope claims that Intel boards are cheaper.

Thanks for keeping the tradition of dopes alive.

Re:it's the boards!

[hairyfeet]

Meh, if you are building yourself a new system it should last for years, therefor spending a few extra bucks on the board at build time is well worth it.

I'd suggest an Asrock board as they have all the nice features and their XFast USB really does kick up the speeds on your USB 2 sticks and their OC Tuner software is nice and simple to use. Really great boards all around.

In the end its a whole $34 more than the ones you posted but has Crossfire support, double the RAM slots, plenty of SATA slots and S'PDIF, its just a nicer board and if you are gonna have a machine for years I'd say its worth the $34 to have more upgrade options down the line.

Re:it's the boards!

[Targon]

You have missed the difference between the higher end "expensive" boards and the lower end consumer boards. Things have changed a fair bit over the past two years since consumer level processors from AMD are the A series(E and C make no sense on the desktop) with the GPU built into the CPU. The socket AM3 and AM3+ boards are intended for machines that will be higher performing(video card, not integrated video), so you end up paying more in that segment these days.

You can find cheap boards that support what you want for the most part: http://www.newegg.com/Product/Product.aspx?Item=N82E16813157319

You will generally get what you are looking for, but like Intel, there are low end chipsets, and then you have medium and then high end chipsets. If you go cheap, don't expect the newest features.

Haswell

[Anonymous Coward]

It is supposed to come out when Haswell does next year. Who will buy it? No mention of AVX2 either.

AMD boards have better PCI-E lanes then intel chip

[Joe_Dragon]

AMD boards have better PCI-E lanes then intel chips.

With Intel you need to go high end to get more then 16 lanes + DMI

well, it better do

[smash]

... or amd are facing irrelevance. ARM is eating their lunch in mobile, the core series is eating their lunch on the desktop, and the atom isn't standing still in the low power market.

Intel's integrated GPUs are now "good enough" for most people. Those who game won't want integrated AMD if integrated intel isn't good enough...

Re:well, it better do

[Osgeld]

AMD is Intel's only direct competition in the desktop market, they are not going anywhere

Re:well, it better do

[smash]

Unless they start turning a profit (i.e., steamroller actually works this time), they won't be able to afford new fabs, and the investment required in CPU design to even try to keep up. Things have not been good for AMD lately.

Re:well, it better do

[afidel]

AMD doesn't own any fabs, that was spun off to Global Foundry, and AMD has made some noise about moving to TSMC for their next CPU despite TSMC having their own problems at the current process node and the fact that AMD will take a hit on the stock they own in GF.

Re:well, it better do

[dywolf]

The AMD APU chips are pretty damn good. I used one for my HTPC, and it runs Diablo 3 and WoW, and most anything else I throw at it more than acceptably. Now, for me, acceptably, on an HTPC, doesnt mean everything maxxed. But its an HTPC, in the living room. a 2nd machine to complement my other rig. So I dont care about being maxxed.

I -love- AMD. I havent used Intel since my first system I built with a pentium3, and that system gave me nothing but grief. My current rig machine uses a black chip, forget which atm (at work) clocking around 3.3GHz.

Sadly though I've kept reading how these last few lines of chips have been having issues. First its a quality problem, causing them to not be able to make the 4cores they wanted (so they release some as 3 cores with the faulty 4th disabled). Then its just plain yeild is too small. Then its bugs in the chip design, or it underperformed. And Ive been planning a new game rig for sometime in the next year (say, March ish). The way AMD has been going though, it's looking like I might end up making an Intel machine, which I really dont want to do. Previously AMD has always met my needs in terms of quality, power and lower cost. Now? Well. we'll see.

Re:well, it better do

[Targon]

You are dealing with some outdated information. The AMD three core processors mostly were gone by the time the Phenom 2 generation came out and once the process technology was a bit more mature. New designs are always problematic, so more "failures" are expected. The Bulldozer issues are the same way, initial batch of a new design was a bit problematic, which Piledriver will fix.

Notice that the A10 parts from AMD have NOT had production issues, and those are based on Piledriver, so now it is just about getting the Piledriver based chips out at high speeds(FX 8350 will be an eight-core chip running at 4GHz with turbo to 4.2GHz stock without overclocking at 125 watt max power draw and due out in October of 2012). That doesn't sound problematic, and should also resolve the performance issues, even under Windows XP/Vista/7.

Re:well, it better do

[smash]

HD3000 can run Diablo 3, WOW, etc as well. In terms of CPU though, intel kicks butt at the moment.

Re:well, it better do

[smash]

I just don't understand how AMD could have gotten to the point of volume production without determining the inferior performance to their previous generation first. Do they do no testing? It's not like Windows 7 is rare or hard to obtain.

Re:well, it better do

[Anonymous Coward]

You are dealing with some outdated information. The AMD three core processors mostly were gone by the time the Phenom 2 generation came out and once the process technology was a bit more mature.

6-core Bulldozers are the same 4-module chip as 8-core Bulldozers, just with one module (2 cores) disabled.

In Linux drivers, Intel is still king.

[BadgerRush]

AMD may be getting its shit together when in regards to chip design. but I'm still going Intel on my next PC because of their superior Linux drivers. At the moment I'm an unhappy owner of a laptop with a AMD graphics card that can't do anything because the drivers are useless. I'm looking forward to a new laptop with an Intel Ivy Bridge processor (I don't think I can wait Haskell).

Re:In Linux drivers, Intel is still king.

[cbope]

Wait a minute... weren't all the ATI (now AMD) fanboys claiming a couple years ago that because ATI was developing more "open" drivers that they would rule the linux landscape?

Re:In Linux drivers, Intel is still king.

[cgt]

I'm quite sure that Intel's drivers are free, too.

Re:In Linux drivers, Intel is still king.

[fa2k]

AMD are quite good about open drivers (well, 0/4 steam games work for me, Portal sort of works, but everything non-windows gaming is great). Get the Intel for the better speed, it will still be a lot faster than the AMD, but don't expect the GPU to be a lot better.

Re:In Linux drivers, Intel is still king.

[thue]

While AMD is releasing documentation, Intel is releasing actual open source drivers. And now that Intel's graphics hardware is no longer a complete joke, Intel is becoming a real alternative for some users.

AMD is still better than NVIDIA, which doesn't release documentation.

Re:In Linux drivers, Intel is still king.

[Anonymous Coward]

So, you're saying your car works great, it's just that when you drive it on the highway the wheels fall off?

Re:In Linux drivers, Intel is still king.

[socrplayr813]

Maybe in principle, but in my experience using the hardware, the drivers that NVIDIA is providing are far superior to the AMD drivers available for all but the most basic uses. This seems to be the general consensus, at least where I tend to spend my time.

If you're more concerned about software freedom than I am, maybe you'd rather have AMD. My Linux boxes are much happier with NVIDIA, especially my HTPC. If I get enough cash to throw at it, I might try a low power Ivy Bridge or one of the new Atoms for a new HTPC, but the low-power standalone NVIDIA cards are just so easy...

Re:In Linux drivers, Intel is still king.

[Anonymous Coward]

AMD is better for documentation. NVIDIA is better for something which actually reliably works.

If you want awesome drivers, go Intel. Performance will suffer and many games will be unplayable. If you want performance first, go NVIDIA which has good drivers and awesome performance. If you want crappy, unreliable drivers and slightly superior performance with slightly less heat than NVIDIA, go ATI.

Choice exists. It so happens that Intel and NVIDIA sit at the intersections which matter for the well informed.

Re:In Linux drivers, Intel is still king.

[Anonymous Coward]

At the moment I'm an unhappy owner of a laptop with a AMD graphics card that can't do anything because the drivers are useless.

Install Windows and then do whatever you want.

Re:In Linux drivers, Intel is still king.

[Cajun Hell]

weren't all the ATI (now AMD) fanboys claiming a couple years ago that because ATI was developing more "open" drivers that they would rule the linux landscape?

That merely killed Nvidia. Intel dodged that bullet and then shot it right back at AMD, better aimed and with an explosive warhead attached. It passed through AMD (wounding them) and and then exploded inside Nvidia's skull.

Open is necessary, but if you intend to open (AMD) while your competitor (Intel) actually does it and then also writes the drivers too (!), then you're too slow. Too little, too late.

Last I heard, talking to AMDs video decoder was still a proprietary trade secret. Video decoders, as late as 2012! As if people are still waiting to build their HTPCs. AMD hasn't even entered the market yet, while Intel is selling. There go your customers.

Re:In Linux drivers, Intel is still king.

[Jeng]

This story is regarding AMD CPUs, not AMD GPUs.

Currently Linux supports the features of AMD's current CPUs better than it is supported on the Windows side of things.

Re:In Linux drivers, Intel is still king.

[amorsen]

Yes, but ATI didn't develop open drivers after all. Intel did.

Re:In Linux drivers, Intel is still king.

[cynyr]

but i like my GPUs to draw 3D things (mostly via wine) so i got a nvidia.

Re:In Linux drivers, Intel is still king.

[unixisc]

Well, they could be using Intel CPUs and AMD GPUs.

Re:In Linux drivers, Intel is still king.

[BadgerRush]

I'm afraid you are wrong, the story is regarding the AMD Bulldozer family of CPUs and APUs, not only AMD CPUs.

Re:AMD boards have better PCI-E lanes then intel c

[smash]

The thing is, the low end don't care for masses of PCI lanes. They run integrated video. The high end want a fast CPU as well.

Re:AMD's next CPU will crush protesters into rubbl

[Anonymous Coward]

It's an interesting story but I think in general Americans are tired of what happens in the Middle East, and some idiot protester getting herself killed garners as much sympathy as those idiot pot-smoking hikers who got caught trespassing. If somebody dies doing something dumb, they tend to get a yawn in response.

The best policy for this country is to stay out of other affairs and work on our own problems. That's why support for the war, occupying Afghanistan, and the Israel/Palestine thing are so low on the radar for most Americans.

I wish we had world peace, but to be honest I'm more concerned about fixing the problems in my backyard and not half the world away. Plus those Israeli/Palestine conflict will never end and both sides do some heinous stuff, so it's a waste to care too much about it. Both sides are monsters, there can't be a winner. It's like having Hitler wrestle Stalin, hard to root for either one.

AMD's in deep trouble with Steamroller

[PhrostyMcByte]

I think AMD's work here will provide some great evolutionary speedups that will be significant to many people. Unfortunately for them, at the same time AMD is bringing out these small "free lunch" general improvements, Intel will be bringing out Haswell -- which in addition to such evolutionary improvements has some really fantastic, significant new features that'll provide remarkable performance boosts.

• Integer AVX-256. For apps that can take advantage of it, this'll be a massive speed-up. Things like x264 and other video processing will take huge advantage of this.
• SIMD LUTs. One of the major optimization tricks programmers have been using for ages, look up tables, have been thus far out of reach to SIMD code without complex shuffle operations that usually aren't worth it.
• Transactional memory. This is not quite the easy BEGIN/COMMIT utopia people are hoping transactional memory will eventually get us, but it's a building block that'll enable some advanced concurrent algorithms that either aren't realistic now or are so complex that they're out of reach to most coders.

These are all pretty specialized features, yes, but they service some very high-profile benchmark areas: video processing and concurrency are always on the list, and AMD will get absolutely crushed when apps start taking advantage of it.

I'm a developer, a major optimization geek both micro- and macro-. I thrive playing with instruction latencies, execution units, and cache usage until my code eeks out as much performance as possible. Of course we'll never know until the CPUs are released for everyone to play with, but right now my money is on Intel.

AMD is in serious trouble here. I hope I'm wrong.

Re:AMD's in deep trouble with Steamroller

[VortexCortex]

I'm a developer, a major optimization geek both micro- and macro-. I thrive playing with instruction latencies, execution units, and cache usage until my code eeks out as much performance as possible. Of course we'll never know until the CPUs are released for everyone to play with, but right now my money is on Intel.

Yeah, I'm a developer too. However, my simulations run on desktops not super computers so it doesn't matter how optimal the code is on a single particular piece of hardware... Wake me up when there's a cross platform intermediary "object code" I can distribute that gets optimised and linked/compiled at installation time for the exact hardware my games will be running on.

We need software innovation (OS's and Compilers) otherwise I'm coding tons of cases for specific hardware features that aren't available on every platform, and are outpaced by the doubly powerful machine that comes out 18 months later... In short: It's not worth doing all that code optimisation for each and every chip released. This is also why Free Software is so nice: I release the cross platform source code, you compile it, and it's optimised for your hardware... However, most folks actually just download the generic architecture binary, defeating the per processor optimisation benefits.

Like I said: In addition to hardware improvements we need a better cross platform intermediary binary format (so that both closed and open projects benefit). You know, kind of like how Android's Davlik bytecode works (processed at installation), except without needing a VM when you're done. I've got one of my toy languages doing this, but it requires a compiler/interpreter to be already installed (which is fine for me, but in general suffers the same problems as Java). MS is going with .NET, but that's some slow crap in terms of "high-performance" and still uses a VM.

Besides, I thought it was rule #2: Never optimise prematurely?
(I guess the exception is: Unless you're only developing for yourself...)

Re:AMD's in deep trouble with Steamroller

[lightknight]

*Looks around* AAAAAAAnd, how does this AVX-256 compare to OpenCl transcoding of video?

Re:AMD's in deep trouble with Steamroller

[TeXMaster]

*Looks around* AAAAAAAnd, how does this AVX-256 compare to OpenCl transcoding of video?

That's a stupid question. OpenCL by itself does nothing whatsoever to improve video transcoding. OpenCL is an API, so the performance of an OpenCL kernel for video transcoding highly depends on which hardware you're running it on. On Intel CPUs supporting AVX-256, OpenCL kernels will be compiled to use those instructions (if Intel keeps updating its OpenCL SDK), on GPUs and APUs it will use whatever the respective platforms use. What OpenCL does is make it easier to exploit AVX-256, just as it makes it easier to exploit SSE and multiple cores.

Re:AMD's in deep trouble with Steamroller

[aaron552]

Wake me up when there's a cross platform intermediary "object code" I can distribute that gets optimised and linked/compiled at installation time for the exact hardware my games will be running on.

I hear LLVM is heading in this direction

Re:AMD's in deep trouble with Steamroller

[Anonymous Coward]

From what I've seen, GPU transcoding is currently shit.

Sure, you have plenty of GPU transcoders.. But the quality of the output is bad. Really Bad. Worse, it's inconsistent.
This may improve as coders get used to programming on GPUs, but as it stands you can get faster output from traditional CPU encoders if you tune down their quality to the level of what most GPU encoders spit it now.

Re:AMD's in deep trouble with Steamroller

[Anthony Mouse]

I believe the point is that an OpenCL transcoding algorithm running on a typical GPU will make doing it on a CPU look silly and pointless, so who cares how fast the CPU can do it when you're going to do it on the GPU anyway?

Re:AMD's in deep trouble with Steamroller

[PhrostyMcByte]

Your question is a bit... difficult.

GPUs can definitely excel at many forms of video processing. Encoding, thus far, hasn't proved to be one of them. Currently, CPU-based encoders are faster and of significantly higher quality. I'm sure someone smart will make a fantastic GPU-based encoder eventually, but so far nobody has come close. A few companies have lied and/or used faulty benchmarking to help people believe they have, though!

Resonant Clock Mesh?

[elfprince13]

Is that new resonant clock mesh technology still planned to launch with this new series? I remember reading they were planning to break the 4GHz barrier?

Re:Resonant Clock Mesh?

[Osgeld]

I just got a 4.2Ghz AMD chip for 140 bucks last week ... Its not stupefyingly awesome, kicks the shit out of my old phenom II 720

Re:Resonant Clock Mesh?

[edxwelch]

Why wouldn't it be? It's already in Trinity

Re:Resonant Clock Mesh?

[elfprince13]

oh, huh, I guess I missed that somehow.

Re:Resonant Clock Mesh?

[DL117]

What 4Ghz barrier? I have an Intel i7 3700k at 4.6Ghz, and before that I had a Bulldozer at 4.5Ghz(which I promptly returned due to it's horribleness.)

Re:Resonant Clock Mesh?

[Targon]

Overclocking compared to the stock speed of 4GHz is two different things. There WERE some issues that held back overclocking in older chips(4GHz was almost a hard limit for some reason), but that has been fixed in the newer chips. Still, the AMD FX 8350 running a stock 4GHz with turbo mode to 4.2GHz should be interesting with the new Piledriver improvements over Bulldozer. That is something that should be interesting to see, just because it may fix all the performance problems with Bulldozer.

It's a Start

[ndykman]

Having dedicated decoders for the IPU is definitely on the right rrack, but a shared fetch is still means there is a bottleneck in getting those cores fed.Also, apparently, the changes hit the L1 performance, so they had to add some cache to make up for it. So, there is some room for improvement, and this does help, However, I just don't see it as the big step that AMD needs against Intel. Intel's dies are smaller, they are making better use of space, and this is a huge advantage. Intel has 10 core dies, and has room to go to 12. In the server space, this is a big gain in density.

Also, they have a lead in hyperthreading, which does seem to boost multithreaded performance enough in highly parallel server scenarios to hold up against the chip multithreading concept that Bulldozer supports.

What is strange is that AMD need to really get a vision of a unified platform. AMD could really have a great ultrabook and media PC platform, but against, they let Intel take a lead in this area.

Re:It's a Start

[lightknight]

Bah, what AMD needs to do is just keep doubling cores on the Phenom-line of chips. A 12-core Phenom III, in the next 12 months, could keep them going for another two years or so. Of course, then they'd think that it might impact their server offerings, but lets be honest, I've looked at their server offerings, and while I love the number of cores, I need at least 3 Ghz cores; the 2.1 Ghz cores make me question whether it's worth just buying multiple machines with Phenoms, as opposed to buying Magny-Cours (or whatever we're up to).

Do you hear me AMD? Fix the goddamn FX processors, I want full decoders & all the bling, I want more cores for the Phenoms, and I want faster server processors. Yes, yes, I know MS and the OSS community are trying to fix the FX kludge (which I upgraded to from a Phenom II X6; the FX's also appear to hate Opera, as it keeps freezing on my machine (as well as Explorer (File Explorer, not Internet Explorer)...keep getting cross-threading exceptions...IDFK, but it shouldn't be there).

And look into optical chip components.

Re:It's a Start

[Anonymous Coward]

A faster server processor doesn't necessarily mean faster cycles, but more cache. IBM's Power7 chips can have up to 32MB with 8 cores. Yes, they also get up to 5Ghz, but I'd much rather have the cache. There's only so much work a processor can do before it needs to dump its results and load new data.

I've was eyeing Intel's E3-1220 17 watt processor, which runs at 2-something Ghz. If it had more cache it'd be awesome. As it is I'm gonna go with one of the 69W, 1220v2's. I don't really care what the speed is. The biggest expense with colocation is power, anyhow.

Re:It's a Start

[Shinobi]

The biggest cost will ALL server facilities is power, floorspace and cooling. No matter how many idiots will keep bleeting "manpower is expensive", the fact remains that manpower is generally a fixed cost no matter what hardware you get, while power, cooling and floorspace will be variable depending on what hardware you get. And right now, it doesn't matter if AMD gives you more cores because Intel does more through a more efficient architecture that can be more easily fed to maintain maximum throughput, at less overall power use, and requires less cooling.

Re:It's a Start

[Targon]

Piledriver(not in an APU) comes out in October of THIS year(2012) with the 8350 set to be released at 4GHz and a turbo mode to 4.2GHz speed without overclocking. Steamroller will be the next step after Piledriver for next year. It is almost a given that improvements to performance per watt will happen every YEAR, so what comes out in a 125 watt max this year will be a 90 watt chip next year for the same performance, possibly even going below that level depending on improvements in the process technology.

The big questions will be in overall performance improvements, such as moving from modules(with shared resources between two cores) to each core getting its own resources. Fab process advances outside of Intel will be a key for that, since a Piledriver design on a 22nm fab process would be a huge improvement in size and power which could also allow for larger cache sizes, which would help with overall performance.

You know the Intel tick-tock strategy, fab process one year, core design improvements the next, back and forth. AMD has been limited in not getting those fab process improvements to allow for better speeds and to allow for design improvements that would require more transistors to make work.

Other things that could happen

[Anonymous Coward]

Unicorns *could* pop into existence.
Cheap flying cars *could* be invented.
Gravity *could* cease to exist.
The dead *could* come back to life.
Etc., etc., etc.

Lots of things *could* happen. It doesn't mean that they will happen or even that they are remotely possible. "Could" in a headline is a cheap way to make you think something is likely to happen or is just around the corner. It isn't.

AMD Brewin Something

[Anonymous Coward]

Steamroller

What are you talking about?

[Sycraft-fu]

That is like asking "What colour do you want this database in?" OpenCL doesn't do anything for transcoding video, it is an API for talking to graphics cards. Now, GPUs can be used for video transcoding, of course, using OpenCL or other APIs (CUDA, DirectCompute). However how well they do depends on what the graphics card is. An AMD 7970, that'll throw down some serious performance. An ATi 3400 doesn't even have driver support for OpenCL and if it did would be very slow.

So there isn't going to be any way to directly compare speeds because the speed of something using OpenCL will depend on the GPU that runs underneath it. On some systems that may be very fast on others it may not be available.

Also along those lines a difference from the user standpoint is that it is integrated in the CPU. You don't have to go buy a GPU, the CPU itself just does it. This is something that is rather nice, and hopefully some day we reach the point where CPUs are powerful enough that we don't need GPUs. The more a CPU can handle well by itself, the less people that have a need for an addon GPU.

Re:What are you talking about?

[lightknight]

+10 for being pedantic (the best kind of correct, technically correct), -1000 for knowing exactly what I was groping for, but choosing to be pedantic.

Just got back from a late-night concert, and my head hasn't stopped pounding yet (and there is some question of sobriety -> Jimmy Buffet with margaritas). Besides, and I am summoning my inner BOFH here, who teh f*ck would run OpenCl code on a CPU? I've tried, and the only thing I've succeeded in doing is giving my laptop a grand mal seizure.

And no one sane does video-transcoding on a 7-year old machine. No, no, just don't go there.

And ponying up an extra $500 on top of the regular CPU going rate ($200-300) for a new chip, from Intel, when a $150 four-generation displaced video card could / would spank it is a thought not even worth considering. But I digress, someone out there will decide that running a video transcoder, on a non-upgradeable laptop (which they will pay way too much money for this chip), with Intel HD integrated graphics (does it even support OpenCl? Is it still a separate chipset, or has it been integrated on-die?), and absolutely need this feature; they will also probably save the processed video onto a 4,000 RPM USB-1 portable value hard drive.

Comments subject to revision if / when I wake up tomorrow, and shake off that last of this Tequila. I think it's Tequila.

Re:What are you talking about?

[hairyfeet]

And if you look at the reviews you'll also see that doing video transcoding on the GPU? Really kinda shoddy. Every test I've seen had some okay, some lousy, but none of them did as good a picture quality or as small a file size as pure CPU.

This is why I'm happy with with my Thuban, 6 cores let me chew through transcodes without any trouble and with great picture quality. While I'm all for using GPUs for decode, my E350 really saves battery life by using the GPU, but transcodes? The tech just isn't there yet.

Re:What are you talking about?

[FreonTrip]

Heads up: The x264 project's incorporating OpenCL support for certain parts of the encoder. Take a look over here - initial results are very promising.

Re:What are you talking about?

[Atzanteol]

+10 for being pedantic (the best kind of correct, technically correct), -1000 for knowing exactly what I was groping for, but choosing to be pedantic.

Welcome to Slashdot!

Re:What are you talking about?

[DuckDodgers]

Agreed. I think some of the things Intel (and AMD, and Qualcomm, and nVidia) put into their CPU instruction sets are 10% for real use and 90% so they can put an item on the bullet list for the marketing literature and get fanboys to buy something new when the thing they already had is just fine.

It was a great Buffett concert, wasn't it? But I'm getting too old for this, after two hours I couldn't hear much beyond the ringing in my ears. I still have a bit of it today. And also, I thought movie theaters with bars charged a lot for booze, but that was ridiculous.

Re:What are you talking about?

[Anonymous Coward]

Where mental masturbation is done out in the open.

Re:What are you talking about?

[lightknight]

Indeed. Jimmy had an amazing amount of energy; I'm jealous of him, at that age, displaying more energy than I currently have. Plus, all the backgrounds made me kind of sad I won't be visiting the various paradises any-time soon.

And yes, the booze was ridiculous. $14 for a guitar, $30 for a bottle of wine, $45 for a 'premium' bottle wine, ~$14 for a drink with two shots in it, more for a drink with two 'premium' shots in it, $6 for a red bull. For contrast, the wine & spirits shops down the road from me sells a very drinkable reitzling for ~$7.

I can only imagine that the markups must be something ridiculous (1000% profit over costs). Although I must admit that despite the venue being in Camden, NJ, it was nicer than the Mann center (Further); I feel like I am going to die every-time I visit there (the heat is unbelievable, and the food is just terrible). I remember buying a ~$5 power-ade at the Mann during the heat wave, realize it was warm (i.e. environmental temperature), asking for a cold one, and the concessionaire looks me straight in the eye and tells me that they have no cold ones left. So, I just bought a ~80-90 degree thirst-aide, because the Mann is too cheap to put a few extra freezers in, when it's the middle of summer, and we've been having 100-degree heat waves. Thanks guys. And the hot dog...Oh God....I think it died.

And it did not escape my notice that a bank had naming rights to the place in Camden. Not a good time, economically speaking, to have your bank associated with such frightful gouging of customers. Gah, even Regal has gotten smart about the insane pricing of food, and has been offering free popcorn / drinks / candy as of late. Going out to the movies, for two, it's now around $40, while attending a concert (not on the lawn) is around $200-300 for tickets, plus $30 if you and a friend only have one drink each. My wallet is, heaven help me, safer in a casino than these venues.

     

Re:What are you talking about?

[DuckDodgers]

I think one of the many things that makes Jimmy appealing is that he clearly is grateful to his fans for the life they let him lead. The only thing I did not like is the lack of at least one request that nobody drive home drunk. I don't think he should be forced to do that, but I think it's a sensible thing to do. There were a lot of blitzed people at the concert.

I had to pay a babysitter to watch the kids for the 90 minutes it took to drive to the concert, the concert itself, and the 90 minutes to drive back. I don't live near a train station, so traveling that way was not an option. Total cost of the evening, tickets, drinks, parking, gas, and tolls was $450. I had a great time, but I don't think I'll ever do it again.

And yes, I think having a bank name associated with those prices was ludicrous. I was also miffed that they had dozens of staff members and lots of police directing traffic before the event, and almost none after. Clearly they were only trying to make life easier on the customers until they had our money - once they had our money, they had no interest in facilitating anything.

I hadn't realized the Mann center was so bad. Yikes.

And an average user would care because?

[Sycraft-fu]

Sorry but lots of PCIe lanes are just not the kind of thing that matters to non-high end users or people who focus on stats rather than real world performance. To even have a situation on a desktop board where it could theoretically matter you have to have multiple graphics cards. The 1x slots hang off the southbridge and have their own bandwidth separate from the lanes on the CPU for the video card. So if you stick on two GPUs then yes, you don't have enough to give them both 16 lanes.

However it turns out to not matter. We have more bandwidth than we need with PCIe, particularly now with 3.0. You test a card in 16x vs one in 8x and you find no difference in performance. So it just doesn't matter even if you have multiple GPUs.

Of course multiple GPUs are rather a high end proposition. Many people don't even bother with a GPU at all, they just use the onboard graphics which these days are surprisingly good (I've played with the integrated graphics on my new Ivy Bridge laptop since it has switchable graphics and for many games, you don't need anything more). Even those that do choose to have addon GPUs, most choose just one. I've got a GTX 680 in my desktop and there is just no need for anything more, it handles all games superbly. I'd have to move up to multiple surround monitors or something before I'd start needing more than one GPU.

So it is a situation where you only end up needing more lanes in a high end environment, thus I don't see the big deal in not having them. It is the kind of thing you won't notice.

Re:AMD boards have better PCI-E lanes then intel c

[Z34107]

What kind of workload needs more than 16 PCIe lanes, but doesn't similarly need a higher-end processor?

The real question

[fa2k]

Will it be faster than the Phenom II 985?

I would like to stay with AMD because they don't usually limit the features to the high-end or server CPUs. For example, I can use ECC on my workstation. Performance matters, though, and it doesn't seem great unless they do something like 20 cores for a reasonable price.

Re:The real question

[Targon]

Piledriver in October of 2012 should answer your question about performance, so you won't need to wait for next year. 8 cores at 4GHz without the scheduler problems SHOULD beat the Phenom 2 generation, but we have another 1-2 months before we know for sure what the performance will be. Socket AM3+ does mean that DDR2 will finally be fading away, so many of us with older systems WILL need all new motherboards and memory on those older machines that didn't get updated yet.

Kaveri

[tchiwam]

Interesting as this is the Finnish word for friend. Finnish cooperation here ?

Re:Kaveri

[Anonymous Coward]

Finnish? Kaveri and Krishna (the project that died alongside Brazos) are both rivers in India.

One more reason why APU in the high end is stupid

[Lonewolf666]

Personally I thought the whole idea was retarded except for the mobile chips like Brazos, on the desktop the idea was completely stupid and on the server even more so. For those that don't know the original plan was to go "Full APU" and have the GPU take the place of the FP on chip, which would be a much simpler and weaker design than in years past thus freeing up more TDP for more cores. Why is this dumb? Well what if you want to use the GPU AND do some floating point heavy task? Or what if you don't want the integrated GPU because you can't OC worth a crap with the GPU built in?

All correct, but I could live with those aspects. I usually don't OC, and if I know I want the GPU AND do some floating point heavy task, I could get an additional discrete GPU. There is, however, a worse one:

Memory bandwidth congestion. A typical lower midrange graphics card with 128 bit data bus and GDDR3 is significantly slower than the same model with GDDR5. In an APU, the GPU part has to share the even lower bandwidth of the DDR3 main memory with the CPU part.

When the LLano was new, Anandtech published a preview:
http://www.anandtech.com/show/4448/amd-llano-desktop-performance-preview
It shows some comparisons to discrete graphics cards, including the HD 5570 which represents the lower midrange graphics card w/128bit mentioned above.
In gaming frame rates, the HD 5570 beats the LLano even when it runs on DDR3-1866 RAM, which was not a JEDEC standard at the time. With standard DDR3, the difference gets bigger. Which shows clearly the LLano is limited by memory bandwidth and really could use four-channel memory as in Intel's socket 2011.
With bigger and faster GPUs, the bandwidth demands will only grow, and for a Bulldozer APU with matching GPU part even four-channel memory may be insufficient..

High School

[gooner666]

I had a steamroller in high school. That thing kicked your ass.

Re:AMD boards have better PCI-E lanes then intel c

[Targon]

I am guessing those who want to do more GPGPU type stuff, so if you get a supported video card or multiple video cards, then you can potentially get some great performance. Of course, if running dual-GPU stuff is what you want, you should NOT be bothering with a cheap motherboard.

Re:One more reason why APU in the high end is stup

[hairyfeet]

Wow...ouch. Thanks for the heads up, I knew they were having bandwidth problems (the net is filled with advice and tutorials on getting the absolute fastest memory your Brazos or Liano will take because it seriously boosts performance, which I can attest to when I swapped the stock 1066 for 1333 in my Brazos netbook) but I hadn't thought ahead to what it would take to feed an 8 core AND a GPU on the same die...ouch.

But that just proves my point though, that in any place other than mobile situations it is a DUMB idea. It works in mobile because battery life is king and by being able to shut down a good portion of the GPU and FP when you are just doing basic web surfing you cut down on both power and heat, while still having the ability to activate the GPU when you wanna watch that video on the bus or use your netbook over HDMI as a media tank.

The problem is what works in mobile does NOT work in desktop and server, its different use cases. And in desktop and server roles as the link you provided shows (thanks again, good read) the bandwidth problem takes an otherwise powerful unit and gimps the living hell out of it. Since we know they will have to have even more powerful parts to compete with Intel that means by using the GPU as the FP the bandwidth is just gonna get worse, as now you have both the more powerful GPU AND a multicore CPU both fighting over a seriously limited pipe...ouch again.

Yeah...no bones about it the APU design for anything other than laptops and basic office boxes is a BAD design.

Reading comprehension failure

[dbIII]

Please reply to something related to what I've written instead of just attaching some stuff about the low end servers I was not writing about. The words "massively parallel" should have been a clue. The cost difference IS that large in the sector I was writing about. There is nothing in my words above that even hints at anything outside of that niche.

Re:Reading comprehension failure

[trilliwig]

AMD doesn't exist outside of the low end server space. Precisely what sector were you talking about, where you can price out an AMD machine on the web? The machines I was talking about scale out to greater than 4 sockets, and possess features aimed at the highest end of the market such as RAS and multi-year service/software support contracts. AMD's latest Interlagos models don't even support more than 4 sockets without custom interconnects and glue logic, and if you're going for more sockets than that, you're in Top500 supercomputer territory, where cost is not a consideration anyway. If it were, AMD would be doing better than it is, as that is the only reason to buy an AMD server nowadays.

Top 500 stats by processor family:

x86-64 - 435

Intel Xeon - 372

AMD Opteron - 62

If you look at the historical graphs, AMD has been falling in share ever since 2009.

One more thing

[dbIII]

The cost difference is not nearly an order of magnitude when you take into account all the other components of the server besides CPUs you have to pay for anyway.

The above costs are from quotes for complete servers, so the cost difference is as described. Do I have to put things in bold with illustrations in this place?

Re:One more thing

[trilliwig]

Please do. Otherwise, you might as well be describing invisible pink unicorns. Your statements about pricing are wildly divergent from the reality I am seeing.

Re:One more thing

[dbIII]

SuperMicro boards, 4 sockets for the AMD with 64 cores and the 40 core Xeon, 8 sockets for the 80 core Xeon. 128GB in each, chassis with less than $10 difference in price included, disks ignored.

Re:One more thing

[trilliwig]

You can't ignore disks and claim order of magnitudes of price difference, just like you can't ignore costs of power and cooling and software. You won't be running without disks, electricity, or software, I trust?

Re:One more thing

[dbIII]

Assume identical disks because that's what happens when you have systems doing identical tasks - thus disks can be ignored in a comparison. I thought that would be obvious so I just stated that as "disks ignored". Is that enough to satisfy that nitpick?
Anyway, that's where the price comparison came in around March or April for machines to run some geophysical computation that is easily done in parallel - as compute nodes they typically only have 1 or two disks. It was $9000 for a mid range 64cpu amd and $89000 for a mid range 80cpu intel, $49000 for the 40 cpu. There are more expensive amd and intel cpus that are marginally faster so you could pay even more. It's still over $8000 each for some Xeons so the price would be fairly similar now.
That's still nowhere near supercomputer territory.

Re:One more thing

[trilliwig]

Assume identical disks because that's what happens when you have systems doing identical tasks - thus disks can be ignored in a comparison. I thought that would be obvious so I just stated that as "disks ignored". Is that enough to satisfy that nitpick?

No, because the price ratios are completely different once you add in disks. X over Y is not equal to X+N over Y+N if we're dealing with positive numbers.

To bring in a car analogy (I know, everyone's favorite thing, right?), this is like subtracting out the cost of non-engine parts and claiming an electric car is 10 times more expensive than a similar non-electric car. Well, such price comparisons are disingenuous, because no one ever buys just an engine; they need the entire rest of the car!

Re:One more thing

[dbIII]

That's a very petty nitpick. Identical disks are identical, and in this case just over 1% of even the lower price.

new suggested name for the next AMD chip

[robbie73]

ASSDOZER! (from the movie "Idiocracy")