Core i5 and i3 CPUs With On-Chip GPUs Launched

MojoKid writes "Intel has officially launched their new Core i5 and Core i3 lineup of Arrandale and Clarkdale processors today, for mobile and desktop platforms respectively. Like Intel's recent release of the Pinetrail platform for netbooks, new Arrandale and Clarkdale processors combine both an integrated memory controller (DDR3) and GPU (graphics processor) on the same package as the main processor. Though it's not a monolithic device, but is built upon multi-chip module packaging, it does allow these primary functional blocks to coexist in a single chip footprint or socket. In addition, Intel beefed up their graphics core and it appears that the new Intel GMA HD integrated graphics engine offers solid HD video performance and even a bit of light gaming capability."

Comments

To GPU bandwidth?

[Anonymous Coward]

Itching to see how good these chips are at some number crunching on the GPU portion. I've always had an issue with the traditional bandwidth of system memory to GPU memory. That northbridge pisses me off.

I realise these particular chips are mobile processors.

Re:To GPU bandwidth?

[William Robinson]

I realise these particular chips are mobile processors.

I didn't RTFA, but I always wished to use Intel chips for mobile/portable devices. Last time I checked they were so power hungry almost making them useless for battery based applications (maybe except notebooks/laptops where huge battery could be afforded.) Their Canmore chip almost consumed over 30W of power...settling down only for STBs and Gaming stations.

Re:To GPU bandwidth?

[Narishma]

By mobile the parent probably meant laptop.

Intel branding considered harmful

[wisty]

Grrr ... I wish Intel would go back to their system of giving new names to new chips then adding a MHz (and if that's not enough, maybe a cache size and number of cores) to distinguish them, rather than using a weird combination new names (for their top-tier chips) and old names (for their low-end gear).

I only just realized that Pentium no longer means "crappy NetBurst", but now means "low end C2D". And later this month, there will be "Pentiums" and even "Celerons" built on the same architecture as the i5. How do you let your friends know that the "Pentium" is either a worthless, power-hungry dinosaur; or a cheap version of the i5? Should people memorize the chip serial numbers? Because that seems to be the only way of figuring out what the chip is these days.

Re:Intel branding considered harmful

[NoNickNameForMe]

That is not the only problem nowadays, even processors within a given family may or may not have specific features (VT, for example) disabled. You'd think that there is a conspiracy going on...

Re:Intel branding considered harmful

[mvar]

Indeed things back then were so much more simple. You had pentium & celeron, now you have celerton, celeron dual-core, c2d, c2q, i5, i7 not to mention all those different cores..oh god

Re:Intel branding considered harmful

[dingen]

I fully agree with this, it's absolutely impossible to fully understand Intel's CPU product line up. And why make all those different models anyway? I understand you have a branch of products focussing on power consumption and another on speed, but the current amount of different processors, brand names, code names, series, serial numbers is completely insane. Especially, as you point out, because the meaning of these names keep changing all the time!

Re:Intel branding considered harmful

[Anonymous Coward]

I think the CPU lineup goes like this:
8088, 8086, 80286, 80386, 80486, Pentium, Athlon, umm not sure if there is anything faster than that.

Re:Intel branding considered harmful

[Anonymous Coward]

You could compare Pentium 4s to each other pretty reliably based on clock speed. Sure, the Northwoods were a bit faster than the Prescotts, and the Extreme Edition chips had a nice speed boost from the cache, but generally clockspeed made em' match up.

However, turbo boost / new architectures can give a 50% speed boost on tasks like x264 encoding when you're talking core 2 vs i5. The frequent changes necessitate new naming schemes.

Re:Intel branding considered harmful

[hairyfeet]

That is one of the reasons I ended up switching to AMD. With Intel it was getting to be a PITA to figure out which were the "good" chips, which were the "okay" chips and which were the cheapos. Especially since some of their chips have VT and some don't. I like how AMD only has three lines-Phenom = (best) Athlon = (good) and Sempron = (cheapo). Plus I remember what it was like when Intel was a monopoly and do NOT want to go back!

And lets be honest, once we hit dual cores for the average Joe the PC ha passed good enough a few miles back. Checking the logs on my customer's PCs on followup even the duals are spending a good amount of their time twiddling their thumbs, because the average user just doesn't come up with enough work to keep them fed. And with the economy in the crapper my customers like how cheap the new AMDs are. Hell you can get a quad for $99!

And as far as these new chips go, does Intel want to get a monopoly charge dropped on it? I mean here they are, being investigated left and right, and the come out with a whole new line of chips with onboard GPUs which looks like it is just another shot at locking out Nvidia. It sure as hell smells to me like trying to lock up the chipset market for themselves. I predict if Intel doesn't get a serious smack down from the EU or Justice Dept that it is gonna end up just them and AMD unless Nvidia buys Via and tries to get in the game that way. Does ATI even make chipsets for Intel boards since being bought by AMD? I know they locked Nvidia into the dead end LGA775 and basically give up. So is there anyone besides Intel making chips for the new socket?

Re:Intel branding considered harmful

[cowbutt]

Even worse than that, at least one model, the Q8300 Core2Quad both does and does not have VT, depending on the sSPEC code; SLB5W doesn't, SLGUR does. Good luck trying to buy one of those online and being sure of what you're gonna get!

Re:Intel branding considered harmful

[Anonymous Coward]

All the 5xxx Pentium Dual Cores as well AFAIK. It's turning into a really annoying crapshoot, especially given that a number of retailers have had cpu mobo combos on sale with no way to determine which SSPEC they're actually stocking :(

Re:Intel branding considered harmful

[wisty]

You could compare Pentium 4s to each other pretty reliably based on clock speed. Sure, the Northwoods were a bit faster than the Prescotts, and the Extreme Edition chips had a nice speed boost from the cache, but generally clockspeed made em' match up.

However, turbo boost / new architectures can give a 50% speed boost on tasks like x264 encoding when you're talking core 2 vs i5. The frequent changes necessitate new naming schemes.

That was the problem, I think. They were rubbing Moore's law in people's faces. Now, the i7 can always be $999, the i5 can always be $250, and the Pentium can be always be $100. People won't feel like idiots for lashing out on a marginally more powerful system, because their i7 will always have a superior sticker to the lowly Pentium.

Re:Intel branding considered harmful

[71bigblock]

Nvidia has a license to make QPI-based chips. whether they have any real plans to do so.....

Re:Intel branding considered harmful

[DrMrLordX]

No, ATI/AMD does not make LGA1156 or LGA1366 motherboard chipsets. Nobody but Intel does, in fact.

Re:Intel branding considered harmful

[sznupi]

What? With Pentium it was easy - there was a year or so long break between using this brand for Netburst and for Core architecture. For around 2 years already anything new & under Pentium brand gives you nice, cheap, C2D CPU...perfect in typical laptops. Yes, it's slightly slower, but together with Intel GFX and slow HDDs it doesn't matter.

Intel of course wasn't really promoting those CPUs, wishing from you to overpaid for full C2D, but they weren't secretive about them either.

Re:Intel branding considered harmful

[tyrione]

Nvidia has a license to make QPI-based chips. whether they have any real plans to do so.....

No they do not. That was the reason they bailed out of the controller market.
Intel licensed SLI from Nvidia for their QPI based Core processors.

SANTA CLARA, CA—AUGUST 10, 2009—NVIDIA Corporation today announced that Intel Corporation, and the world’s other leading motherboard manufacturers, including ASUS, EVGA, Gigabyte, and MSI, have all licensed NVIDIA® SLI® technology for inclusion on their Intel® P55 Express Chipset-based motherboards designed for the upcoming Intel® Core i7 and i5 processor in the LGA1156 socket.

http://www.electronista.com/articles/09/12/16/ftc.ignores.amd.settlement.in.intel.suit/

The FTC notes the lawsuit is not a direct antitrust case and only accuses Intel of violating competition and monopoly rules under Section 5 of the FTC Act. As a result, it prevents other companies from 'piggybacking' on the lawsuit by using an antitrust decision to demand triple damages in any private cases. In pursuing the complaint, the government commission is hoping to ban Intel from engaging in unfair bundling, pricing and exclusionary licenses and could, if victorious, force Intel to allow NVIDIA chipsets like the GeForce 9400M and Ion for Core i3, i5 and i7 processors as well as future Atom designs. Companies like Apple have faced the possibility of mandatory major reworkings of their computers to continue using modern processors.

Re:Intel branding considered harmful

[Anonymous Coward]

As was the case with Core 2 Duo T5x00/7x00

Re:Intel branding considered harmful

[Kjella]

And as far as these new chips go, does Intel want to get a monopoly charge dropped on it?

The writing has been on the wall for a while, it will all be integrated into one chip at least on the low end. Oh sure Intel might get slapped one way or the other but by the time the dust settles it'll all be on a <30nm chip and no court will manage to force them to create discrete chips again.

The other part is games but the chips are running ahead of eyes and displays and developer time, if you looked at the latest reviews they only test at 2560x1600 with full AA/AF. I'm sure Fermi will be impressive but 30" displays is a tiny niche and the rest don't need it.

nVidia is talking about supercomputers and GPGPU but they're going the way of Cray and SGI, into some niche where they'll slowly wither away. AMD will hang in their because their CPU/GPU combos beat Intel on the GPU part.

Re:Intel branding considered harmful

[Anonymous Coward]

you missed i3, atom and xeon (in it's various X and W guises) ....

Re:Intel branding considered harmful

[TheKidWho]

Last I checked, 1080p displays where becoming the norm for PCs...

Re:Intel branding considered harmful

[Anonymous Coward]

It's just Schroedinger's processor. You have to watch it closely to permanently enable or disble VT.

Re:Intel branding considered harmful

[Rockoon]

1080p is quite a bit less than the 2560x1600 that the poster was talking about. In consumer terms, its comparing 2 megapixels vs 4 megapixels.

Also, last I checked, the largest PC gaming segment still runs at 1280x1024 (presumably on commodity 5:4 aspect LCD's which stormed the market several years ago.) Only 12% run at 1080p or higher resolution. (source)

The 512MB NVIDIA 8800GT is probably still the best bang-for-your-buck card on the market given the resolutions people are gaming at. The 8800GT handles every game you can throw at it just fine at 1280x1024.

Re:Intel branding considered harmful

[Anonymous Coward]

Don't be dense. Reducing component count by integrating functionality provides enormous improvements in cost, reliability, power efficiency etc. There is no good reason for external GPUs except the immaturity of the technology.

Re:Intel branding considered harmful

[PopeRatzo]

And lets be honest, once we hit dual cores for the average Joe the PC ha passed good enough a few miles back.

When has that not been the case with PCs?

They've always been way past good enough for the average user.

Re:Intel branding considered harmful

[TheLink]

If VT matters to you, don't buy it. Buy something else instead.

Re:Intel branding considered harmful

[Andy Dodd]

Yup. And for a while, Pentium meant "Core Duo, not Core 2 though".

It's why I've been VERY wary of anything but "Core whatever" branded CPUs, to Intel's detriment - I've held off on an HTPC purchase for a while because the cost was more than I was willing to justify, and Adobe Flash is iffy unless you have a LOT of CPU horsepower...

Re:Intel branding considered harmful

[MikeBabcock]

No flame intended to the Intel fans, but this is one thing I find much simpler with AMD's nomenclature.

Re:Intel branding considered harmful

[gTsiros]

is VT really significant? VMware on a CPU without VT works just fine, doesn't it?

Re:Intel branding considered harmful

[KarmaMB84]

That is until developers decide they have uses for the spare CPU cycles and memory. Don't worry, we'll be bogging down 8-core CPUs with 12 GB of RAM like pros within a couple years.

Re:Intel branding considered harmful

[Sinning]

And you also missed core duo (not the core 2 or the celeron dual core). Hooray for clarity in marketing.

Re:Intel branding considered harmful

[alc6379]

is VT really significant? VMware on a CPU without VT works just fine, doesn't it?

Yes, but there are many applications where having VT will improve the performance of the VM. If you do a lot of virtualization, you'll definitely want it.

Re:Intel branding considered harmful

[ae1294]

number of retailers have had cpu mobo combos on sale with no way to determine which SSPEC they're actually stocking

Send it back and demand a full refund else charge it back. Let the retailers deal will Intel's BS.

Re:Intel branding considered harmful

[Elbows]

VT lets you run a 64-bit guest OS on a 32-bit host OS. It probably has some performance benefits, too.

Re:Intel branding considered harmful

[Khyber]

nVidia is guilty of this as well. You'd think that any 8-series GeForce would outperform any GeForce 6 series card - but my 8600 is actually running slower than my older 6800. Even something as simple as Frets on Fire just CHOKES on what should be a superior GPU.

Re:Intel branding considered harmful

[Anonymous Coward]

hmm, guess i got bad info:

http://www.bit-tech.net/news/hardware/2008/08/29/nvidia-has-a-qpi-license/1

Re:Intel branding considered harmful

[apoc.famine]

Heh...I just got into this exact discussion yesterday with a friend of mine.

Should people memorize the chip serial numbers? Because that seems to be the only way of figuring out what the chip is these days.

That was pretty much our end idea. There's just no categorization in the names anymore. The best you can do is find a processor comparison chart at Tom's and pick one which looks like it will run your specific app well for the price.
 
It's really unbelievable that we're at this point - there was a time where I was well versed in the specs of each chip and socket. Now, it's a crap shoot when I go to buy a processor. When model numbers and parts of names get reused, and cache sizes and bus speeds change all the time, there's just no good way to keep track of them anymore.

Re:Intel branding considered harmful

[Hurricane78]

Well, I wonder why this is so hard for you... Just look at the release dates of the chips. Newer chips are nearly always better, when they have the name name. Usually you can even guess the release date, by looking at the frequency times the cores times perhaps one big feature.

I never had problems. And I remember AMD having Athlons with no difference except the serial number showing what core actually was inside. So I just bought those with the higher frequency (and nowadays more cores).

(Ok, I also used to turn 550 MHz CPUs into 950 MHz ones with only a pencil (to reconnect cut bridges). :)

Re:Intel branding considered harmful

[Hadlock]

The largest segment is technically 1280x1024 @ 21.2%, typically the highest resolution available on consumer CRTs. The next largest is 1680x1050 @ 19.98%, which is most definitely an LCD display. Technically it's not 1080p, but it's damn close for most applications. If you include all the resolutions from 1680x1050 all the way up to 1900x1200, HD, or "damn close HD" makes up a full third (36.69%) of the displays being used. The 8800 is no doubt a stellar card (I wish I'd bought one two years ago, instead of a "hold me over" 8600 until the next gen was released) but with a modern display, the 8800 is only mediocre at best for most people's gaming use.

Re:Intel branding considered harmful

[drsquare]

Not even close. The average PC has maybe a 17" monitor, driven by an AGP graphics card. Maybe you have a different definition of 'norm' to the rest of us.

Re:Intel branding considered harmful

[Neil Hodges]

Of course, the current version of Microsoft's Virtual PC requires VT in order to work at all.

Re:Intel branding considered harmful

[Neil Hodges]

My HTPC has a 65 W AMD Athlon II X2. Seems to work just fine for me, especially with VDPAU.

Re:Intel branding considered harmful

[thtrgremlin]

Like I just found out that the nVidia 8800GX is technically a 9000 series processor, having full support for directX10, but that the 8800GTX, the most powerful 8000 series card does not support directX10. Something about not wanting to use the old chips when they made the watered down 8800 cards. Makes sense, but why not just discontinue the line rather than rebrand a "compatible" card?

Re:Intel branding considered harmful

[sproingie]

The first digit in an nvidia model number is the generation, the second denotes how hopped-up the card is. The 6800 Ultra had some pretty decent memory bandwidth that isn't even matched by a mid-range 8600, but an 8800 will blow it out of the water. Suffice to say that the difference between the 8600 and the 8800 is way more than it was between the 6600 and the 6800. Your 8600 supports shader model 4 however, while your 6800 does not, so that would come in useful to newer games that take advantage of it (which is unfortunately not all that many).

Fact is though, a game like Frets on Fire shouldn't choke, and I suspect your problem is driver-related. NVidia's 19x series of drivers frankly just suck, and I would downgrade to the 18x series if you can. If you're on Linux, there's a utility that makes it easy to switch to older versions, but its name escapes me.

Re:Intel branding considered harmful

[Hadlock]

am... am I reading this in 2010? does intel even sell core 1 anything anymore? Core 1 came and went in about six months. Back in 2006 or so. You can buy a core 2 solo on the budget end, but nowadays you can buy something like a Dell Zino for $350 in basically a mac mini formfactor that will push 1080p no problem. Where have you been the last two years? Core 1 Solos had no problem pushing 1080p with a proper video card, a Core2 solo (when you can find them) shouldn't have any problem either with a more midrange card. Your post is the most bizzare piece of FUD I've read in a long time.

Re:Intel branding considered harmful

[Rockoon]

Yes, but even "close to" 1080p is nothing like the resolution the poster was talking about. Literally twice as many pixels or more. If the game isnt bottlenecked on transforms or something on the CPU end, then we are talking about a 2x difference in framerate between 1920x1080 and 2560x1600.

Yes, the 8800GT isnt going to be doing great at these high resolutions, but the latest cards do well even at 2560x1600 and are twice as good at only 1920x1080

Re:Intel branding considered harmful

[Kjella]

If you include all the resolutions from 1680x1050 all the way up to 1900x1200, HD, or "damn close HD"

"High definition" seems to be one of those ill-defined concepts like "broadband", but the usual definition is 720p+ which even the 1280x1024 screens beat. Graphics cards have been supporting HD resolutions forever but now they just seem to go up, up and away. The latest AMD HD58xx class could probably drive a 3840x2160 (2160p) display in games - they already push that number of pixels in Eyefinity - if you could find one for less than 10,000$. That HDTV/BluRay tops out at 1920x1080 probably means we'll see many screens being 1920 pixels wide but few wider than that.

Re:Intel branding considered harmful

[TheKidWho]

AGP? What decade are you living in?

The average PC has an IGP from Intel, but they aren't exactly the target market for gaming companies. Also the Average PC most likely has a screen size less than 15" if you take notebooks into account.

Re:Intel branding considered harmful

[hairyfeet]

You ought to try the ATI 4xxx and 5xxx cards. Rock solid stable drivers, hardware acceleration for just about everything, hardware trascoding software, really nice. And they have models from $35 all the way to $500 so you are sure to find something in your budget.

After the whole bad solder fiasco and the way Nvidia tried to bone their customers I would wait awhile and make sure they got their shit together for spending decent money on one of their cards. The last thing you need is to be gaming and have the card blow and possibly damage the motherboard with it.

Re:Intel branding considered harmful

[Hadlock]

ATI's new cards are incredible for their bang for the buck; it's no suprise ATI is having trouble keeping their 5-series cards in stock. My computer uses a 380w power supply which will power my 8600gt, but I'd have to upgrade my power supply (Which probably means a new case as well) to get a modern 150w+ video card to run in there with no problem. I doubt I'd buy a new Nvidia until they release a new core that isn't just a variation of the 8800gts core. $150 buys you a lot of video card these days from ATI. Nvidia, not so much.

Re:Intel branding considered harmful

[mhajicek]

Who would need more than 640K of memory anyway?

Re:Intel branding considered harmful

[Khyber]

I've run it from 14x.xx up through 18x. It just doesn't compare in any form. Yes, games using SM4 do run better than the 6800 with the better looks, but you just can't get the resolution. on the 8600, I'm stuck at 1024x768, on the 6800 I can do a full 1920x1080.

I also know my 6800 has DDR3, my 8600 is using high-end DDR2.

Re:Intel branding considered harmful

[yuhong]

Anyone confused by this should read Intel's Product Change Notifications on the transition that has the exact dates, as well as the exact CPU models affected:
http://content.intel.pcnalert.com/dm/d.aspx/8A3733FB-4EEA-43E5-8493-89C7DB7B80EF/PCN109359-00.pdf
http://content.intel.pcnalert.com/dm/d.aspx/A3E94F65-56CE-4E4D-B272-2C59B2F90B0B/PCN109337-00.pdf
(Yes, Intel PCNs are public, you can search for them at http://intel.pcnalert.com/)

Re:Intel branding considered harmful

[mgblst]

Pretty sure that means you are still dealing with bullshit, unless you are just buying chips for a laugh. Getting a chip, and then having to send it back and wait for a replacement is a major headache if you actually want to use that chip.

Re:Intel branding considered harmful

[yuhong]

But note that AMD will do the same with Fusion, just much later.

Re:Intel branding considered harmful

[ae1294]

Pretty sure that means you are still dealing with bullshit, unless you are just buying chips for a laugh. Getting a chip, and then having to send it back and wait for a replacement is a major headache if you actually want to use that chip.

ugh, OK fine I'll call Intel in the morning and ask them to stop being asses but only because you said something.

Re:Intel branding considered harmful

[hairyfeet]

Have you looked at the cheap 4xxx cards to maybe give you a little more bang until you can get around to getting a new PSU/Case? I have the Gigabyte 4650HD, with 1Gb of RAM, and I ran it for nearly 4 months on a 350w PSU until I had more SATA drives than I had slots on the PSU. Games like Bioshock, HL2, the latest MoH, Swat 4, all ran cranked up and as smooth as butter, and having hardware acceleration for all the popular formats as well as hardware transcoding is a nice bonus.

But I have to agree that the "bang for the buck" on both the AMD CPUs and ATI GPUs is just nuts right now. For less than $700 after rebates I got an AMD quad with 8Gb of RAM (max 32Gb), a 4650 1Gb, dual 500Gb HDDs, and Win7 HP x64 along with dual burners. I mean you can get a fully loaded quad for less than $600! I like having the biggest ePeen in benchmarks but at that price? The amount of horsepower you get is just nuts.

Re:Intel branding considered harmful

[El Tonerino]

I have that exact processor. Only found out about the VT thing after I got it home. I was not happy. The booklet even says that it doesn't have the VT extensions.
Virt-Manager and Virtual Box say otherwise though.

Re:Intel branding considered harmful

[toddestan]

Under 15"? Are you living in the 80's or something? Netbooks haven't been around long enough to have that big of an impact, and besides I'm sure for every netbook sold, there is another desktop sold with 22-24" screen. I'd put the average at around 17", mostly due to the piles of regular 15-17" laptops around, plus all the older desktops still in use with monitors around 16-19" in size.

Re:Intel branding considered harmful

[TheKidWho]

Who said anything about netbooks?

I'd agree with about a 17" average size anyways when you compare all the notebooks with desktops...

Personally, I use a 28" screen, but that's me. Once I got used to that screen size from work, I couldn't use anything smaller.

Video decoding under Linux

[sajjen]

What's the state of video decoding support under Linux for these integrated GPUs? I've been looking for something to update my HTPC with...

Re:Video decoding under Linux

[0100010001010011]

Not sure about Intel. But Nvidia has VDPAU which is very nice. Feature Set C even added MPEG4 decoding and SD content upscaling, all in GPU (http://en.wikipedia.org/wiki/VDPAU#NVIDIA_VDPAU_Feature_Sets)

Broadcom finally released Crystal HD drivers for Linux, which means if you have a mini PCI-E slot, you can get HD content. (http://xbmc.org/davilla/2009/12/29/broadcom-crystal-hd-its-magic/)

If you want to know what is available for what GPU/Platform, keep an eye out on the XBMC guys are doing. They seem to be at the forefront of getting hardware acceleration working on different setups
http://xbmc.org/wiki/?title=Hardware_Accelerated_Video_Decoding

Re:Video decoding under Linux

[daoshi]

I just got a HTPC for me this Xmas. It's Intel Atom N330 dual core + Nvidia ION You can either build it youself or buy a system from some of the vendors. If you build youself, it's cheaper and you can get a much bigger hardrive (1TB), the pre-built systems these days usually ship with 320GB HD. But they usually got a better form factor. Mine got pefect and smooth 1080p playback. I use XBMC (xbmc.org) on ubuntu 9.10 You just need to install the lastest Nvidia driver: https://launchpad.net/~nvidia-vdpau/+archive/ppa Get youself a MCE remote.

Re:Video decoding under Linux

[Kjella]

Not sure about Intel.

I don't know about these chipsets, but the current Intel chipsets with HD acceleration support like Poulsbo and GMA X4500HD have had extremely poor Linux support. nVidia and VDPAU was really the first viable solution.

Re:Video decoding under Linux

[sajjen]

I'm currently running a HTPC based on a AMD Athlon 64 X2 4850e on a 780G based motherboard. All media is stored on a NAS. This setup works great for 720p, but it's not as silent as I'd like. I've seen a few postings about XBMC on BeagleBoard, but it doesn't seem to be in a functional state as of yet. What I'd really like is to have a passively cooled box that's able to play 1080p H.264.

Re:Video decoding under Linux

[MemoryDragon]

Passively cooled, forget it unless you can find a Tegra based system (which do not exist out of the box) Tegra can do it, but no one pushes out boards for that, because it is ARM based, and you know how easy it is to get a decent ARM board (outside of the beagle board there are none you can get), the ION is the closest you can get, add a silent active cooling and you are off even with the cheapass ATOM processors.
You might be better off in a half years time, NVidia is working on a Via based ION solution which would be better than the ATOM core.

Re:Video decoding under Linux

[asc99c]

Passive cooling hopefully isn't necessary. Consider just getting a big HSF with a big fan, to run at minimum speeds. I've got a Core 2 E6400 HTPC in a Silverstone LC11-M case. I could unplug every single case fan, set the stock cooler to the lowest possible speed (about 920rpm) and play videos without it overheating, or even getting close to overheating. At that point the noisiest item was the hard disc, even though its got soundproofing panels around it.

I've recently bought a 30GB SSD to replace the noisiest bit, along with a Nexus LOW-7000 cooler. I've now got the single 120mm fan on the Nexus running about 750 rpm as the only moving part. I know it isn't technically going to be silent, but it is now silent to my ears at least. I think big fans are the way forward rather than passive cooling. I've also got an older 80mm Papst fan in my desktop PC, which at minimum speed I just can't hear, even with my ear 1 inch from the fan. Have a look for Papst 8412NGLE - over about 10 'silent' fans I've bought over the years, this seems to be the one that remarkably does do what it says.

Re:Video decoding under Linux

[threephaseboy]

What I'd really like is to have a passively cooled box that's able to play 1080p H.264.

You mean like this?

Re:Video decoding under Linux

[sajjen]

Didn't know there was ION boards with dual core Atoms and passive cooling. Time for a late Christmas gift for myself.

Re:Video decoding under Linux

[daoshi]

There is an ION micro-ITX mother board that you can buy with heatsink only (i don't want to mention the vendor's name here). Google it.

Re:Video decoding under Linux

[LordKronos]

That one comes with a cooling fan, and from what I've heard there's a good chance you'll probably need it for the dual core model. But I think the single core is cool enough based one what I've read.

Re:Video decoding under Linux

[sajjen]

After looking closer, the board is only fanless in the pictures. The box contains a 60mm fan for the CPU.

Re:Video decoding under Linux

[MikeBabcock]

Quiet SATA drives are nearly silent these days, but fans still aren't. A large small fan draws a lot of air and while you reduce the sound of the bearings in the fan, you still have the noise of air flow over the uneven surfaces.

I have a couple completely passive boards and the noise difference from nearly-silent Sonata case + very slow CPU fan to truly silent is still very notable.

Re:Video decoding under Linux

[sunking2]

Why not get a popcorn hour or other comparable media tank? They are cheap, can be fanless, and I've not run into anything it wouldn't play.

Re:Video decoding under Linux

[b0bby]

What I'd really like is to have a passively cooled box that's able to play 1080p H.264.

This may be what you're looking for:
http://www.newegg.com/Product/Product.aspx?Item=N82E16883103228&Tpk=aspire%20revo

Lifehacker & others have instructions to put XBMC on it using VDPAU, seems really nice.

Re:Video decoding under Linux

[asc99c]

You're right of course - there is another point I forgot here. The Papst fan is silent when held up in clean air next to your ear. Once mounted into a case directly onto a grill, it is not.

On my desktop, the solution was just cut out the grill and replace with a special grill that sticks out 2cm before interrupting the airflow. In the HTPC, the fan is mounted on the CPU cooler and is separated by ~1cm from the grill in the case anyway.

Re:Video decoding under Linux

[FreonTrip]

I'm not sure what OS you're running, but good progress is being made with ffmpeg-mt, a natively multithreaded implementation of the ffmpeg libraries. The ffdshow-mt is equivalent for Windows. A friend of mine has had very good luck with that, and I hope it helps you, too.

This may not help with the silence, but should work fine for 1080p H.264 playback - if all else fails, snagging a passively cooled GeForce 9 card would give you access to VDPAU or accelerated playback through Windows.

Re:Video decoding under Linux

[Randle_Revar]

Intel chips only do XvMC (partial MPEG-2 acceleration).

ATI chips with UVD2 and running recent Catalyst drivers can make use of XVBA as a backend to VA-API. VA-API also has a VDPAU (Nvidia) backend, but VA-API is currently implemented in fewer projects than VDPAU. Only S3's Chrome 400/500 and Intel's (PowerVR's) GMA500 have native VA-API support, but their drivers are a cluster anyway.

Bottom line, unless you like binary Nvidia drivers, don't count on good video acceleration until after Gallium3D is in full effect.

Re:Video decoding under Linux

[threephaseboy]

AFAIK it's optional

Re:Video decoding under Linux

[Jorophose]

Is it running Linux? I'm pretty sure XvBA works with 1080p h.264 without a problem except maybe a possible issue with colours. (unconfirmed)

Netbook

[jlebrech]

So it's a netbook cpu/gpu combo? On a desktop isn't that a waste of transistor, because who will use this GMA POS.

Re:Netbook

[DarkofPeace]

businesses.

Re:Netbook

[beelsebob]

No, it's a laptop CPU/GPU combo, these things are aimed squarely at high end laptops like MacBook Pros.

Re:Netbook

[Anonymous Coward]

these things are aimed squarely at high end laptops like MacBook Pros.

High end? "offers solid HD video performance" is hardly a boast you'd make about high end hardware, even for a laptop. That's basic functionality. This only makes sense if you're comparing it with netbook performance.

Re:Netbook

[beelsebob]

no, it makes sense to advertise it when it's a single chip that does everything. Netbook CPUs don't offer *any* video decoding support (other than software), that's all done on the chipset (assuming you actually got an ion).

Re:Netbook

[ceoyoyo]

Midend laptops like (the lower) MacBooks. MacBook Pros have always had decent discrete graphics, which is one of the primary factors differentiating them from the cheaper Macbooks.

Re:Netbook

[Randle_Revar]

Anyone that is not a gamer and some gamers, such as myself, that play only a few older, less intensive games

Re:Netbook

[HuguesT]

Please try to keep up, even the very lower end of Apple laptops, and even the Air have had reasonable discrete GPUs for a while now.

Re:Netbook

[ceoyoyo]

Perhaps you should try to keep up.

Integrated graphics are those that share system memory. Have a look at the technical specs for the low end Macbook. See the part where it says "NVIDIA GeForce 9400M graphics processor with 256MB of DDR3 SDRAM shared with main memory?"

Now take a look at the listing for the Macbook Pro and Macbook Air. See the part where it says "NVIDIA GeForce 9400M integrated graphics?" Note that the ONLY Macbooks with discrete graphics are some of the Macbook Pros (not all of them), which have BOTH integrated and discrete graphics.

Whoopsie. Might want to check your facts next time.

What does "light gaming capability" mean?

[Anonymous Coward]

How does it stack up compared to nVidia's chips? I would consider "light gaming capability" to imply pre-GeForce performance; is that what the editors intended to convey?

Re:What does "light gaming capability" mean?

[hairyfeet]

Well if you read the specs here you will see that it has 12 execution units, which I'm guessing is Intel speak for stream processors, which considering a $30 ATI card has 320, I'm guessing like all Intel GPUs its gonna be of the uber-suck.

About the only ones I saddle piss poor Intel GPUs on anymore is the housewives, who at most are playing a browser game on Facebook. Everyone else gets an Nvidia or ATI onboard so if they decide to do a little light* gaming they can.

* The new ATI onboard GPUs are surprisingly good at gaming. I personally was playing Bioshock and Swat 4 on my 780v until I could get time to order a 4650 discrete. While these games aren't cutting edge, the fact that an onboard could actually game blew my fricking mind! Compared to the horrible chips that Intel calls GPUs it was actually nice, and it had full hardware acceleration for the most popular formats out of the box. I was impressed, and unlike so many horror stories I had heard the ATI drivers were just as solid and stable as could be.

Re:What does "light gaming capability" mean?

[Vigile]

You can see the Intel IGP compared to a low cost NVIDIA solution here:

http://www.pcper.com/article.php?aid=849&type=expert&pid=12

Re:What does "light gaming capability" mean?

[FreonTrip]

The execution units / shader cores / stream processors are unfortunately misleading. All three companies have different approaches to running shader code; while I don't know what Intel's strategy is, I'm under the impression that the large number of units in ATI's cards makes up for the fact that they run synchronously with the rest of the GPU. By comparison Nvidia has literally componentized its chips, with the "shader core" running dramatically faster than the rest of the chip.

I haven't had the opportunity to use an Intel graphics chip for any longer than a few minutes at a time in 3D, but their biggest problem seems to be driver optimization and quality. According to theoretical benchmarks their x4500 integrated part should be about level with a GeForce 6800 in terms of raw speed, but in practice it's rarely half as fast, and the more demanding the scenario, the greater the degradation in performance becomes.

Re:What does "light gaming capability" mean?

[PitaBred]

You can't compare stream processors between manufacturers. For example, the Radeon 4670 has 320 stream processors, but the GeForce GTX 280 (which is a MUCH faster card) has only 240. So really... your entire comment is based on nothing of substance.

Re:What does "light gaming capability" mean?

[ZosX]

My notebook has a Nvidia 9100M, which is on the low end of the spectrum for 9x gpus. It only has 8 shader/vertex units and seems to get along fine with anything up to Morrowind, X2, IL2, KOTOR, etc. I don't think it would handle bioshock all that well, but it does play a mean game of half-life 2. Still it is leaps and bounds over any GMA solution intel has.....

Re:What does "light gaming capability" mean?

[DrXym]

It probably doesn't stack up at all. IGP chips are fine for netbooks and budget PCs but don't expect to be able to run any graphically demanding game on them. I expect that some performance benefit comes from integrating the CPU & GPU into one chip.

Re:What does "light gaming capability" mean?

[Randle_Revar]

I prefer Intel for my light gaming, such as StarCraft, Civ3/4 and Quake3. Currently I have a X3500, which is less powerful than the X4500, which is less powerful than the GMA in Clarkdale. As the Anandtech review shows, Clarkdale holds up fairly well against the AMD 790GX (which in turn is comparable to a GeForce 9400M).

Re:What does "light gaming capability" mean?

[Anonymous Coward]

I don't have a straight answer for you but in the mobile segment, I have the previous-generation part, a GMA 4500MHD (GM45) w/DDR2 memory. On Ubuntu 9.10, it easily runs Quake Live at 800x600 at 60fps. 1024x768 bogs it down to 37 fps. You can use that as a baseline I suppose. The new part AFAICT is an evolutionary improvement.

Sockets and mobos

[grimJester]

The average consumer has little chance of realizing an i7 may need a 1156 or a 1366 socket depending on what the model number is. Those should really have been named differently.

Re:Sockets and mobos

[beelsebob]

The average consumer doesn't give a shit what socket their CPU is in either, so it's all okay.

Re:Sockets and mobos

[phillips321]

I bet they do give a shit when they try using a hammer to fit a 1366 pin into a 1156 socket!

Re:Sockets and mobos

[mooglez]

I bet they do give a shit when they try using a hammer to fit a 1366 pin into a 1156 socket!

Average consumers don't try to build their own computer from parts.

Re:Sockets and mobos

[Rockoon]

..and now we know why.

Re:Sockets and mobos

[beelsebob]

No, the why is because they're not interested.

You don't build your own car, why? Because you're not interested in building cars.

You don't build your own house, why? Because you're not interested in building houses.

They don't build their own computers, why? Because they're not interested in building computers.

Re:Sockets and mobos

[DrMrLordX]

That's a faulty comparison. Cars and houses take many well-trained hands to build, whereas a PC can be built by a single individual with little to no training in a few hours time (or less). I don't change my oil, I don't paint my house, hell I can't even fix the leaky faucet downstairs, but I can certainly build my own PC.

Re:Sockets and mobos

[jonbryce]

They buy a computer that is described as having a Core i7, and would like to know whether or not that allows them to run XP Mode in Windows 7. They don't care which type of socket the motherboard comes with.

Re:Sockets and mobos

[Deosyne]

Yeah, that PC can be built with little training if you hand that person the exact collection of parts that they will be assembling and then supervise them. Picking parts, on the other hand, is a whole other ball of wax. I'm a systems engineer whose job it is to keep apprised of new PC technologies and yet it took me quite a while to determine the final configuration for my latest build. There are a crapton of different options for most components, which is awesome for those of us who spend a lot of time learning about them as we can get exactly what we are looking for, but it is totally intimidating to the vast majority of people who can only tell you what kind of computer they have by the brand name of the case.

Re:Sockets and mobos

[JamesTRexx]

Because you're not interested in building houses.

Only until you find out you can get yours at IKEA as well.

Re:Sockets and mobos

[LordKronos]

See, that's funny to me because changing oil, painting a house, or fixing a leaking faucet take FAR less knowledge and ability than assembling a computer. Hell....my WIFE changes the oil on the car.

Re:Sockets and mobos

[Anonymous Coward]

You probably could fix that leaky faucet, change your oil, and even paint your house, but you are not interested enough in it to find out how, or to put the effort into it... You don't need to be very well trained either to change oil in most cars, nor paint a house, nor fix a leaky faucet. There are a lot of people who find those things easy to do, especially when compared to building a computer from parts. The point gp makes is that the majority of people is not interested in building their own computer, even if you (and many people on /.) find it easy. You bring cour car in to get an oil change, because that is nice and easy, not expensive, pretty fast, and you get some kind of guarantee that it will work after they are done, so you are not interested in doing it yourself. And most people feel the same about buying a computer versus building one from parts.

Re:Sockets and mobos

[PitaBred]

Really? You just have to match parts together for a PC. Leaky faucet? Is it the rings, something rusted through, do you know where the water shut-offs are, do you know what wrench to use for what nuts (they aren't labeled like the inside of a PC). Changing the oil, gotta make sure you know which is the oil pan under the car, the proper nut to remove, get the right oil filter, know how to use an oil filter wrench given most newer cars, know how much oil your car takes and how to fill it... they certainly take just as much knowledge. You just take what you know for granted.

Re:Sockets and mobos

[MeanderingMind]

That's kind of like saying you can paint the Mona Lisa by just dabbing the right colors on the canvas. If you want to make something good and avoid making potentially damaging mistakes there's an awful lot of work involved.

Another important point is that the knowledge for changing oil or fixing a pipe doesn't change unless you buy a new car or completely overhaul your plumbing. Every time you build a new computer you have to get up to speed on all the latest developments, find reviews for many individual components, check system comparisons, monitor prices etc. Unless you habitually track all of that in your spare time or plan to just throw a bunch of parts together and hope it makes sense, it's a fairly long and involved process.

Re:Sockets and mobos

[joggle]

I don't know, I never had any training and can build computers--and it used to be somewhat harder in a way because RAM used to be a PITA to install in the old days. You also don't have to deal with master/slave settings on hard drives and usually don't even need to worry too much about which SATA port you plug the drive into (especially on low-end motherboards that don't offer RAID).

Now you can go to a website like Newegg, pick a CPU for a new computer and then you'll be presented a list of combos you can buy to go with the CPU (ranging from complete systems to various combinations of RAM, motherboard, power supply, etc). You're essentially given a parts list simply by picking any component of a new computer to start with. If you aren't too concerned with the details of your new computer and aren't too knowledgeable about computer components this can be very helpful.

I know building a computer is intimidating for some people but when I compare the complexity of building a computer to other things I know how to do it seems pretty simple. Even when I compare it to changing the oil in my car it still seems pretty similar in complexity.

Changing oil:

1) Find out what kind of oil you need (depending on the season/climate you live at and the kind of car you drive), what quantity you need and what kind of oil filter to get (can determine which oil filter at the store of course)
2) Find out where the bolt in the oil pan is that you will remove to drain the oil.
3) If it's very cold outside, run the car for a minute or two so that the oil will drain without too much viscosity. If the car has been running for more than a few minutes, especially on a warmer day, let the engine cool down for a while so that you aren't burned when the oil drain plug is removed.
4) Put a oil drain pan underneath the oil plug and carefully loosen the bolt.
5) After the oil has been completely drained, remove the old oil filter. (can definitely be easier said than done on some cars where the oil filter is difficult to reach)
6) Rub some new oil around the gasket of the new filter and install, being careful to not screw on the new filter too tightly.
7) Clean the drain plug washer and rub some new oil on it, then reinstall the drain plug, being careful to not tighten the bolt too much.
8) Add the correct amount of oil to the car.
9) Turn the car on for a minute, then check the oil level, adding more oil as necessary.

I could write the instructions for building a computer, but it would probably be about the same number of steps and would certainly be easier to do from a mechanical point of view.

Re:Sockets and mobos

[Randle_Revar]

but changing oil is more frequent, and also a giant pain in the ass, despite being technically simple. Painting a house is fairly simple and rare, but very time consuming, and again, a giant pain. Fixing a faucet is rather nice in comparison, and I think building a computer would be about the same, if you don't actively enjoy it.

Re:Sockets and mobos

[Anonymous Coward]

But how can she change the oil when shes in the kitchen

Re:Sockets and mobos

[mhajicek]

I fix my car and am working on designing a new one. I fix and modify my own house. And yes, I build my own computers. The average consumer doesn't build their own computer because an off the shelf $300-$500 jobber does what they want, so building their own isn't worth it.

Re:Sockets and mobos

[mhajicek]

Kudos to parent. I spent tons of time researching my last build. Unless you're building for a specific purpose and trying to push the envelope it's really not worth it for most people.

Re:Sockets and mobos

[mhajicek]

Just match parts eh? Not booting up? Is your registry corrupted, or is your hard drive failing? Loose SATA cable? Need to adjust your memory timings? Etc. ad nausium.

Re:Sockets and mobos

[mgblst]

Really going to trust someone who can't even change their own fucking oil. Who useless can you get?

Re:Sockets and mobos

[LordKronos]

Really? You just have to match parts together for a PC.

LOL. Yeah sure. Is your power supply powerful enough for the hardware you chose? Does it have the right connector for the high end graphics card you might have bought? Do you know to install the standoffs in the case first rather than trying to screw the motherboard right to the backplane (shorting it out in the process)? Do you know whether your core i7 needs the 1366 pin socket or the 1156? Do you need DDR2 or DDR3? Does the brand of memory you bought have a compatability issue with your motherboard? Are you aware that the memory sockets on your motherboard are actually supposed to take a lot of force to insert, and that you aren't going to break it? Are you aware that you shouldn't touch the top of the cpu or the thermal paste? Do you know how to hookup all the pin headers on the case...even the ones that are poorly labeled on a lot of cases (ie: do you know whether that white arrow on the plastic means positive or negative, or that it's alright that your 2 wires for your power led connect to the header that's typically diagramed as a 3 wire connector and you should just ignore the middle connector)? That's just the beginning of the implied knowledge you don't even think about but the average person would have no clue.

Leaky faucet? Is it the rings, something rusted through, do you know where the water shut-offs are, do you know what wrench to use for what nuts (they aren't labeled like the inside of a PC).

The shutoff is right below the sink or toilet in 99% of houses. You grab it with your hand and turn it. It's very familiar because in most cases it looks exactly like the shutoff valve you've got on the outside of your house where the hose connects. Most faucets are either
1) remove a screw, remove a handle, remove another screw, slide out the cartridge or the shutoff assembly, or
2) remove a screw , remove the handle, unscrew the cap, take out the ball and springs.

Then you just take that up to the hardware store, match up the parts to the identical ones, take it home, and put it back together in the reverse of the above steps. There's usually only a few types for kits for each brand, and most hardware stores have a guy that knows all about plumbing. Most of them will even help you switch the washers in store if you need help. Most repair kits will have instructions right on the back, just in case you get the slightest bit confused. But it's not that complicated at all. The first time I replaced a leaking faucet, I was 16, had never touched plumbing in my life, and I managed to replace it easily, and I didn't even have to look in a repair book (the public internet didn't even exist). But if even a single part of the above is at all confusing, you can buy one of those DIY books that covers dozens of household repairs. Those books will generally tell you how to repair just about any type of faucet, and they'll do in int about 3 pages of pictures with maybe 2 paragraphs of text.

Changing the oil, gotta make sure you know which is the oil pan under the car, the proper nut to remove, get the right oil filter, know how to use an oil filter wrench given most newer cars, know how much oil your car takes and how to fill it... they certainly take just as much knowledge. You just take what you know for granted.

I don't want to go through the whole faucet explanation again, but 95% of that is documented in the owners manual, the guy at the auto parts store will tell you exactly what parts you need (you don't even need to know what make or model your car is...just give him your registration and he can most likely look it up from the vin), and any auto parts store will sell a book that explains the simple process in less than 2 pages (mostly pictures).

I learned to change the oil in 9th grade, almost 20 years ago. Every single car since then has been exactly the same, so it's a process that never changes (except for some high end foreign cars that make it really difficult). Compare that to computers, where the parts/connectors/etc change every few years.

Solid huh?

[John Betonschaar]

In addition, Intel beefed up their graphics core and it appears that the new Intel GMA HD integrated graphics engine offers solid HD video performance

Solid HD video performance? I see 35% CPU load in the Casion Royale 1080p trailer screenshot, on a fast Quad-core CPU. My puny single-core Atom 1.6Ghz with NVidia graphics does 6-10% max on any 1080p content I throw at it in XBMC.

It's better than what Intel offered before: nothing, but I still wouldn't recommend Intel graphics for any HD video player.

Re:Solid huh?

[hedwards]

You shouldn't be recommend Intel graphics for pretty much anything. Unless I missed the memo, Intel is really the worst choice for graphics cards, sure it's available on whatever platform you like, but AMD's been releasing documentation on its cards and the Intel graphics chips haven't been good. You're really far better off going with either nVidia or AMD for graphics chips at this point.

Re:Solid huh?

[MadKeithV]

Solid, as in "frozen solid".

Re:Solid huh?

[MemoryDragon]

Actually what you see os the entire cpu load including the gpu part, what you see in the second case is the pure cpu load and offloaded gpu, the end result is pretty much the same if you sum both up...

Re:Solid huh?

[DoofusOfDeath]

You shouldn't be recommend Intel graphics for pretty much anything.

I disagree. I've had a few laptops that were primarily used for programming. On those, the basic, build-in Intel graphics (GMA950 and X3100, iirc) were just fine.

In fact, they were even better than ATI or nVidia graphics for me: those computers were running Linux, and I could always count on the Intel drivers being available for the most up-to-date Linux kernels, whereas I couldn't make that assumption for the closed-source nVidia or ATI drivers.

Re:Solid huh?

[John Betonschaar]

You're wrong. What you're seeing is that apparently the Casino Royale video is not fully accelerated by the GPU. The GPU load is never factored in the Windows performance monitor, you need CPU-z or something like that.

Re:Solid huh?

[Anonymous Coward]

There was one, and only one worse card that is a distant memory. S3 Virge

Re:Solid huh?

[Anonymous Coward]

You shouldn't be recommend Intel graphics for pretty much anything.

I disagree. I've had a few laptops that were primarily used for programming. On those, the basic, build-in Intel graphics (GMA950 and X3100, iirc) were just fine.

In fact, they were even better than ATI or nVidia graphics for me: those computers were running Linux, and I could always count on the Intel drivers being available for the most up-to-date Linux kernels, whereas I couldn't make that assumption for the closed-source nVidia or ATI drivers.

The current intel laptop that I have doesn't perform as well well as the one I got 4 years ago with an nvidia chip. Memory usage gets to high, and the sucker just crashes.

Re:Solid huh?

[Anonymous Coward]

*CPU* load graphs DO NOT reflect *GPU* load. It's like saying that your car's gas gauge reflects your home heating oil usage. They are unconnected.

What's more likely happening is pure software video decode. Getting GPU video decode working can be fiddly, and I remember the Quantum of Solace trailer being encoded in a fashion (L5.1, 5ref frames) that made it entirely outside the spec, and thus very hardware-decode unfriendly. Nvidia got hardware decode of it working, but AFAIK, no one else even tried.

I don't remember seeing that screengrab in the QoS x264 trailer, though, and it says they were using a rip of the entire Bluray, so it's most likely that the reviewer simply did not take the necessary steps to enable GPU acceleration.

Re:Solid huh?

[Hurricane78]

That’s because you got no hardware acceleration in your graphics card. It has nothing to do with the GPU. And i can tell you that on my system, a 1080p video with 5.1 just sound barely runs at all, eating the whole CPU alive.

So 35% without hardware acceleration is actually pretty good.

Have you looked up if the card and driver actually support acceleration, and if you really have it enabled? VDPAU? Something like that?

By the way: That’s the reason you should never ever use a recent ATi card with Linux. There just is zero hardware video acceleration. And if there is, it looks so bad, that you have to disable it to get rid of huge blocking, and contrast set up so high that there is just either black or white, with, like, 3 levels in-between. ;)

Re:Solid huh?

[Randle_Revar]

No, Intel is great. They have the best drivers right now (AMD OSS drivers are caching up, though), and as long as you don't play a lot of Wine Crysis they are plenty powerful.

Re:Solid huh?

[ratboy666]

But 35% is far better than "can't do it at all". Which is what an Atom with the older gen Intel would get you.

And 35% still gives you plenty of CPU for background tasks (which you probably won't even utilize).

So, you would then have to judge the two solutions based on power performance (especially since you brought in the Atom processor). Do you have any insight into the relative power performance? (to clarify, a unit like "BlueRay/hour").

Reviews online at anandtech.com and techreport.com

[IYagami]

DESKTOP PROCESSORS
http://techreport.com/articles.x/18216/1
"As a CPU technology, Clarkdale is excellent. I can't get over how the Core i5-661 kept nearly matching the Core 2 Quad Q9400 in things like video encoding and rendering with just two cores. We've known for a while how potent the Nehalem microarchitecture can be, but seeing a dual-core processor take on a quad-core from the immediately preceding generation is, as I said, pretty mind-blowing. Clarkdale's power consumption is admirably low at peak
(...)
The integrated graphics processor on Clarkdale has, to some extent, managed to exceed my rather low expectations."

http://anandtech.com/cpuchipsets/showdoc.aspx?i=3704
"For a HTPC there's simply none better than these new Clarkies. The on-package GPU keeps power consumption nice and low, enabling some pretty cool mini-ITX designs that we'll see this year. Then there's the feature holy-grail: Dolby TrueHD and DTS HD-MA bitstreaming over HDMI. If you're serious about building an HTPC in 2010, you'll want one of Intel's new Core i3s or i5s."

NOTEBOOK PROCESSORS
http://anandtech.com/mobile/showdoc.aspx?i=3705
"From the balanced notebook perspective, Arrandale is awesome. Battery life doesn't improve, but performance goes up tremendously. The end result is better performance for hopefully the same power consumption. If you're stuck with an aging laptop it's worth the wait. If you can wait even longer we expect to see a second rev of Arrandale silicon towards the middle of the year with better power characteristics. Let's look at some other mobile markets, though.
(...)
If what you're after is raw, unadulterated performance, there are still faster options.
(...)
We are also missing something to replace the ultra-long battery life offered by the Core 2 Ultra Low Voltage (CULV) parts. "

Do Not Want!

[A12m0v]

Anyone else suspicious of this? Intel trying to use its CPU monopoly to gain a GPU monopoly?

Re:Do Not Want!

[Anonymous Coward]

Too late, Intel already owns most of the integrated GPU market. It is by far the most common vendor of GPU's in the average machine, just not in gamer machines...

Re:Do Not Want!

[kramulous]

We breath oxygen. What do you breath there?

Re:Do Not Want!

[petermgreen]

It will further cement thier already near monopoly in the integrated graphics for intel based systems segment. Whether it will have much impact on the gamer graphics segment depends on how well it performs. It seems that they have more or less caught up with AMD integrated graphics but I don't think that in itself is enough to seriously impact on sales of discrete graphics cards.

Unfortunately TFA jumps straight from integrated graphics to a £130 card and uses completely different settings for the two tests. What i'd really like to see is a comparison of the integrated graphics on these things with say a 8400 GS (a £25 card).

anyone here got an 8400GS and one of the games used in TFA and prepared to run some benchmarks at the settings they used for the integrated graphics test? (yeah I know the rest of the system won't match but all i'm interested in are ballpark figures)

Re:Do Not Want!

[MemoryDragon]

Jepp they already said they want to bankrupt nvidia, every move in the last year was in this direction, first shutting out the ION chipset by illegal pricing now trying to push the gpus into the core so that the cheap enough solution ends wherever nvidia (and ATI but they are less bothered since they can do the same) got its core money from, third fighting a patent war on them to shoot them out of the chipset market.

The entire thing started when NVidia was blabbering about you dont need CPU upgrades anymore just use the GPU for everything, that woke Intel up, and as usual with cheapass solutions which are worse but cheaper they kill off the competition!
Worked in the past works again.
I wonder if we will see NVidia in 5 years at all in the PC market they might end up being a second PowerVR still healthy in the embedded sector but not at all present on the PC side of things.

Re:Do Not Want!

[Anonymous Coward]

Not an official benchmark, but my old Pentium 4 (3.0ghz) with a gig of ram and an AGP GeForce 7600 runs Quake Wars Enemy Territory at around 30-45 fps. Looking up pricing on this card was rendered irrevalent when I realised that it's so old that it's getting more expensive (most AGP cards are though). Judging from gaming performance alone I'd put this card just below a GeForce 6400, which is also so old that it isn't being sold anymore. I'm sure the 8400 would blow this thing straight out of the water.

So congratulations Intel, you have produced a cutting edge graphics chip that gets half the performance of a budget gamer card released about 4 years ago, and gets itself absolutely dominated by your competitors cheapest low end crap.

This is only an article because Intel finally integrated a GPU and CPU on the same chip.

Re:Do Not Want!

[Anonymous Coward]

Jepp they already said they want to bankrupt nvidia, every move in the last year was in this direction, first shutting out the ION chipset by illegal pricing now trying to push the gpus into the core so that the cheap enough solution ends wherever nvidia (and ATI but they are less bothered since they can do the same) got its core money from, third fighting a patent war on them to shoot them out of the chipset market.

The entire thing started when NVidia was blabbering about you dont need CPU upgrades anymore just use the GPU for everything, that woke Intel up, and as usual with cheapass solutions which are worse but cheaper they kill off the competition!
Worked in the past works again.
I wonder if we will see NVidia in 5 years at all in the PC market they might end up being a second PowerVR still healthy in the embedded sector but not at all present on the PC side of things.

Conversely it seems as if you think Nvidia should be allowed to compete in the CPU arena, or destroy it in favor of some bastard child called GPGPU. Why then does this become an issue when when their main competitor does it, and seemingly does it better?

One thing you shouldn't do is brag. Another thing you shouldn't do is bring a knife to a gunfight. Finally you should never, ever, brag about bringing a knife to a gunfight. You must be retarded if you didn't see this coming from the very moment Nvidia said they wanted to compete with Intel.

Re:Do Not Want!

[MemoryDragon]

There is a huge difference between a company which has about 20% marketshare and trying to compete in legal ways by simply providing good products in its own area and another one openly trying to bankrupt it by leveraging illegal tactics and its pure 90% marketshare (ATOM pricing).
Thats exactly the situation we have here. Intel has done similar things in the past, they are a very paranoid company and do not allow competition to thrive. They openly said, they dont see AMD as competition, but they have NVidia on their radar, and their tactics are as illegal as in the past several times.
The Atom pricing scheme to kill of the ION chipset simply is illegal and there is a reason why Intel has the FTC on their necks currently and why they had to pay literally billions to the EU anti trust regulators, they applied the same illegal tactics to kill off AMD to some extent when they were not able to compete on legal means by having decent products!

What we now face here is a situation where Intel simply wants to damage NVidia as much as possible to gobble up the remains to get into the last market in the PC world they cannot compete at all.

After that and once they got the FTC off their back say goodbyte to AMD as well,
and ARM also is on their we want to kill them list.

Re:Do Not Want!

[MemoryDragon]

Jepp they now apply the same tactics they did on the notebook market in an even worse way for the desktop market, simply by providing a SOC solution with a good enough video processor integrated is enough that 80% of all sold desktop systems simply wont get dedicated graphic adapters anymore.

NVidia should get into the x86 business asap by buying via or developing their own solution or they will have to get out of the PC business entirely, and after that it will be AMD.

ARM is not yet on the list of companies openly being targetted, but ARM will get a hard time in 3-4 years.

What the hell...

[NervousNerd]

What the hell is up with their model numbers? Quick, is that i5 you have a dual core or a quad core!? At least Intel's older Core 2 processors differentiated with "Duo" or "Quad", and AMD's simply uses "X2","X3" or "X4".

Re:What the hell...

[Anonymous Coward]

If there is any sanity in Intel, an i5 is obviously a penta-core ... oh, no, wait ...

Re:What the hell...

[Tim C]

I bought an i7 as part of a general upgrade a few months ago; it wasn't until I had it installed and happened to check Task Manager that I realised it was a quad core chip.

Re:What the hell...

[NervousNerd]

Shouldn't your i7 show 8 threads in the task manager as hyper-threading is on (by default, I would think, but I do not own an i7 system)?

Re:What the hell...

[MikeBabcock]

Probably dual-core, with hyper-threading turned on. Try one of the many CPU-ID programs to view your CPU's full features (or cat /proc/cpuinfo on Linux)

Re:What the hell...

[wdebruij]

Yes. Core 2 Quad made so much more sense.

How much does Intel pay its marketeers? Is that in bananas?

Re:What the hell...

[Anonymous Coward]

I have an i7 laptop (so basically an i5) with hyperthreading, and indeed it shows up as 8 threads. However, I assume the original poster did the same thing I did, saw 8 threads, went "WTF," then, "Oh, right, Intel brought back hyperthreading. I have 4 cores x 2 threads per core."

Re:What the hell...

[Anonymous Coward]

Quick, is that i5 you have a dual core or a quad core!?

Quad core. The dual core ones didn't come out until today.

Or anything under Linux?

[MacroRodent]

Is the graphics unit a derivative of the notorious Poulsbo (no good open-source Linux support), or of GMA9xx (open drivers on Linux)?

Re:Or anything under Linux?

[Randle_Revar]

It is evolved from the X4500HD (GMA9xx series)

Re:Or anything under Linux?

[Randle_Revar]

Forgot to say that Intel doesn't really like GMA500 much, as they don't own it and didn't invent it, but rather licensed it from PowerVR, because they needed a very low power graphics chip in a hurry (I guess when they started Atom, they forgot to start a GFX chip at the same time). I expect Intel will kill off the GMA500 soon enough.

Re:Or anything under Linux?

[MacroRodent]

Thanks. That's a relief. Makes the new Intel CPU a reasonable prospect when I next need to update my home box (now with Pentium Duo + ATI 9200). I'm not a big user of 3D apps, but I like to fly with the Google Earth occasionally, and the kid needs some games...

Multitasking

[aclarke]

It's conceivable that one might want to be, say, ripping or transcoding one movie while watching another. Or running a web server while watching a movie. Maybe you want to watch a movie while you're compiling some code, so you want extra CPU for that. There are any number of things one might want to use one's CPU for while watching a movie.

Re:Multitasking

[chrisG23]

Yo dawg, I heard you like watching movies, so we made a computer powerful enough for you to watch a movie in hi-def, while you watch a movie in hi-def.

Re:Multitasking

[sznupi]

65% of Core i5 CPU is worth much more than 90% of Atom for "multitasking". Plus, those numbers aren't correlated strongly with how smooth any hypothetical multitasking will be, it's more about OS & the way apps are written.

Re:upgrade treadmill

[OrangeTide]

don't worry, there is little doubt that this is a downgrade. (except for Atom owners)

Fewer bangs per buck...

[Anonymous Coward]

Yeah, I know. Language is a fluid thing and we can all spell stuff how we want and use any word we choose regardless ( the last refuge of the illiterate ).

Go ahead and be dim.

Re:Fewer bangs per buck...

[Anonymous Coward]

we can all spell stuff how we want and use any word we choose irregardless

FTFY

Re:Fewer bangs per buck...

[PotatoFiend]

FTFY

You only think you did.

Vista Ready ?

[emilper]

well ... is it "Vista Ready"?

Re:Vista Ready ?

[DrMrLordX]

More importantly, will it run Crysis?

(trick question; nothing runs Crysis, at least not smoothly at hi res with all the settings maxed out)

Re:Vista Ready ?

[emilper]

my question was a trick, too: Intel sold quite a few onboard graphic chips as "Vista Ready" in the past; I bought one without doing my homework before, and right now I am quite cautious when it comes to Intel hype.

Re:Vista Ready ?

[Anonymous Coward]

Intel did no such thing, MS lied about the requirements, and the MFGs of system boards and PCs took MS at it's word. MS was the one that started that mess.

http://www.tgdaily.com/business-and-law-features/36279-microsoft-caught-pants-down-in-intel-love-affair

MS determined what was and was not Vista-Ready.

Sure -- theoretically

[anti-NAT]

"In theory, practice and theory are the same. In practice they aren't"

Most people don't multitask on their desktop, or better described, "significantly multitask", meaning run multiple programs are once that are intensively using the CPU(s). Typically, they're running one application, which they're focussed on, and other background applications, while they are running, are mostly idle, or utilising no more than the occasiona few percent.

Ripping a movie, on an Atom CPU PC (likely a netbook) at the same time as watch one? I think that's an unlikely event.

Running a highly trafficed web server, on an Atom CPU? I think that's even less likely that ripping a move while watching one.

Remember the OP's criticism? 35% CPU utilsation, which of course still allows 65% CPU for any other tasks, such as ripping a movie, running a web server etc. was unacceptable. So how much unused CPU is enough for more than likely theortical, rather than in practice, use? 70%, 80%, 90%? Any free CPU is CPU you've paid for but aren't getting any value from. The greater the unutilised CPU percentage, the less value for money you're getting.

People buy CPU capacity based on their peak usage, not their average usage. My fundamental point, and why I agree with "Solid HD" performance, is that the typical high load use of a PC while watching a movie is only watching that movie. If these new Intel CPUs with GPUs still have 65% capacity left while the movie is playing, you could say they're significantly overspec'd for their likely peak use - by 65% or so percent.

Re:Sure -- theoretically

[NervousNerd]

I wouldn't quite say that. Lower CPU utilization equates to less electricity utilized, which equates to a lower power bill. I would much rather have a processor use only 15% of my CPU's resources in order for me to be able to view a movie instead of having to use 85% of the CPU's resources in order to view a movie.

Re:Sure -- theoretically

[jonbryce]

They might be watching a video while touching up their photos in Photoshop. That's probably the most likely heavy use scenario.

Re:Sure -- theoretically

[Anonymous Coward]

your video decoder runs on air ?

Re:Sure -- theoretically

[e70838]

I have two screens attached to my PC. This seems standard nowadays. A usual configuration is TV (IP, DVD, BD, ...) on one screen and work (eclipse java, virtual box running redhat and oracle, ...) on the other. Sometimes in the background, I compress some videos to DIVX.

What kind of multitask would you accept as significant ?

Re:Sure -- theoretically

[Anonymous Coward]

Because everybody knows GPUs don't consume power...

Re:Sure -- theoretically

[aclarke]

I don't dispute that this may be the case for you, but it isn't for everyone. I was just making the case that for some people, including the parent of your earlier post, it is a concern. For instance, my media centre box is a 2.13GHz Core 2 Duo. In addition to playing TV for me, it is also 24x7 recording about 60 data feeds. It could also be running P2P, recording live TV, transcoding video, running one or more VMs if my main server is down for some reason, or any number of other tasks. I would consider the computer at this point to be minimally specced for its uses. It is of practical importance to me that my media server has spare capacity when playing HD content.

In contrast, the computer in my truck is an early Intel Atom. As long as it can drive the foreground process (usually GPS) sufficiently I don't care too much about whether or not it can multitask. I don't think I've ever even tried to play video on it.

Re:Sure -- theoretically

[Bert64]

Heat and noise...
When CPU usage gets high, many machines especially laptops get hotter and crank up the fans to compensate...
If the work is being done by the GPU then typically less heat is generated and thus less noise...
Noise when watching a movie can detract from the enjoyment of the movie.

Re:Sure -- theoretically

[Anonymous Coward]

They might be watching a video while touching up their photos in Photoshop. That's probably the most likely heavy use scenario.

'Touching up' photos in Photoshop is in no way CPU intensive, at least not in the way that most users 'touch up' photos (red eye removal, cropping, straightening, minor color adjustments). At this point, that's like saying that running spell check in Word taxes the computer.

Re:Sure -- theoretically

[Targon]

I think you also missed a key factor, and that is the concept of the GPU being improved. Intel has had the advantage in terms of CPU performance, so Intel is trying to improve their graphics performance now. And, if the new GPU isn't up to the task, THAT is where you CAN focus.

Now, let's be honest, the low end of the market is where integrated graphics really comes into play. So, we are talking about the $400 computer towers you can buy from HP, Gateway, and Dell. And for THOSE, you need the combined CPU+GPU price to be reasonable.

The AMD 790GX with the integrated Radeon 3300 is a few generations back now, but the newer generations for integrated are not really that much faster I don't think. Still, you have to compare the combination of low end CPU with the integrated graphics to see the true value.

Will we see the i5s with the better graphics in those $400 towers, or will they end up being up in the $600 range? For all the improvements, Intel may not be releasing anything that will really change the balance of power in the low end of the market.

As far as power usage, I think many people are a bit too hung up on how much a computer adds to the electric bill. The real line is when you can exchange your space heater for a computer tower due to power usage and heat generation. Or, power draw comes into play when you want to save on battery life in a laptop/netbook. Power also comes into play when you need to vent the heat from a computer. For overall use though, most people REALLY don't care about the power draw, especially when they are buying a faster computer.

So, price for a motherboard+CPU+memory on the AMD side, vs. a COMPARABLE motherboard+CPU+memory on the Intel side. Remember, video performance is a bit more important under Vista and Windows 7 compared to XP if you leave Aero turned on, so the GPU performance DOES impact how good a machine feels to use now. Up to this point, the FEEL of a $400 Intel based tower hasn't felt as good as a $400 AMD based tower overall. That is what you have to look at in the low end.

Re:Sure -- theoretically

[MikeBabcock]

I won't bother pulling up the numbers but I'm pretty sure you'd find that a CPU spec'd to 15% of your current CPU's capacity uses a lot less power than the current CPU running at 15% capacity.

There's a reason they don't throw Core i3's in cell phones and just under-clock them. Low power CPUs exist for a reason.

Re:Sure -- theoretically

[b0bby]

They might be processing .mts files from their HD camcorder - that's the only thing that my ~2 year old machine is killed by. Photoshop isn't that bad, huge compressed video files really are the worst case I come across.

Re:Sure -- theoretically

[PitaBred]

Screw the power bill. We're talking about netbooks.... what about the battery?

Re:Sure -- theoretically

[anti-NAT]

The thing is, Intel don't make CPUs for "some people", they make CPUs for "most people". IOW, the common case. People on slashdot, including the OP I originally replied to, usually aren't the common case. When judged by what I think the market is for this CPU, then I think 35% CPU utilsation while displaying a HD movie is arguably lower than it needs to be. If it was 65% utilsation, it'd still be acceptable.

Re:Sure -- theoretically

[vcompiler]

which of course still allows 65% CPU for any other tasks -- You don't take count into responsiveness. Yes, you still have 65% for other tasks. But that means the point at which your "other tasks"assigned is likely to fall into the duration of 35% and therefore those "other tasks" must wait a while for those 65% comes. If you want your CPU be utilized more than 50%, then you should not use any interactive UI applications, you should treat your machine like an old time batch-processing monster.

Interesting implications

[rpp3po]

While you might have missed that Intel already is the largest GPU vendor in the world for years (gaming is small compared to B2B sales), you are right, anyway. When offering intel CPUs implies having to buy their GPU, the air will become thin for excellent integrated chipset offerings as Nvidia's. Instead of pushing customers through secret, anti-competitive contracts, they have just changed their product lineup. Want a CPU? Fine, but you can't have it without a GPU.

It will be interesting to see, wether Apple will get special treatment. The have already semi-officially let a word slip out, that they are not interested in the Arrandale GPU and won't use it. It's just not powerful enough for their GPU-laden OS and application lineup compared to Nvidia's chipset offerings.

Re:Interesting implications

[TheRaven64]

And remember the last time they did this? Before the 486, there were a few x87 manufactures (including AMD). The 486 came with an integrated 487, so there was no need to buy one from a third party (they later split the line into 486sx and dx, where the sx was a 486 with the broken 487 disabled).

AMD survived by ramping up investment in their x86 clones and shifted to selling x86+x87 cores, rather than just x87 cores, as their primary market. ATi is doing the same thing by being purchased by AMD. nVidia is trying to do something similar with ARM cores in the Tegra line.

Re:Interesting implications

[vcompiler]

There was one time people bought resistances, capacitances, and transistors to build up their own circuit.

So here's the summary:

[Anonymous Coward]

The new Core i5 Clarkdale-based CPUs might be interesting, but they're overpriced. These are dual core CPUs with Intel integrated graphics built into the CPU package, and they cost roughly as much as a quad core Lynfield (Core i5-750) or AMD Phenom II X4 965 CPU, both of which will trounce it in any benchmark. They are somewhat based on the Nehalem architecture, but they moved the memory controller off of the CPU core (while leaving it on the CPU package) introducing more latency and lower memory bandwidth.

The Core i3 CPUs offer more of a value proposition with prices starting under $130. These might be the chips to go for if you want an HTPC, though the CPU utilization for HD media decodes is much higher than on similar platforms (i.e., nVidia ION).

The integrated graphics performance is nothing to get excited about and is really only suitable for business use/HTPC use. You're still not going to game on this GPU, nor will it be suitable for high performance computing.

The single most interesting thing about these CPUs are the inclusion of the AES-NI instructions which accelerate AES encrypt/decrypt functions. When paired with full disk encryption solutions that utilize AES these CPUs see a roughly 15% decrease in disk performance as opposed to the usual 30% or so. Of course, you might just as well buy a quad core CPU and let the extra cores handle encode/decode too.

Realistically these are going to be used in business-class PCs. You get decent dual core performance, competent business graphics performance, and integrated support for accelerated AES functions. They might also be suitable for home brew VPN endpoint solutions with their AES acceleration and relatively low power requirements as well.

Oh yeah...while these will work in existing Socket LGA 1156 boards (with a BIOS update, of course) you will need a completely new motherboard if you want to take advantage of the integrated graphics capability, as existing boards do not have connectivity from the CPU socket to a video out port. Of course, if you have an LGA 1156 mainboard already then any of these new Clarkdale CPUs would be a downgrade, so probably no worries there.

Re:Unrealistic expectations

[emj]

Even if you only use 80% of your CPU I'm pretty sure it will choke at some point, because there is always something running.

Not that different

[Anonymous Coward]

Intel also has three lines that more or less directly correspond to AMDs: Core/Phenom (good), Pentium/Athlon (ok) and Celeron/Sempron (cheap), plus the server Xeon/Opteron. The real pain is the amount of different model numbers and numbering schemes. The secret decoder ring for Intel models is:

A) old three number codes
E.g. Pentium 965, Celeron 450, ...
First digit is the model, second digit corresponds to the speed
These are usually old crap and should be avoided. Celeron 743 and Celeron 900 fairly recent low-end chips that you can still buy.

B) Letter plus four numbers codes, e.g. SU7300:
* S = small form factor
* U = ultra-low voltage (5-10W), L = low-voltage (17W), P = medium voltage (25W), T = desktop replacement (35W), E = Desktop (65W), Q = quad-core (65-130W), X = extreme edition
* 7 = model line, tells you about amount of cache, VT capability etc. Scale goes from 1 (crap) to 9 (can't afford).
* 3 = clock frequency, relative performance within the line. Scale from 0 to 9.
* 00 = random features disabled or enabled, have to look up for specific details.

C) New Core i3-XYZa
Similar to scheme B, with the added dash and more confusing
* i3 = Line within Core brand, can be i3 (cheap, but better than Celeron or Pentium), i5 (decent) or i7 (high-end)
* X = the actual model, tells you the amount of cache and number of cores, but only together with the processor line (i3-5xx is very different from i5-5xx)
* Y = corresponds to clock speed, higher is better
* Z = modifier, currently 0, 1 or 5 for specific features
* a = type of processor: X = extreme, M = mobile, QM = quad-core mobile, LM = low-voltage mobile, UM = ultra-low-voltage mobile

Re:Not that different

[Rudeboy777]

This secret decoder ring is exactly where Intel has got it all wrong. Who needs to charge them with antitrust violations when the marketing and product departments will run them into the ground anyway

Re:Not that different

[kestasjk]

My laptop sticker says "Centrino 2", and I just happen to know that that's a Penryn (and what "Penryn" means). I think it's safe to say Intel's naming scheme sucks.

Re:Not that different

[Anonymous Coward]

Somebody put this in an faq on gamefaqs.

Deciphering weapons in Borderlands is easier than this.

Ridiculous...

Re:Not that different

[hairyfeet]

No shit! And which of those by looking at his list have VT (required for XP Mode) and which don't?

Now let us compare that big mess to AMD, shall we? Phenom = best, with the higher number being the fast chips. As you can see here it is pretty simple, with the faster chips having a higher number, such as my 925 being faster than an 805 but slower than a 940.

Then we have Athlon Xx = good. With the Athlon the second x is the number of cores, be it X2, X3, X4. Again as we see here the bigger the number the faster the chip. And notice they have 45w quads, which makes for a nice machine that sips power and can still kick butt when needed. I have found the Propus makes a damned good office machine.

Then finally we have the Sempron = Cheap. They are only offering a single Sempron with the current so there is no need to compare, and frankly you'd have to be insane to buy a Sempron when you can get an Athlon dual for $51. And most telling, did you notice that ALL of their current chips have AMD-V? No need to guess, if it is a current chip, from the cheapest Athlon dual to the highest Phenom X4 they all have AMD-V. You would think with a major feature of Windows requiring it that Intel would be the same, but last I looked there was plenty of their low midrange chips without VT enabled.

Intel may have the highest benchmarks but trying to keep up with their naming schema is a PITA. Especially with some having VT and some not, whereas with any of the AMD chips I don't have to worry about a customer deciding they need Windows 7 Pro for XP Mode, because all the chips support it. Much easier IMHO to deal with than the new Intel schema.

Article is terrible

[sammydee]

The article is awful. There is only one game benchmark and that compared to an integrated AMD GPU that hardly anybody has heard of. There is also no way of telling from the article whether the integrated intel graphics actually has HD video decode acceleration or not. The modern core i5 chips are pretty capable of decoding 1080p content by themselves without any gpu assistance.

I think the article writer misunderstands how hardware video decode assist actually works. It isn't magically engaged when you play any HD movie in any media player (usually it has to be turned on in an option somewhere with a media player app that supports it) and it isn't a sliding scale of cpu usage. Modern decoding chips either decode EVERYTHING on the card, reducing cpu usage to 1% or 2%, or the app decodes EVERYTHING in software, resulting in fairly high cpu usage.

I still have no idea if the new intel graphics chip actually offers any HD video acceleration at all. If it did, it would make it a nice choice for low power and HTPC solutions. If it doesn't, it's just another crappy integrated graphics card.

Re:Article is terrible

[nxtw]

I think the article writer misunderstands how hardware video decode assist actually works. It isn't magically engaged when you play any HD movie in any media player (usually it has to be turned on in an option somewhere with a media player app that supports it)

DXVA acceleration works automatically with Windows 7 and any application using the proper built-in decoder and EVR renderer. It should also work with Media Player Classic Home Cinema, if the default renderer is compatible.

I still have no idea if the new intel graphics chip actually offers any HD video acceleration at all.

It does. The G45/GM45 chipsets released in 2008 also have full decoding.

solid HD performance?

[StripedCow]

with 20.7 frames per second?
that's not what i call solid performance...

tough day for nvidia stock

[tjstork]

a wsj analyst has to be looking at this, and concluding that the gpu business is doomed.

Re:tough day for nvidia stock

[TheRaven64]

It is, and has been for a long time. It's now in the same place that the discrete math coprocessor market was in 1989. That's not necessarily a problem for nVidia for two reasons.

First, Intel is licensing the Atom microarchitecture to SoC manufacturers. They can also license a GPU core design from nVidia, and maybe a DSP design from someone else, and build their own integrated SoC with an nVidia GPU and an Atom CPU. This is Intel's attempt to compete with ARM. When you buy an ARM chip, you almost always buy a SoC that contains a few other cores to suit your particular application. The diversity in the ARM marketplace is something Intel has difficulty competing with.

Secondly, nVidia doesn't just make GPUs for x86. They also make their own line of ARM SoCs; Tegra. These take a fairly stock ARM core and combine it with a low power nVidia GPU core. I'm not sure whether they will keep developing Tegra, or whether they will just license the IP to other SoC manufacturers. The margins are higher with the first option, but the quantities are bigger with the second.

If I were in charge of nVidia, I would look seriously at following ARM's business model and transitioning to selling IP cores to SoC makers. Even Intel buys some of these - some of their chipsets incorporate a rebranded PowerVR GPU - and they could probably sell some cores to AMD and Intel, as well as the likes of Samsung, TI, and so on who produce ARM SoCs.

Re:tough day for nvidia stock

[Fujisawa Sensei]

a wsj analyst has to be looking at this, and concluding that the gpu business is doomed.

Intel is going to have to come out with a GPU that's better than a 4 year old nvidia gpu first.

Re:tough day for nvidia stock

[MikeBabcock]

Intel claimed the video acceleration market was doomed years ago with their first AGP cards that accessed system memory and were about as useful for gaming as a Gameboy that's been through the laundry.

Integrated video has been available for a long time, and it keeps getting better, but the gap between that and truly good gaming hardware is still very wide. The difference is that video card makers simply need to refocus and won't have the low-end market to supply anymore for those easy profits.

Re:tough day for nvidia stock

[TheRaven64]

Did you used to work for SGI? Your post sounds exactly like the arguments that SGI management was making in the '90s.

Discrete graphics chips are luxury items. For most users, integrated graphics are more than enough; there's a reason why Intel has more than 50% of the GPU market. I rarely come close to stretching my three-year-old laptop's GPU. Integrated graphics keep improving. The gap between integrated and discrete isn't the important thing, it's where on the line 'good enough' sits. For a lot of users, 'good enough' is well within the integrated range. For a few hardcore gamers it isn't, but integrated graphics (especially ones integrated onto the CPU die, which get much better CPU to GPU bandwidth) can play most games, especially ones a couple of years behind the curve, which includes new games that are released with licensed engines.

The market for stand-alone graphics hardware is going to keep shrinking, just as the market for graphics workstations did. You'd think nVidia would learn this lesson well, because they were the company that killed SGI by making hardware that was not quite as good, but good enough, and an order of magnitude cheaper. Before nVidia you basically had the choice of software rendering (including things like 3Dfx cards, which only accelerated as small part of the pipeline) or a really expensive workstation. Then nVidia came along and produced 3D accelerators that were a tiny fraction of the price of a workstation and ran in a commodity PC. They weren't as good as SGI workstations, but you could buy ten PCs with nVidia cards for the price of one SGI machine.

Now nVidia is seeing the same thing, but with one extra constraint; power. GPUs integrated onto CPUs can share a lot of hardware with the CPU (memory controller and so on). This gives them a power advantage in the laptop and handheld markets (the bits of the computer market that are growing quickly).

The low end is eventually the only bit of any high tech enterprise. You have three choices. Control the low end and watch it consume the rest of the market, let one or more of your competitors do that and go out of business, or create a new market. nVidia's best choice is option 3, which is why they should be strengthening their ties with ARM.

Re:tough day for nvidia stock

[LUH 3418]

>> Intel is going to have to come out with a GPU that's better than a 4 year old nvidia gpu first.

Not to flame or anything, but the GeForce 6800 GT was released in 2004. I owned one for some time, and it could comfortably play Half-Life 2/Left 4 Dead at 1920x1200 with 2xAA. I also owned a laptop with a GeForce 7600 Go (2006 videocard) which did O.K. in Prey, a rather demanding video game at the time. Can any Intel GPU even manage to play those games comfortably at 1280x800? My girlfriend owns a recent Sony Vaio laptop with a GMA chip, and it's not really good enough to play WoW.

Re:Unrealistic expectations

[John Betonschaar]

Well for one, the machine can be passively cooled and will jump over 70 degrees Celsius if I tax the CPU for more then a few percent, during GPU accelerated playback it stays nicely around 60C. Also, the thing is in use as a home-server/personal web server, which means there's all kinds of stuff running in the background. 35% of a core i5 = around 300% of a single core Atom, you do the math.

Last but not least I like the idea that the most efficient part of my computer is used for the most appriopiate task. The Atom is barely able to do full-screen standard-def flash video, while the Nvidia GPU does silk-smooth 1080p content. How on eart would someone _not_ want that.

Just the mere mention

[Mattskimo]

of integrated graphics makes me shudder. Didn't we get over that in about 1999? Seriously though this looks like a fairly terrible solution unless you feel like running Vista on something the size of an iPod.

Re:Just the mere mention

[kramulous]

While I generally agree, I would like integrated graphics in their high end chips. That north-bridge gives be the shits when trying to do some numerical computing. It takes too long to move data from system memory to gpu memory. An integrated chip will alleviate this.

OK can someone clear this up

[ZERO1ZERO]

I have no idea of what intel are calling their cips and which is the best etc

Can someone answer these 'simple' questions - In terms of regular geek activities, movie playing/encoding, gaming, compiling, rendering, desktop use, all the regular things

1. Which processor is the all out fastest, best (money no object)

2. Which processor is the best bang for buck (money and object)

3. how do intel chips compare to amd on the bang per buck level.

Re:OK can someone clear this up

[Jesterace]

I found it's hard to find an intel chip that's best bang for your buck. I ended up going from a Athlon x2 6000 to a intel e5200 at the time was the closest in price. Though I wanted to build a hackintosh which didn't last too long. Next step is going back to AMD as I should have stuck with in the first place.

Re:OK can someone clear this up

[W2k]

Of course. Every PC hardware site worth a penny does regular articles on which CPU is currently the fastest and which will give you the most for your money. As well as comparisons between Intel/AMD. My favorite site for such things is Tom's Hardware, though Google will likely find you many more.

Which CPU is actually fastest heavily depends on what you will be using it for. Your list of "regular geek activities" does not narrow it down enough. Also, many applications contain optimizations that target a particular CPU family or architecture.

CPU articles: http://www.tomshardware.co.uk/review/Components,1/CPU,1/

Best (gaming) CPU for the money as of dec 09: http://www.tomshardware.co.uk/best-gaming-cpu,review-31755.html

All CPU performance charts: http://www.tomshardware.co.uk/charts/processors,6.html

Re:OK can someone clear this up

[sydneyfong]

1. The latest Intel Core i7s
2. I bought a AMD Athlon x4 620 recently. Touted as the "$99 Quad Core Processor". If you can find a way to use up all the four cores it's basically the cheapest per clock.
3. Intel chips are generally more expensive. But Intel's high end stuff are hands down faster than AMD. But then considerations would include availability of supporting motherboards and the price of those too.

Re:OK can someone clear this up

[petermgreen]

Which CPU is actually fastest heavily depends on what you will be using it for. Your list of "regular geek activities" does not narrow it down enough. Also, many applications contain optimizations that target a particular CPU family or architecture. .............
All CPU performance charts: http://www.tomshardware.co.uk/charts/processors,6.html [tomshardware.co.uk]
I just picked a few of those at random and they were all topped by the "Intel Core i7-965 Extreme Edition (Bloomfield 4c) * 3.2 GHz, DDR3-1066, 1 MB L2, 8 MB L3"

Seems if raw performance is what you desire the i7 9xx stuff (with triple channel ddr3) is basically the top of the pile regardless of application category.

Re:upgrade treadmill

[PopeRatzo]

don't worry, there is little doubt that this is a downgrade. (except for Atom owners)

Care to elaborate on how an i5 on a laptop is a downgrade for anyone?

Re:Reviews online at anandtech.com and techreport.

[Kjella]

Wow, you're cherry-picking in favor of Intel, how about some quotes like:

When I first started testing Clarkdale I actually had to call Intel and ask them to explain why this wasn't a worthless product. The Core i5 661 is priced entirely too high for what it is, and it's not even the most expensive Clarkdale Intel is selling!

Repeat after me: A monopoly isn't illegal

[Joce640k]

Monopolies are only illegal when you abuse them.

Re:Repeat after me: A monopoly isn't illegal

[Anonymous Coward]

you show me any monopoly, and we'll all show you how it's abusive

99.9% of the time it's as obvious as higher prices and poor service

Re:upgrade treadmill

[Anonymous Coward]

You do know that you can buy Core i7 laptops for relatively reasonable prices, right?

only have 1 x16 + DMI IS bad as boards with usb 3.

[Joe The Dragon]

only having 1 x16 + DMI IS bad as boards with usb 3.0 / sata 600 have to cut pci-e lanes or use pcie switches to have the bandwidth to run them.

Excellent point!

[Anonymous Coward]

The only thing companies care about is for you to give them money. Confusing you into doing so is obviously a benefit to them. Remember the Megahurtz war?

to bad intel wants to make QPI to be in $500+ cpus

[Joe The Dragon]

to bad intel wants to make QPI to be in $500+ cpus only and maybe exons as well.

The desktop i5 / i7 should have QPI as well.

Apple better not use this gpu as it is slower 9400

[Joe The Dragon]

Apple better not use this gpu as it is slower then the 9400m and much slower then the newer 9400m gen 2.

apple has to much in to the gpu / cuda to go back to Intel GMA POS.

and if they do intel is just asking for people to user mac os x86. Come on a $1200 aio with this? $1500 - $1700 laptops with this and 13" / 15" screens? a $800 desktop with this. When you get a core i7 (920) 5770 ati video 6gb ram 1TB HD and more for $1000 - $1200. Apple better not even think of this at $800+.

Nvidia does a real thing, Intel is fake

[Ilgaz]

...or, being a journalist, he uses a Mac and he has purchased the i950/945 based scandals from Apple.

Trust me on that, Apple figured they made the biggest mistake by trusting to Intel's "graphics". There are games who has to carry "Intel graphics based Macs aren't supported".

Imagine, you fix the endian issue, claim to have "best opengl" and you base your OS to GPU acceleration features. Some CPU monopoly who you stupidly relied on as a single vendor offers you a graphics solution and your "living room computer" (Mini) can't even display Cover flow on iTunes.

If I stay on Apple brand and figure out there is no way a Quad Core Intel CPU is planned in it, I may switch to Mac Mini (Nvidia 9400M) from a Quad G5. I was that impressed with the performance, at least the DX/OpenGL feature support from the GPU. On the other hand, my cousin ended up with almost no games on "white" Macbook. They told him "Black" one needed to play games. Guess what the White has? Intel GPU. Black, Nvidia.

As early as today, I can watch 1080p youtube videos with 4-6 % CPU load with a very little load to GPU too. If I want to play a game, everything is supported, there is no "supported games" list. This GPU (!) decision by Intel will do nothing than further drive people to AMD/ATI. Add the insane pricing of i5/i7 too, in this economy, where business guys are happily using netbooks at cafes. Ignore the hype, these are very serious mistakes of INTEL. By GPU/CPU integration, they may trigger a MS like situation.

Re:Nvidia does a real thing, Intel is fake

[petermgreen]

On the other hand, my cousin ended up with almost no games on "white" Macbook. They told him "Black" one needed to play games. Guess what the White has? Intel GPU. Black, Nvidia.
Either you are misremembering the story or the person who advised your cousin was wrong. The black macbook always had intel graphics as did thier contempory white models.

Maybe you were thinking of the brief period when the aluminium unibody macbook had nvidia graphics while the white macbook still had intel graphics.

i5/i7 am I missing something?

[Ilgaz]

I finally understand, i5 is lighter version of i7 with less cache. OK, it is what Intel did for years with Pentium/Celeron.

What is the basis of not enabling "HT" on lower end while it is on higher end which is already in use by high end Workstations and apps actually using the cores and doesn't need some fake virtual CPU to fill threads?

Re:i5/i7 am I missing something?

[nobodylocalhost]

It goes deeper than that. i5 is limited to dual channeling. This means less through put between your ram and cpu. So if you are doing something memory intensive, you'll feel the difference. It also affects disk caching.

As far as HT goes, this is my thought on it. HT is great for those who use assembly or in-line assembly as it enables you to keep more data on registers and cache as oppose of keeping them in RAM or Disk (better logical processors have separate registers but shares the same adder array, functional execution area, as well as cache; cheaper logical processors utilize a fetcher trick that sends data to two paths of the same physical core). If optimized correctly, a program can be much much faster than running it on a single physical thread. However there are timing issues due to sharing of certain resources. This is especially noticeable for cache refreshing. High-end and Low-end CPU really doesn't matter as much as how you utilize your computer and how much load your processor can physically handle. If all you do is e-mail, browse, and single threaded gaming; it makes no sense to have HT enabled since you are clobbering cache left and right. If you are running a database server or encoding video on multiple threads, HT is fairly good at keeping programs running fast. Using HT correctly doesn't mean using all threads in totally unrelated manner; more hardware threads are better for that.

Let's get to why people dislike HT on low end CPUs. This is because people try to overclock these CPUs, and the timing mechanism in HT was affected negatively on certain processors. In short, it made overclocking unstable and prone to errors.

Mainboards/chipsets for i5/i7 are expensive too

[Ilgaz]

I tried to make 2 alternatives for my imaginary new PC (as I am getting sick of Apple), one AMD Athlon 2 Quad based, other i5 or i7 based.

If you go with a trusted brand like Asus, the mainboard may cost more than the CPU itself! AMD mainboards are way cheaper and has an integrated but a REAL gpu, ATI 4000 something which really supports up to directx 10.1 and has several 2d acceleration features.

I couldn't see usual suspects offering i5/i7 supporting chipsets, VIA etc... Or they are a bit late...

Re:Mainboards/chipsets for i5/i7 are expensive too

[hairyfeet]

And you can actually game on the ATI IGP! I know, its a shock after dealing with Intel crapola GPUs, but I played games like FEAR and Bioshock for nearly 3 months on my 780v (which is a 3xxx series IIRC) before getting around to picking up a 4650HD. The gameplay was smooth, and what was extra nice is all the major codecs were hardware accelerated out of the box.

For the CPU I picked up the 925 and frankly it is scary how well that baby runs for just $140. Even after gaming for hours it rarely gets above 114f on stock cooler, averaging 108f under load and 82f idle. With 8Mb of total cache this things really flys and at 95w it isn't sucking down the juice either.

Of course if you want to go really cheap they have the Athlon X4 for $99! Now the only difference I can tell is the Athlon has the L3 cache disabled (which i have heard if you get the right Gigabyte board you can often re-enable cache and/or cores from BIOS) but you are still talking a 2.6Ghz quad for under $100. I have built a few of these for customers and they just rave how fast it is.

But if you are thinking of building a new PC the bang for the buck with AMD now is just insane. All told IIRC after rebates I paid less than $700 for my quad core with 8Gb of RAM, a 4650 1Gb, dual 500Gb HDDs, and Windows 7 HP X64. I am used to having to build myself a new machine every three years or so, but honestly this thing is so fast I just don't see it happening for quite awhile. Hell the board will take up to 32Gb of RAM! You really can't beat the bang for the buck at AMD and the last thing we need is Intel to end up with a monopoly again like they had in the old days. And with the socket AM3 being AM2+ backward compatible you can keep a board for quite a long time with only a BIOS update to support the new chips. As a former Intel guy I highly recommend the new AMD chips. They are some truly scary tech for truly cheap money.

Re:upgrade treadmill

[nxtw]

You can't get a small/lightweight Core i7 laptop, though. I doubt any of them have spectacular battery life.

Re:Reviews online at anandtech.com and techreport.

[Vigile]

Couple of other reviews worth noting:

Clarkdale (desktop): http://www.pcper.com/article.php?aid=849

Arrandale (mobile): http://www.pcper.com/article.php?aid=850

turbo

[Anonymous Coward]

turbo mode alone is a huge marketing scam. along with itty bitty L2 cache. and terrible GPU performance.

Perhaps ironically

[N Monkey]

I wonder if we will see NVidia in 5 years at all in the PC market they might end up being a second PowerVR still healthy in the embedded sector but not at all present on the PC side of things.

Ironically, if the folks on Beyond3D are right PowerVR are still in the PC market in the Intel Lincroft chipset with dedicated 3D, video decode and video encode IP.

Re:Perhaps ironically

[MemoryDragon]

Yes Indeed, Intel has licensed the PowerVR design again... but they did a lousy job, the chipset they integrated has been way more capable than the drivers Intel has delivered are capable. I have to revert my statement somewhat because I forgot PowerVR again is present somewhat.

Re:upgrade treadmill

[Anonymous Coward]

You can't get any mainstream x86 processor with spectacular battery life. Even the Atom pales in comparison to ARM.

Re:upgrade treadmill

[nxtw]

You can't get any mainstream x86 processor with spectacular battery life. Even the Atom pales in comparison to ARM.

The i7 has bad battery life relative to i5/i3 and mobile Core 2 Duo CPUs, which have bad battery life relative to Core 2 LV/ULV CPUs.
And ARM systems do not have the performance of mainstream x86 processors and chipsets.

Re:to bad intel wants to make QPI to be in $500+ c

[Neil Hodges]

The Exon (Xeon) Valdez?

Don't bet against integration

[Anonymous Coward]

Didn't we get over that in about 1999?

No. The semiconductor business never gets over integration. Integration is the reason you can afford to own or use computers.

Discrete graphics cards represent a small and shrinking fraction of the personal computing world. They have high margins for manufactures so they are still being supplied. These margins make them targets for Intel, however.

Probably more than 90% of the laptops in the world use integrated graphics, and laptops represent more than half of all personal computers sold. Integrated graphics will ultimately win because it's cheaper, more reliable and uses less energy. That's why disk controllers, sound devices, network devices and even wifi radios have been integrated on to motherboards. All of that stuff use to be discrete. Graphics is being integrated directly into the CPU package due to the high bandwidth needed.

Don't ever bet against integration; you'll lose every time you try.