Slashdot Mirror
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."
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.
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.
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. "
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
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.
The average consumer doesn't give a shit what socket their CPU is in either, so it's all okay.
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".
"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.
The Internet's nature is peer to peer - 20050301_cs_profs.pdf
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.
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
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.
This is how the loudness war is killing music.
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.
Monopolies are only illegal when you abuse them.
No sig today...