he just said tablet and netbook, and you go right back to phones. do you really have any understand about the market of ontario?Quote:
Originally Posted by savantu
Printable View
he just said tablet and netbook, and you go right back to phones. do you really have any understand about the market of ontario?Quote:
Originally Posted by savantu
Yes, and it seems like Ontario is a business opportunity for AMD. Filling a market segment effectively is a good way to generate revenue, despite low margins.Quote:
Originally Posted by savantu
Power is exactly the reason. They would have much lower battery life with Atom.Quote:
If Atom is weak, what about current ARM designs like Ipad and Iphone ? Those CPUs are much slower than Atom. Maybe the whole point is not performance, but power ? Atom has still too much performance per clock to be able to go down to a few hundred milliwats of power. The uarch needs to be simplified or moved aggressively to new processes. Performance is the last issue.
Why reinvent anything? Maybe because it would be better. ARM is a more effective low-power architecture. A low-power instruction set from Intel plus their process would be more efficient then just their process alone. The MID/smartphone segment software is still an evolving platform at this time so backwards compatibility is not as important.Quote:
Why bother and reinvent the wheel when you already have x86 ? The whole problem is getting it into the right envelope. And from then on, for Intel at least, its process advantage will leave ARM&foundries in the dust.
The atom IS weak, it can't even decode HD properly. Nor can it handle flash. SoCs for mobile platforms work because they have dedicated silicon for handling tasks or are very well optimized. The atom is just a pathetic all rounder. What's the point of trying to stuff x86 into that segment?Quote:
Originally Posted by savantu
None, except if you're intel and want to make money. But it's not working very well right now.
Are you saying this from experience or just pulling it out of your south pole ?Quote:
Originally Posted by cegras
Besides, what stops Intel to implement dedicated silicon for those tasks ? They are already doing it.
and where is your south pole to prove it does..... you already acknowledge it can't by stating that they are doing it right now in new silicon....Quote:
Originally Posted by savantu
Nothing and they already have it in the pipeline:Quote:
Originally Posted by savantu
http://www.anandtech.com/show/3929/i...system-on-chip
Granted its for embeded solutions, you see where they are heading.
http://ark.intel.com/ProductCollecti...codeName=32201Quote:
Originally Posted by savantu
Think this is the link that you were quoting it from. Most of them are single core atoms and well, as far as i'm concerned, priced slightly higher as there was no competition as yet.
Then again TDP is indicative of a thermal design envelope where the rating in W is the maximum temp that the given chip could operate in. AMD quotes TDP as the maximum as allowed by the engineering design. Intel quotes a recommended design point, which is not indicative of a potential maximum draw by that chip.
see the linked articles for reference...
http://www.silentpcreview.com/article169-page3.html
http://www.silentpcreview.com/article169-page4.html
&
http://arstechnica.com/hardware/news...esign-wins.ars
Also, if you look at their i series chippery, let us take the most popular, i7 920, the Acore is 100(max) and Vcore is 1.37(max)... the formula to calculate TDP is rather simple, multiply both... it comes to 137W, whereas Intel quotes only 130W :P
Also, there's a lovely quote about how Intel measures TDP in the arstechnica article. Now AMD, when it says Bobcat's TDP is 9W, you know maximum drawn will be 9W, which is the worst case scenario and almost never happens. Atoms however, expect them to be running at that maximum allowed in that envelope. Then again, that is not the maximum that an atom chip could draw... I couldn't find Vcore and Acore info on Intel site as well, what could they be hiding?
TDP =/= max power drawnQuote:
Originally Posted by tifosi
See PII 965BE 125W version, P9950BE, x2 6400+
All of them have higher power consumption versus TDP.
Source?Quote:
Originally Posted by qcmadness
Which is why AMD came out with ACP... now let us not get into that argument. However, i think i read somewhere that this is how the thermal design envelope is calculated based on which cooling solutions are designed. It is quoted as it is by AMD, but not by Intel... :PQuote:
Originally Posted by qcmadness
Also...
http://www.lostcircuits.com/mambo//i...1&limitstart=6
http://www.firingsquad.com/hardware/...e_performance/
in the above links you could see that at load 2 of the 3 chips you quoted draw less power than they quote as TDP... nm, leave this be! :confused:
http://www.xbitlabs.com/articles/cpu...3_3.html#sect0Quote:
Originally Posted by -Boris-
It is not the real power consumption, it is way more:Quote:
Originally Posted by qcmadness
Quote:
In this case we measured the power consumption along the 12 V power line connected to the CPU voltage regulator, which means that the numbers above include the power losses that occur in the mainboard voltage regulator too. Nevertheless, they are quite sufficient for comparative purposes.
You can't compare that to ampere numbers.Quote:
Originally Posted by qcmadness
In case you haven't noticed, there are heatsinks and heatpipes on the power regulators. Which means that they consume a considerable amount of power.
So that measurement isn't reliable, besides, some of the power in that cable can be used for other stuff, it depends on motherboard design. Compare i7 with Phenom II. i7 often uses less power in load than a Phenom II when measuerd at the 12V, but uses a whole lot more when measured before the PSU. And that is in CPU intensive tasks with otherwise equal hardware. Intel might prefer to use more power from the ATX-connector, and AMD might prefer to use some power from 12v for memory and NB. You can't know that if you isn't involved in the design of those motherboards.Quote:
In this case we measured the power consumption along the 12 V power line connected to the CPU voltage regulator, which means that the numbers above include the power losses that occur in the mainboard voltage regulator too
The only things we do know in the matter is that motherboard layout matters, power regulation uses a whole lot of power, and there are some big differences between AMD and Intel on 12v power draw vs PSU draw.
So, with that taken into consideration, It's safe to say that you can't use those numbers to determine actual power usage for the CPU.
article says 1.4v and 1.45v for the C2, i thought that was an issue as all retail chips were 1.35v MAX after proper bios was released or something like that. the stock cooling on amd chips is horrible and they have a low maximum temp, and generally have run way under their tdp.Quote:
Originally Posted by qcmadness
I'm pretty sure the 4 or 8pin 12v motherboard connectors are just for the CPU...unless I've been taught incorrectly. The issue here is that the power they're measuring is the power draw BEFORE the VRMs for the CPU. So whatever power is wasted as heat, whatever the VRM effiency is, that is also being added to CPU power in their measurement. SPCR attempts to estimate VRM losses, but it's dynamic, since the efficiencies of the power regulation, like all power circuits, reduce with high temperature.Quote:
Originally Posted by -Boris-
http://www.anandtech.com/show/2889/4Quote:
Originally Posted by savantu
http://www.anandtech.com/show/2889/10
And overall that article in general. What a pathetic, useless chip. Mind you, that's a review of the desktop solution, not a netbook platform.
http://techreport.com/articles.x/18167/7
http://www.digitimes.com/news/a20100401PD202.htmlQuote:
Netbooks have always been an obvious example of compromise, and the 1005PE is no different. You give up a lot of performance and functionality in exchange for awesome battery life, budget pricing, and a tiny footprint. Intel would probably be quite content to let that compromise remain intact, since netbooks have already eaten into sales of notebooks that carry higher average selling prices.
Looks like people are starting to realize it, after getting burned on original netbooks.Quote:
Hewlett-Packard (HP) and Dell have both significantly reduced their investments in the 10-inch netbook segment, with HP reportedly even considering quitting the 10-inch netbook market and turning its focus to AMD-based 11.6-inch notebooks because profits from Intel Pine Trail-based netbooks have been lower than expected, according to sources from notebook makers
I guess that precludes anyone with a career in science ever using a netbook.Quote:
Have you ever tried to open up a couple of pdf's, outlook and maybe word 2007 on an atom netbook? Not very pretty..
Although the N550 dual core mobile Atom is coming out, it's still probably worthless unless paired with that Broadcom chip (e.g. reference Anandtech's article about the Dxx for dual core performance). I'm open to seeing if it will cut it for a tablet style solution (and therefore netbooks) but fact is people have been screwed for the past two years with nothing but fluff.Quote:
I changed to foxit reader, the footprint is smaller however, there really isn't any good substitute for office 2007. I also use chrome and upgraded to 2 gb of ram. Even so called light weight usage can be taxing. Try opening an excel worksheet with a few graphs.. ouch. The single core atom is anemic, its nothing but an expensive toy that will eventually fill up our land fills with more computer garbage. Its not even fast enough to do today's apps. It like using a comp from early 2002 in 2010. I bought into the whole low power, portability thing and bought one of these but I do too much useful work on a computer in a given day to use this. Its such a niche product (in terms performance / watt) but it sells so well because its cheap. You may have bought one of these as a toilet seat computer but many people will get attracted to the price and buy themselves back to the late pentium 3 era in performance.
http://www.engadget.com/2010/09/07/h...rmance-review/
Not bad, getting better, but expensive.
Nothing, except power consumption. Any additional dedicated circuitry draws additional power. A generalist arch chip + dedicated hardware will still be less efficient the a low-power arch chip + dedicated hardware.Quote:
Originally Posted by savantu
And from Anand:
Mmmmmyep. Looks like I'll either be waiting a while for a good tablet or purchasing an ARM + Broadcom combo. At least ARM based ones will have optimized software like Android, whereas the neutered x86 Atom will just chop everything to pieces.Quote:
On the netbook side of the mobile device booth, they had an array of some of the newer netbooks, all stuff that’s been already released. A couple were running MeeGo v1.0, and most of them had the Atom N450 underhood. The only really interesting ones were the Asus EeePC 1015PEM and the not-sold-in-America LG X140, both of which were fitted with the dual core N550. Intel was showing off the newfound 1080p playback ability for the dual core-equipped netbooks, a feat that previously required the Broadcom Crystal HD chip. Unfortunately, HD Flash video is still a no-go, and even videos playing from the hard drive max out the CPU and still manage to drop a few frames here and there. So dual core Atom netbooks can handle 1080p playback....kind of.
The reason why Atom sucks and why I don't think anyone should use one is because the Core 2 Duo really set the standard for modern computing performance (and K8 too). Taking a step back means sacrificing all the speed in loading and multitasking that we've grown to take for granted.
Apparently Anand agrees with me as well.Quote:
Moorestown and Oak Trail based tablets running lighter operating systems like MeeGo or Android 3.0 are where the future of Intel internet tablets lie.