Slashdot Mirror
Intel's New Chips, High Power And Low
sebFlyte writes "Centrino has been one of Intel's major successes of late, and they've just released the replacement, Sonoma. ZDNet has stripped the new chipset, and published a review of the new kit with all the technical details of what this new chipset will do for your laptop."
ZeroOne42 adds a link to Hardware Zone's exhaustive look at Sonoma, "complete with benchmark results between a Sonoma notebook (Fujitsu E8020) and a Centrino one (Gigabyte N512). Looks like Sonoma is closing up the technological gap between desktops and notebooks."
And on the desktop side, foxalopex writes "It seems that Intel's new dual-core CPU chips will have some of the highest wattage ratings ever seen on the X86 CPU market, which, according to Tom, wasn't what they initially said would happen. I guess this isn't too surprising seeing how AMD's been beating them on power usage in the last several revisions of chips."
This is even worse then it seems. The actual power usage (from intel.com) is at 5mWA/ms, compared with the target of 3mWA. Thats a full 2 milliwatt amps over their target, and much higher then AMD!
Answer: 100%
From an energy perspective, the CPU only converts electricity into heat; it's not like a light bulb that converts x% of the power into light and y% into heat. Energy can't turn into processed instructions.
Now, if you want to know how power efficient is a processor, you'd have to obtain the MIPS-Watts or FLOPS-Watts ratio, and compare these numbers between CPUs.
atcurtis:
I think that they may be taking the wrong approach by putting 2 whole processor cores on the same die...
We have SMT (HyperThreading in Intelese) which in my opinion is a pretty decent idea... just a crying shame about how they set about doing it. They sacrificed the silicon used by the original P4's integrated RamBus memory controller and put in the necessary silicon for their HT technology. The idea of getting an extra CPU for 'free' in the current HT processors doesn't work because in a demanding application, most of the execution units will be busy anyways.
No P4 has ever had an integrated memory controller. The original Williamette cores interfaced to RDRAM through the i850 chipset. In fact, there were a number of Williamette machines that had PC-133 memory (slow as Christmas, but they existed). I owned an RDRAM one and I worked on a PC133 memory one at a job site.
Because of this, many old RamBus P4 machines can outperform their newer P4 siblings - mostly because the newer P4 do not have an integrated memory controller and have to go through the IO originally for peripherals. (ok, they are fixing this with the much higher pincount chips than the 1st gen P4 which did not need all that IO due to integrated RDRAM controller)
I've never seen a Williamette outperform any later Pentium4 core and again, P4s haven't had IMCs.
Instead, I believe that they need to design a processor with the original intent to be hyperthreaded (instead of the P4's original intent, to use RDRAM). What this means, is perhaps provide many execution units, maybe 50% more than what a single processor requires, and then make it look like 2 CPUs. Or perhaps double or triple the number of units and make it look like 4 CPUs to the software.
Eventually, if you provide 2x the number of execution units that are "needed", what difference do you have from a dual core processor? Some units would be shared - fetch, decode, memory stages, etc, but you'd be getting close anyway because of the added interconnect logic for more execution units (it's an n-squared problem with the number of execution units for data forwarding and data hazard detection/resolution).
So... What they need for the consumer is a high-pincount device which is truely designed for hyperthreading (ie, has enough execution units available to be able to perform nearly as good as having a whole 2nd CPU)
And for the server market, bring back the integrated RamBus controller, still have plenty of pins so that the server can have perhaps 4 or more RDRAM channels to keep the data flowing fast enough to keep the 4 SMT logical processors occupied. (IIRC, the original P4 has 2 RDRAM 800 channels)
You also have to remember that RDRAM isn't as wide as DDR, for example. It's fairly narrow in the scheme of things. Having two channels can make is wider just like dual channel DDR memory.
And while I am in my Intel rant mood, I'll criticise the Itanic... Surely with the EPIC architecture, all that branch-prediction and other crud they have in the processor is unnecessary... They need to cut away 2/3rds of the silicon, and get people to write compilers which really do work for them. IIRC, the whole point of all that extra cruft is to make it perform ok for brain-dead compilers. Either they get decent compilers out there (perhaps, open-source their Itanic compiler optimiser) or admit that EPIC was "another nice idea, pity it doesn't work in practice".
Branch prediction is hardly "crud". If your CPU is capable of performing dynamic branching (branches based on the results of an operation as opposed to a hard static branch such as BRA (branch always)) and has a penalty with pipeline flushes on branching, then you probably can benefit from branch prediction). Also, EPIC *hardly* performs "OK" with brain-dead compilers. You have to have a decent compiler to get the performance out of it. A brain-dead compiler won't get 25% out of the CPU's max performance
Ok, so since Laptops were first created, their whole idea was to be a way someone could do work remotely or in areas where having a desktop was hard or near imposibe for one reason or another. in 2003 "November 2003 survey of Penn State University undergraduates found that freshmen were more likely to own laptops than upperclassmen. Of 1,838 respondents, 73.7 percent own a desktop, 32.2 percent own a laptop, 9.2 percent own both, and 3.4 percent own neither." http://pittsburgh.bizjournals.com/pittsburgh/stori es/2004/08/23/focus1.html
e ID=1513
e gory=31558&item=6736232824&rd=1&ssPageName=WDVW
(sorry if i am not suposed to link to auctions, it was the only english page i could find since its a japanese laptop)
"t). Currently, one-quarter of Americans own a laptop or notebook computer (23 percent)" http://www.thegoodsteward.com/article.php3?articl
What does this tell us? it says that laptops are becoming more and more popular, and as they become more and more popular, there is a bigger drive towards creating the "ultimite laptop"
As far as i see it, Apple and IBM are the only good laptop companies. i know thats a dramatic statement, but look at it, any laptop over 6LBS is WAY TOO HEAVY to carry around. Most HP/Compaq laptops are 7 to 9 lbs. Most dells are in that same range.
what the laptop industry needs is a re-working of laptops. Sony has just released the X505 VAIO laptop, this laptop is built in the way all laptops should be built. It weighs a mear 1.73lbs WITH battery in it, and has enough power to run almost anything exept games and video editing.( But trying to play games on a laptop is just stupid anyway, small screen and no mouse or full keypad) http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&cat
The new centrino chips are amazing, to put it simply. If you put a 2ghz Pentium M into a desktop and slap on a liquid cooling system, you can overclock it to be way more powerful then a 3.46EE or even a 3.8ghz P4. The pentium M is just years ahead of its time, and people having figured it out yet! Its kinda like black lotus for all you magic players!
The prices aren't that bad on the low-end Dothans if you put them in, say, DFI's board. They overclock QUITE nicely ;-)
That said, if you don't like the board prices, wait for the DFI 852GME-MGF. Identical to the 855GME-MGF, but with the much cheaper, much more overclockable, and not being phased out 852GME chipset (which is pin-compatible, but not supposed to run with a P-M)