ISSCC 2012 Preview online

Article: ISSCC 2012 Preview
By: goose (go.delete@this.away.now), December 1, 2011 8:28 pm
Room: Moderated Discussions
Vincent Diepeveen (diep@xs4all.nl) on 12/1/11 wrote:
---------------------------
>goose (go@away.now) on 12/1/11 wrote:
>---------------------------
>>Vincent Diepeveen (diep@xs4all.nl) on 11/29/11 wrote:
>>---------------------------
>>>David Kanter (dkanter@realworldtech.com) on 11/29/11 wrote:
>>>---------------------------
>>>>Back from Thanksgiving, we have a new article ready to go. For those who can't
>>>>quite wait to hear what ISSCC has in store in February:
>>>>
>>>>http://www.realworldtech.com/page.cfm?ArticleID=RWT112811200830
>>>>
>>>>Highlights of the upcoming 2012 ISSCC include the first 22nm disclosures from Intel
>>>>and several SoC papers from AMD, Cavium Networks and Oracle. Looking out further
>>>>to the future, the clear focus is power consumption. There are several papers from
>>>>Intel on low-power logic, one from IBM discussing 3D integration of embedded DRAM
>>>>and a third from Fujitsu on system level power for the K supercomputer.
>>>>
>>>>David
>>>
>>>You mean they are gonna eat more power again meanwhile saying loud they do not?
>>>
>>>I remember how some years ago entire desktops power consumption was 60 watt. Now
>>>laptops are 60 watt, a desktop is well over 250 watt and a 2 socket machine or
>>>an overclocked single socket box is well over 400 watt.
>>>
>>>That's excluding the gpgpu card, which eats another 400-500 watt from the power tap.
>>>
>>>
>>
>>You really think that a standard desktop without a discrete GPU eats 250 watts?
>>Get one of those plug packs that is a watt meter, i think you will be rather surprised.
>
>Yes we have a lot of those over here. But i get the impression you do not.
>
>>
>>http://www.anandtech.com/show/4083/the-sandy-bridge-review-intel-core-i7-2600k-i5-2500k-core-i3-2100-tested/21
>>
>>
>>If in case of TLDR; a 2600K only used 128 watts at the wall at full load.
>
>2600k is just a 4 core chip and expensive. Not interesting therefore.
>
>Only a very old machine mine has 4 cores. It's from 2005.

Given that only enthusiast (read: repackaged server CPUs) have more than 4 cores, to throw out my argument is short sighted.

That being said, even an i7 3960X system doing an x264 transcode uses less than 250 watts

http://www.anandtech.com/show/5091/intel-core-i7-3960x-sandy-bridge-e-review-keeping-the-high-end-alive/7

>
>>
>>Also, NO and I mean NO, GPU eats 500 watts by itself.
>
>Watt meter here proves you wrong.

Your statement is retarded. A watt meter shows complete system draw, not the video card. You have no way of isolating the GPUs power draw.

>
>>Below is a link to an article about the GTX590, the current big boy, and it has
>>a total system draw of less than 500 watts.
>
>HD 6990 is of course the king everywhere. 1.4Tflop double precision and 5.4 Tflop single precision.

We were talking about power, not performance which is why I chose the 590.

>
>As most codes i run are not capable of multiply-add, you can from *all* manufacturers
>divide that capability by 2 by the way.
>
>As you know the GTX590 is crippled in double precision by factor 4. So it has just
>25% of the double precision capacity of tesla times 2. meanwhile you pay a big price for the GTX590.
>
>It's fast for 32x32 multiplication.

Off topic rant.

>
>>http://www.bit-tech.net/hardware/graphics/2011/03/24/nvidia-geforce-gtx-590-3gb-review/8
>>
>>If you think that I cherry picked those results, know that I simply googled the
>>one on the 590 and the other one was an assumption that Anand had tested the power usage of the 2600K.
>
>Yes you just can google and over here we test.
>Thank you.

Why do you think I suggested that you invest in a watt meter, which you apparently have? Hint, I tested myself. The links are there as supporting evidence and to allow you to reproduce their results as they have clearly defined methodologies.

This is a whole lot more than what you have done. Post some system specs and a methodology and I will verify. Otherwise stop talking crap.

>
>>That being said feel free to call BS, as I am certain I can provide plenty more evidence where they came from.
>
>NSA

??

>
>>Finally but most importantly, new systems are dramatically faster than those of
>>yesteryear but power has not increased anywhere near the same rate. This means that
>>they are getting considerably more efficient.
>>
>>Remember that we ask more of our current PCs then we used to. If you gave each
>>system the same workload, the new one will use considerably less power than the old one.
>>
>

Interestingly I noticed that you have posted the same stuff in the knights corner thread. You must be really pissed off with the world.
< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
ISSCC 2012 Preview onlineDavid Kanter2011/11/29 02:47 PM
  ISSCC 2012 Preview onlineVincent Diepeveen2011/11/29 04:10 PM
    ISSCC 2012 Preview onlinegoose2011/12/01 09:36 AM
      ISSCC 2012 Preview onlineVincent Diepeveen2011/12/01 09:42 AM
        ISSCC 2012 Preview onlineJouni Osmala2011/12/01 10:50 AM
        ISSCC 2012 Preview onlinegoose2011/12/01 08:28 PM
    Go write this bullshit on your blog instead (NT)Megol2011/12/01 10:11 AM
    ISSCC 2012 Preview onlinebakaneko2011/12/01 11:09 AM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell purple?