Like Tree7Likes
  • 5 Post By shri
  • 1 Post By huja
  • 1 Post By HowardCoombs

The end of Moore's Law - not such a bad thing...?

Closed Thread
  1. #1

    Join Date
    Aug 2008
    Location
    猴山
    Posts
    23,652

    The end of Moore's Law - not such a bad thing...?

    The future of computing | The Economist

    The ability of CPU makers to continually double the speed (and reduce the energy) of processors every 18 months or so has driven the world's technological advances.

    The new processing power has allowed smart devices from TV's to the powerful cloud servers. Software developers have chased this extra speed offering more services and better experiences.

    So if Moore's Law is ending is that a potentially crippling technology advancement?

    Almost certainly no...

    Software will now start to focus on new ways of being efficient with processing power where it is available.

    Are you worried that chip makers are struggling to make life faster?

  2. #2

    Join Date
    Dec 2002
    Location
    薄扶林
    Posts
    47,963

    I really think its time for software makers to start focusing on optimising their software.

    Back in the day when I used to write software, we were so worried about optimising speed, file sizes etc... it was ridiculous (had to fit the whole installer in a 5 inch floopy else the cost of goods went up and the manufacturing process got more cumbersome!

    There is a renaissance of sorts in some sectors of the interwebs - optimising web response time, reducing payloads, caching, compression etc etc... partly because what was once a "super highway" fuelled by multi-megabit networks has become a multi-lane slow speed congested road thanks to mobile users and high adoption rates...

    Interwebs "software" for the lack of a better term is going to be the next "doubler" ... not CPU or bandwidth.

    Last edited by shri; 14-03-2016 at 09:08 AM.

  3. #3

    Join Date
    Nov 2008
    Posts
    4,574

    Big driving forces in computing power are financial services (I don't need to explain how much money is involved in this sector) and gaming (by some measures, equal to the economic size of Hollywood).

    Financial services pushes things with customers demanding ever lower latencies and higher throughput's, we've also learnt a lot in terms of resilience and network design from FS.

    Gaming also brought us very low latency, think online gaming and "lag", the code that goes into making sure your bullets in COD make sense is a work of art. Gaming also brought us GPU advances, which have lent themselves to other fields such as engineering and bio-sciences where certain types of calculation benefit from the skills of GPU's.

    Agreed, code needs to become more efficient, but let's not assume all code is on par with Acrobat Reader or any number of fat apps out there, there is plenty going on from the software side pushing the needle further.


  4. #4

    Join Date
    Dec 2011
    Location
    Tsim Sha Tsui
    Posts
    3,963

    I do agree with we should concentrate more on software, but learning how to write unique code from scratch. Nearly everything is open source nowadays and everything is a hap dash of copy and paste, ending up with really badly written code.


  5. #5

    Join Date
    Jun 2010
    Location
    Tri-State
    Posts
    11,644

    The bottleneck in advancement is a moving target. At various times it's been CPU speed, internet bandwidth, battery power . . . every link of the technology chain will eventually have it's day in the sun as the problem to solve to move forward.

    Rob2020 likes this.

  6. #6

    Join Date
    Dec 2002
    Location
    薄扶林
    Posts
    47,963
    Financial services pushes things with customers demanding ever lower latencies and higher throughput's, we've also learnt a lot in terms of resilience and network design from FS.
    And I was lead to believe that the earliest drivers of FS learning were from the porn industry - they knew how to scale and deliver.

    Lets not forget...

    Very very little from the FS industry comes back into the consumer web. When you look at anecdotal stats like "25% of the web's content is hosted on Wordpress" you can start to think of the bloat the is being passed around on the wires.

    Even "Long URLs" end up in bandwidth wastage ... (increasing latency) -

    Are Long URLs Wasting Bandwidth? - Slashdot

    (On a personal note, I'd started a project a few months ago, which looks at performance from an OSI model ... down from the DNS to the last piece of text being rendered on your browser and even with my very very limited resources and diminished braincells concluded that the interwebs could be faster, cheaper and better if people avoided what proplus has said...

    For the ubergeeks - I hope . soon-ish-sometime-in-the-near-future to have some time to spec it out and release a "headless cms" which will replace wordpress and make me the king of the hill... or something like that!

    )

  7. #7

    Join Date
    Aug 2012
    Location
    Location Location
    Posts
    1,201

    Am I worried?

    The Luddite in me is somewhat thankful that the machines won’t take over my job quite so soon.
    The progressive in me is somewhat worried that by the time I retire there won’t be enough young people to keep the economy running.


  8. #8

    Join Date
    Mar 2007
    Posts
    13,099
    Quote Originally Posted by East_coast:
    Software will now start to focus on new ways of being efficient with processing power where it is available.
    You mean old ways!

    I grew up with old and slow and tiny computers where games fit into 4k and you had spreadsheet & word processors that fit into 16k of memory. Programmers had to do a lot of optimisations to be able to work in such tiny spaces.

    It will be good to teach the new kids on the block some very old tricks.
    shri likes this.