Page 1 of 2 12 LastLast
Results 1 to 10 of 16

Thread: AMD's comeback with the Cyclos-Clock-Engine arming up the Piledriver

  1. #1
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default AMD's comeback with the Cyclos-Clock-Engine arming up the Piledriver

    Cyclos clock engine is based on the Resonance effect the clock engine don't need to push every clock it only need to fill up the resonance.

    http://en.wikipedia.org/wiki/Resonant_frequency

    Resonance also means it recycling energy.

    this means the 8core-Piledriver will get a base clock of 4ghz and more with the turbo.

    http://www.golem.de/news/prozessoren...202-89965.html

    Last edited by Qaridarium; 02-22-2012 at 03:26 PM.

  2. #2
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    http://www.gamestar.de/hardware/news...iledriver.html

    "Laut Samuel Naffziger von AMD war die sogenannte »Resonant Clock Mesh Technology« problemlos in den bisherigen Herstellungsprozess zu integrieren und die ersten produzierten Prozessoren hätten die Erwartungen erfüllt. Allerdings gab Naffziger auch zu, dass die Investition in die zuvor ungetestete neue Technik ein gewisses Risiko darstellte, das sich nun ausgezahlt habe."

    Samuel Naffziger pointed out that it was a high risk to bet on this untested technique but in the end amd win and now amd earn 10% more clock speed for free with the same power consuming.

    in may point of view this and ZRAM makes it possible for amd to beat intels 22nm cpu with a 32nm process.

  3. #3
    Join Date
    Jan 2008
    Posts
    772

    Default

    I don't understand how this increases clock speeds. Generating a fast clock is one thing; actually meeting setup time on all your flops is another.

  4. #4
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by Ex-Cyber View Post
    I don't understand how this increases clock speeds. Generating a fast clock is one thing; actually meeting setup time on all your flops is another.
    it increases clock speed because it reduce the heat and then you can clock higher in the same TDP.

    and it reduce the heat because it does recycling the electrons in the vibrating laboring and then the stored electrons push in the next cycle.

    all other clock engines only push electrons this clock engine don't push electrons it vibrating pushes the laboring big difference!

    because of this you need coils and condensers for the vibrating laboring directly on the core-die

    if the vibration laboring hits the resonance frequency you need ZERO energy to hold the clock!

    because of this the new clock engine pushes the laboring always to the resonance frequency-

    the resonance is very powerful:


  5. #5
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    AMD FX-Series FX-8150, 8x 3.60GHz Turbo 4.20GHz

    the new Piledriver amd cpu will have 4ghz and 4,6ghz turbo clock.

    and the new CPU name: AMD FX²-Series FX²-8250, 8x 4GHz Turbo 4.60GHz (Piledriver)
    Last edited by Qaridarium; 02-24-2012 at 09:36 AM.

  6. #6
    Join Date
    Jan 2007
    Posts
    459

    Default

    Quote Originally Posted by Qaridarium View Post
    http://www.gamestar.de/hardware/news/prozessoren/2565069/amd_piledriver.html

    "Laut Samuel Naffziger von AMD war die sogenannte »Resonant Clock Mesh Technology« problemlos in den bisherigen Herstellungsprozess zu integrieren und die ersten produzierten Prozessoren hätten die Erwartungen erfüllt. Allerdings gab Naffziger auch zu, dass die Investition in die zuvor ungetestete neue Technik ein gewisses Risiko darstellte, das sich nun ausgezahlt habe."

    Samuel Naffziger pointed out that it was a high risk to bet on this untested technique but in the end amd win and now amd earn 10% more clock speed for free with the same power consuming.

    in may point of view this and ZRAM makes it possible for amd to beat intels 22nm cpu with a 32nm process.

    LOL Qaridarium, "amd comeback" indeed, Cyclos's "resonant clock mesh technology" finally licenced by AMD after ARM already partnered with Cyclos in the early days (now you know one of the reasons why AMD were at the ARM venues and part of that PR) is for lowering the power envelope OR increasing the clocks as a side benefit, and lets be honest here AMD FX-Series FX-8150,Piledriver and all the rest need all the help they can get to lower the power and hopefully speed up the data throughput for a given watt even if they managed to slightly hand optimise the slow throughput "auto design" errors noone at amd bothered to correct the first time around.


    i really cant beleave your still falling for the AMD PR innovators sale pitch today, sure they can advertise 4.6ghz as a valid Number due to this 3rd party licencing deal, and they are first to use it commercially rather than ARM, they have their constantly evolving version sat on a tech bench waiting for the right time to put it into production, might as well let amd go through the pain of ramp up, "Cyclos announced a proof of concept processor implementation based on the ARM926EJ-processor in 2008 under the name "Project Elizabeth" along with the availability of design tool support."


    but ARM dont need to bring down the (milli)watts as a priority, AMD sure do today compared to intel, but thats beside the point , AMD are using this (anyone can)license tech today as it allows them to recycle more product from their yield waffers as in some cases yields actually improve when using resonant clock meshes with better power savings rather than generic clocks today, and even if that wasnt the case, as dan (ganousis) CEO of cyclos says "Most design teams today are quite happy to accept a small yield penalty in exchange for 10% total power reduction anyways."

    and whats your point with Zram, everyone knows MRAM is faster,cheaper and better ,plus already in commercial production OC for the longer term
    Last edited by popper; 02-24-2012 at 06:15 PM.

  7. #7
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by popper View Post
    LOL Qaridarium, "amd comeback" indeed, Cyclos's "resonant clock mesh technology" finally licenced by AMD after ARM already partnered with Cyclos in the early days (now you know one of the reasons why AMD were at the ARM venues and part of that PR) is for lowering the power envelope OR increasing the clocks as a side benefit, and lets be honest here AMD FX-Series FX-8150,Piledriver and all the rest need all the help they can get to lower the power and hopefully speed up the data throughput for a given watt even if they managed to slightly hand optimise the slow throughput "auto design" errors noone at amd bothered to correct the first time around.


    i really cant beleave your still falling for the AMD PR innovators sale pitch today, sure they can advertise 4.6ghz as a valid Number due to this 3rd party licencing deal, and they are first to use it commercially rather than ARM, they have their constantly evolving version sat on a tech bench waiting for the right time to put it into production, might as well let amd go through the pain of ramp up, "Cyclos announced a proof of concept processor implementation based on the ARM926EJ-processor in 2008 under the name "Project Elizabeth" along with the availability of design tool support."


    but ARM dont need to bring down the (milli)watts as a priority, AMD sure do today compared to intel, but thats beside the point , AMD are using this (anyone can)license tech today as it allows them to recycle more product from their yield waffers as in some cases yields actually improve when using resonant clock meshes with better power savings rather than generic clocks today, and even if that wasnt the case, as dan (ganousis) CEO of cyclos says "Most design teams today are quite happy to accept a small yield penalty in exchange for 10% total power reduction anyways."

    and whats your point with Zram, everyone knows MRAM is faster,cheaper and better ,plus already in commercial production OC for the longer term
    wood??? i write nothing wrong and you turn it into a hate speech.. LOL

    I'm happy about Cyclos's "resonant clock mesh technology" and you turn it into : "i really cant beleave your still falling for the AMD PR innovators sale pitch today,"

    and then you come with a Troll move ARM vs x86 but in fact AMD USE IT FIRST the Cyclos engine.

    you really buy ARM if you want a 64bit-highend system ? ? are you MAD man?

    and then you really compare Zram to Mram?? are you mad? you really compare volatile memory to non-volatile memory ???

    and be sure Zram is the cheapest ram ever build because you only need 1 transistor for every bit and Zero-capacitor.

    http://en.wikipedia.org/wiki/Z-RAM

    whatever i buy ARM if they get nativ 64bit for NAS systems and Servers and Routers.

    so be happy i'm not a ARM hater.

  8. #8
    Join Date
    Jan 2007
    Posts
    459

    Default

    Quote Originally Posted by Qaridarium View Post
    wood??? i write nothing wrong and you turn it into a hate speech.. LOL

    I'm happy about Cyclos's "resonant clock mesh technology" and you turn it into : "i really cant beleave your still falling for the AMD PR innovators sale pitch today,"

    and then you come with a Troll move ARM vs x86 but in fact AMD USE IT FIRST the Cyclos engine.

    you really buy ARM if you want a 64bit-highend system ? ? are you MAD man?

    and then you really compare Zram to Mram?? are you mad? you really compare volatile memory to non-volatile memory ???

    and be sure Zram is the cheapest ram ever build because you only need 1 transistor for every bit and Zero-capacitor.

    http://en.wikipedia.org/wiki/Z-RAM

    whatever i buy ARM if they get nativ 64bit for NAS systems and Servers and Routers.

    so be happy i'm not a ARM hater.
    LOL perhaps its a German thing but how you turn that in to a "hate speech" iv no idea...

    i actually like you Qaridarium and ill actually make the time to read your many badly translated posts several times to see if you have a good point to make and agree with it if so.

    I'm happy about Cyclos's "resonant clock mesh technology" and you turn it into : "i really cant beleave your still falling for the AMD PR innovators sale pitch today,"
    you appear happy that a larger ghz number is involved as the reason and the basis for your so called amd to "come back", not the real reason why Cyclos's resonant clock mesh technology is really interesting, and that is all this is "interesting", not something to base a so called amd comeback on, its that simple, dont go reading anything more in to it.

    you really compare volatile memory to non-volatile memory
    again you miss the point, ill spell it out, volatile memory needs constant power to refresh it, non-volatile memory does not, mram is as fast and durable as generic volatile memory dram today and its in production now,the only downside right now is density.

    will it become the standard everywhere to lower the ram power budget, perhaps not, but it does stand a chance sure, one things for sure the industry as a whole needs something right now to replace the crap write cycle endurance of solid-statememory/NAND-based flash memory and their currently best 533-megabit-per-second DDR interface NAND device implemented in sub-20-nm technology
    Last edited by popper; 02-24-2012 at 07:43 PM.

  9. #9
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by popper View Post
    LOL perhaps its a German thing but how you turn that in to a "hate speech" iv no idea...

    i actually like you Qaridarium and ill actually make the time to read your many badly translated posts several times to see if you have a good point to make and agree with it if so.
    yes thank you.

    Quote Originally Posted by popper View Post
    you appear happy that a larger ghz number is involved as the reason and the basis for your so called amd to "come back", not the real reason why Cyclos's resonant clock mesh technology is really interesting, and that is all this is "interesting", not something to base a so called amd comeback on, its that simple, dont go reading anything more in to it.
    Yes I'm happy abour more GHZ and why does this technique does not help amd to "comeback" ?

    Quote Originally Posted by popper View Post
    again you miss the point, ill spell it out, volatile memory needs constant power to refresh it, non-volatile memory does not, mram is as fast and durable as generic volatile memory dram today and its in production now,the only downside right now is density.
    no i don't miss the point because ZRAM is the ram with the highest density ever! (and the energy consuming is very low)
    so we are saying the same with different words.
    you can use mram+zram!

    Quote Originally Posted by popper View Post
    will it become the standard everywhere to lower the ram power budget, perhaps not, but it does stand a chance sure, one things for sure the industry as a whole needs something right now to replace the crap write cycle endurance of solid-statememory/NAND-based flash memory and their currently best 533-megabit-per-second DDR interface NAND device implemented in sub-20-nm technology
    sure MRAM are better than Flash.

  10. #10
    Join Date
    Aug 2007
    Posts
    6,633

    Default

    Well when you read the full article then piledriver will only save 5-10% energy. When you look at fx-61/81 that would mean that piledriver will use at max 1x w less, so still around 110w. When you look at ivb, then intel will use also less energy, from 95w to 77w (quad cores). And intel only needs 4 true cores to match that funny 8/4 core mix of amd amd definitely needs too much energy for the speed you get from it. in order to keep up with intel amd needs to rise the frequency way over 4 ghz. when you look at the single core speed then amd is not even faster than the i3 entry cpus (maybe even a higher clocked pentium g is enough). just by comparing an amd cpu vs intel cpu at same clock speed with cinebench in single core mode you should be able to interpolate the frequency amd needs to keep up with intel - just throwing more cores into the chip does not solve this basic issue...

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •