MacBidouille posts PPC 970 benchmarks

1141517192034

Comments

  • Reply 321 of 665
    snoopysnoopy Posts: 1,901member
    Quote:

    Originally posted by KidRed



    . . . So it looks like 970 around August, so maybe announced at MWNY and shipping immediately or within weeks.






    Why? Here is what he said regarding a question about August. "90% likelihood before then. 100% likelihood within 30 days of then." Looks like he says 'before' August is the most likely.
     0Likes 0Dislikes 0Informatives
  • Reply 322 of 665
    bigcbigc Posts: 1,224member
    Quote:

    Originally posted by snoopy

    Why? Here is what he said regarding a question about August. "90% likelihood before then. 100% likelihood within 30 days of then." Looks like he says 'before' August is the most likely.



    Question is: ¿Is that when you can order them for 6-8 week delivery or get them off-the-shelf?
     0Likes 0Dislikes 0Informatives
  • Reply 323 of 665
    bigcbigc Posts: 1,224member
    Quote:

    Originally posted by Escher

    I haven't done any of these electrical calculations since high school, but it always helped me to think of the units used for the different variables. Using units, the above equation would go like this:



    Watt = Ampère x Volt



    So if we have a processor drawing 19 Watt at 1.1 Volt, it does indeed draw 17.3 Amps. As pointed out above, a household vacuum cleaner would draw 100 times the voltage of a 1.2 Ghz PPC 970 (200 times in Europe). That's why it would draw about a 100 times more Watts than said PPC 970. Does this make more sense?



    Escher




    Try Power ~ ½ CV2Af

    C= Capacitance

    V= Core Voltage

    A= Activity Factor

    f= Clock Frequency



    It's a dynamic switch and not a heater coil.



    edit - that's Voltage squared
     0Likes 0Dislikes 0Informatives
  • Reply 324 of 665
    derrick 61derrick 61 Posts: 178member
    Quote:

    Originally posted by Leonis

    Rage 128 with 8MB VRAM



    Yeah...in the top end
     0Likes 0Dislikes 0Informatives
  • Reply 325 of 665
    programmerprogrammer Posts: 3,503member
    Quote:

    Originally posted by Bigc

    Well sounds like were back to the original schedule August-September release and maybe pre-order in July or at WWDC



    "back to"? Some of us never left.
     0Likes 0Dislikes 0Informatives
  • Reply 326 of 665
    snoopysnoopy Posts: 1,901member
    Quote:

    Originally posted by Bigc

    Try Power ~ ½ CV2Af

    C= Capacitance

    V= Core Voltage

    A= Activity Factor

    f= Clock Frequency



    It's a dynamic switch and not a heater coil.



    edit - that's Voltage squared




    Current times voltage works for any black box. It does not matter what is inside. This calculation will give you the power dissipation for any equipment powered by DC voltage sources, as in computers. AC voltage sources are a different matter, and we need a power factor term to take into account phase differences.



    The formula you give is a special case to estimate power. The one half C V-squared term is the energy stored in a capacitor. Since it is multiplied by a frequency it has a one over time term, so the units look okay. Power is energy divided by time, or rate of energy delivery.



    So your formula estimates energy by the size of capacitor being discharged during a switching cycle, and the duty cycle is what you call the activity factor. I guess if one were to add up all the switches in a processor with some educated guess about how often they switch, you might get in the ball park for power. In the end, they get the real numbers from measurements of the processor as a black box. Power = current X voltage.
     0Likes 0Dislikes 0Informatives
  • Reply 327 of 665
    bigcbigc Posts: 1,224member
    Quote:

    Originally posted by snoopy

    Current times voltage works for any black box. It does not matter what is inside. This calculation will give you the power dissipation for any equipment powered by DC voltage sources, as in computers. AC voltage sources are a different matter, and we need a power factor term to take into account phase differences.



    The formula you give is a special case to estimate power. The one half C V-squared term is the energy stored in a capacitor. Since it is multiplied by a frequency it has a one over time term, so the units look okay. Power is energy divided by time, or rate of energy delivery.



    So your formula estimates energy by the size of capacitor being discharged during a switching cycle, and the duty cycle is what you call the activity factor. I guess if one were to add up all the switches in a processor with some educated guess about how often they switch, you might get in the ball park for power. In the end, they get the real numbers from measurements of the processor as a black box. Power = current X voltage.




    not for a capacitive-discharge system which has a frequency of oscillation, which is what a microprocessor is. If there is no oscillation then no power. Under dead short conditions then maybe Power = volts*amps since there is no oscillation at that time.



    ... and Programmer I'm with you, I've never left.
     0Likes 0Dislikes 0Informatives
  • Reply 328 of 665
    bootsboots Posts: 33member
    er, you might want to google for "leakage current". A gate will leak some current even when not undergoing changes of state.
     0Likes 0Dislikes 0Informatives
  • Reply 329 of 665
    bigcbigc Posts: 1,224member
    Quote:

    Originally posted by boots

    er, you might want to google for "leakage current". A gate will leak some current even when not undergoing changes of state.



    well hopefully it's not much or there will be problems.
     0Likes 0Dislikes 0Informatives
  • Reply 330 of 665
    costiquecostique Posts: 1,084member
    Quote:

    Originally posted by KidRed

    Very interesting thread. I for one believe the Powerjack (MacWhispers guy) to be legit.



    My favourite one from the forum:
    Quote:

    if you are under NDA, why would you tell us you are under NDA, that makes it more suspicious



     0Likes 0Dislikes 0Informatives
  • Reply 331 of 665
    snoopysnoopy Posts: 1,901member
    Quote:

    Originally posted by Bigc

    not for a capacitive-discharge system which has a frequency of oscillation, which is what a microprocessor is. If there is no oscillation then no power. Under dead short conditions then maybe Power = volts*amps since there is no oscillation at that time. . .







    The laws of physics don't change depending on how you design a system. The formula you give, provided your estimates are all correct, and the formula for power based on measurement of voltage and current to the CPU will both give the same answer. The problem with your formula is that you must know the capacitance and duty cycle of each switch. With millions of switches this is a big task, and it is inaccurate because you do not know the values exactly enough.



    If you think about it, you will realize that the very things that increase and decrease power in your formula will also increase and decrease current from the power supply. If the CPU is inactive, the switches have low or zero activity factor in your formula, and current from the supply is also low, resulting in low power either way you calculate it. The 17 Amperes at 1.1 Volts is base on IBM's reference to 19 Watts. This is likely running full out, using some standard conditions for measuring maximum power consumption.



    How do you think IBM got that 19 Watts? (Or the 42 Watts?) They measured it. They didn't do some estimate based on capacitance and duty cycle. There is absolutely no way to "measure" power dissipation using your formula.
     0Likes 0Dislikes 0Informatives
  • Reply 332 of 665
    mr. memr. me Posts: 3,221member
    Quote:

    Originally posted by snoopy

    The laws of physics don't change depending on how you design a system. The formula you give, provided your estimates are all correct, and the formula for power based on measurement of voltage and current to the CPU will both give the same answer. The problem with your formula is that you must know the capacitance and duty cycle of each switch. With millions of switches this is a big task, and it is inaccurate because you do not know the values exactly enough.



    If you think about it, you will realize that the very things that increase and decrease power in your formula will also increase and decrease current from the power supply. If the CPU is inactive, the switches have low or zero activity factor in your formula, and current from the supply is also low, resulting in low power either way you calculate it. The 17 Amperes at 1.1 Volts is base on IBM's reference to 19 Watts. This is likely running full out, using some standard conditions for measuring maximum power consumption.



    How do you think IBM got that 19 Watts? (Or the 42 Watts?) They measured it. They didn't do some estimate based on capacitance and duty cycle. There is absolutely no way to "measure" power dissipation using your formula.




    The confusion here seems to be that people are getting the processor must be drawing 17 amperes directly from the plug in the wall. That is not the case. To power the processor alone, the AC plug in the wall must draw 158 mA, assuming 120 V line voltage. Obviously, this is a small fraction of the current required to power the whole system. The line voltage is transformed and rectified to 1.1 V DC to power the processor. As stated above, the 17 A current through the processor is dominated by the switching rate of capacitive elements. It is not the current through the heating element of a heater or light bulb.
     0Likes 0Dislikes 0Informatives
  • Reply 333 of 665
    anandanand Posts: 285member
    Anybody that believes Macwishpers is a fool. That guy has lied about products that he sells. He is at best a con artist and at worst just a theif.



    The only rumor sites out their that has seems to have any real information are "Think Secret" and "Mac rumors" and they have said nothing about the 970. Nothing. They were right about the iMacs, they were right about the eMacs and they were even right about the 17 inch powerbooks.
     0Likes 0Dislikes 0Informatives
  • Reply 334 of 665
    davegeedavegee Posts: 2,765member
    Quote:

    Originally posted by anand

    The only rumor sites out their that has seems to have any real information are "Think Secret" and "Mac rumors" and they have said nothing about the 970. Nothing. They were right about the iMacs, they were right about the eMacs and they were even right about the 17 inch powerbooks.



    Oh trust me Nick (Think Secret) is well informed about the 970, that you can BET ON. The big difference that I see with Nick when compared to just about all the other rumors sites is that Nick works hard on the stories he posts and it reflects in his site. He doesn't jump the gun on anything and as such we see one maybe two items each week (if we're lucky). When you see something posted on TS, 99.9% of the time 'the story is a lock' and that's what makes TS special.



    In short Nick knows what the words '2nd source' means and he also knows that lots of stuff that he hears that WERE true can change/morph or even die as time goes on.



    Case in point:



    Anyone remember those iBrary screen shots? Well here's a little nugget... I've been told that they were 'the real deal' (a skunkworks project at Apple) but since then for reasons unknown the project was put back in the basement... I'm sure tons more projects just like that are going on at Apple each and every day and only the best of em will ever see the light of day.



    Had Nick jumped the gun done a story on iBrary (and it never came out) it would have made TS look pretty bad even though the project was REAL and his source was on the up and up...



    I take my hat off to very few rumor reporters Nick, Matt Rothenberg and the legend NMR. These folks have GREAT contacts and enough insight to hold back till they have something to say that they KNOW will come true. The rest of the rumor sites are a good read and fun but you can't bank on em.



    Dave
     0Likes 0Dislikes 0Informatives
  • Reply 335 of 665
    kidredkidred Posts: 2,402member
    Quote:

    Originally posted by snoopy

    Why? Here is what he said regarding a question about August. "90% likelihood before then. 100% likelihood within 30 days of then." Looks like he says 'before' August is the most likely.



    And? 'Around August' is 'before August', 'during August' and even 'after August'. Besides, he's only 90% sure it's before August, the within 30 days of August leads us to the original 'around August'. Not sure what your beef is.
     0Likes 0Dislikes 0Informatives
  • Reply 336 of 665
    bigcbigc Posts: 1,224member
    Quote:

    Originally posted by snoopy

    The laws of physics don't change depending on how you design a system. The formula you give, provided your estimates are all correct, and the formula for power based on measurement of voltage and current to the CPU will both give the same answer. The problem with your formula is that you must know the capacitance and duty cycle of each switch. With millions of switches this is a big task, and it is inaccurate because you do not know the values exactly enough.



    If you think about it, you will realize that the very things that increase and decrease power in your formula will also increase and decrease current from the power supply. If the CPU is inactive, the switches have low or zero activity factor in your formula, and current from the supply is also low, resulting in low power either way you calculate it. The 17 Amperes at 1.1 Volts is base on IBM's reference to 19 Watts. This is likely running full out, using some standard conditions for measuring maximum power consumption.



    How do you think IBM got that 19 Watts? (Or the 42 Watts?) They measured it. They didn't do some estimate based on capacitance and duty cycle. There is absolutely no way to "measure" power dissipation using your formula.






    I guess you should just read this or try searching with google.com for some information on CPU power usage. You obviously assume.



    ... and no/yes Watts = Volts * Amps. In a 3-phase circuit, though, don't forget to multiply line current by the square-root of 3 and multiply by the Power Factor.



    Enough said.
     0Likes 0Dislikes 0Informatives
  • Reply 337 of 665
    kidredkidred Posts: 2,402member
    Quote:

    Originally posted by anand

    Anybody that believes Macwishpers is a fool. That guy has lied about products that he sells. He is at best a con artist and at worst just a theif.



    The only rumor sites out their that has seems to have any real information are "Think Secret" and "Mac rumors" and they have said nothing about the 970. Nothing. They were right about the iMacs, they were right about the eMacs and they were even right about the 17 inch powerbooks.




    After reading a macintouch page with emails from quite a few people I changed my stance and opinion. He doesn't sound very knowledgeable at least not in the rumor sence. If anything, it may seem he's in the rumor game to get traffic to sell others and potentially his own products.
     0Likes 0Dislikes 0Informatives
  • Reply 338 of 665
    rhumgodrhumgod Posts: 1,289member
    Regarding that whole Powerjack post on spymac rumor forums, the only info I have come up with has been this one,

    URL here:



    ECS reportedly lands Apple 12-inch PowerBook orders Compiled from outside sources; Chinmei Sung, DigiTimes.com [Tuesday 21 January 2003] Elitegroup Computer Systems (ECS) has landed orders from Apple Computer to contract manufacture the 12-inch PowerBook unveiled early this month at the Macworld Conference and Expo, reported the Chinese-language Economic Daily News. Shipment volume is estimated at 300,000 units a year. The company also produces iBooks for the US vendor. Apple Computer:



    Taiwan ODM manufacturers



    Quanta Computer

    - 15.2? PowerBook, 17? flat-panel iMac



    Compal Electronics

    - 15.4? PowerBook, 17? PowerBook



    ECS

    - 12? PowerBook, iBook



    Foxconn Electronics*

    - iMac/eMac



    Source: compiled by DigiTimes, January 2003. *The registered trade name of Hon Hai Precision Industry
     0Likes 0Dislikes 0Informatives
  • Reply 339 of 665
    snoopysnoopy Posts: 1,901member
    Quote:

    Originally posted by KidRed

    And? 'Around August' is 'before August', 'during August' and even 'after August'. Besides, he's only 90% sure it's before August, the within 30 days of August leads us to the original 'around August'. Not sure what your beef is.



    Sorry about that. I was really responding to a reply to your post, which interpreted you to mean we would not see the 970 until September. In my haste, I thought it would be better to quote your post too, but as soon as it appeared on the message board I realized I was wrong. I apologize. I should have quoted what I was replying to.
     0Likes 0Dislikes 0Informatives
  • Reply 340 of 665
    bill mbill m Posts: 324member
    Quote:

    Originally posted by KidRed

    After reading a macintouch page with emails from quite a few people I changed my stance and opinion. He doesn't sound very knowledgeable at least not in the rumor sence. If anything, it may seem he's in the rumor game to get traffic to sell others and potentially his own products.



    Precisely my point.



    8)
     0Likes 0Dislikes 0Informatives
Sign In or Register to comment.