Re: Pi to 10 trillion places


"Ronald Bruck" <bruck@xxxxxxxxxxxx> wrote in message
news:270120061835489434%bruck@xxxxxxxxxxxxxxx
> In article <43da5ba1$0$45181$892e7fe2@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
> Boris Uion <nospam@xxxxxxxxxx> wrote:
>
>> "Ronald Bruck" <bruck@xxxxxxxxxxxx> wrote in message
>> news:270120060815543687%bruck@xxxxxxxxxxxxxxx
>> > In article <43da2db7$0$85453$892e7fe2@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
>> > Boris Uon <nospam@xxxxxxxxxx> wrote:
>> >
>> >> <robin.bruce@xxxxxxxxx> wrote in message
>> >> news:1138368257.710242.37420@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>> >> > No,
>> >> >
>> >> > I've not done it, but just looking at a graph for pi of decimal
>> >> > places
>> >> > versus time (in years). It looks like we're due another step up from
>> >> > Kanada's 1.2 trillion places in 2002. Anyone know who's working on
>> >> > this
>> >> > stuff and how they're progressing? What's the major challenge to
>> >> > overcome? Is it concepts in computer arithmetic that are lacking? Is
>> >> > it
>> >> > the difficulties in co-ordinating the multiple processing elements
>> >> > required or are there fundamental barriers in theory to be overcome?
>> >> > Is
>> >> > it a weighted combination of the three?
>> >> >
>> >> > I'm coming at this from the perspective of someone developing
>> >> > FPGA-based supercomputing nodes with an interest in potential future
>> >> > applications.
>> >> >
>> >> > Cheers,
>> >> >
>> >> > Robin Bruce
>> >> >
>> >>
>> >> There is no money to fund it. There is no need for it. Where would
>> >> you
>> >> put it? How could you check it?
>> >
>> > No intrinsic need, indeed. But my understanding is that the driving
>> > force behind these extended calculations is the need to test new
>> > hardware. If I run the calculation on new hardware to 2 trillion
>> > places, and it agrees with the previous 1.2-trillion place record, then
>> > the chances are very good that there are no hardware bugs.
>> >
>> > --Ron Bruck
>>
>> but, how you gonna check that it is right?
>> and what do you mean "hardware" at the chip level it is just accumulator,
>> and low level firmware,
>> nothing requiring advance technology.
>
> If it agrees with the previous result to the original number of
> significant figures, the probability that the full result is correct is
> overwhelming.

not at all.
you would be assuming it is correct or guessing.

>
> As for "accumulator and low level firmware"... Ever hear of the
> Pentium FP bug?

early pentiums, but they fixed it.


.

Relevant Pages

  • Re: Pi to 10 trillion places
    ... but just looking at a graph for pi of decimal places ... >> required or are there fundamental barriers in theory to be overcome? ... force behind these extended calculations is the need to test new ... If I run the calculation on new hardware to 2 trillion ...
    (sci.math)
  • Re: Pi to 10 trillion places
    ... > force behind these extended calculations is the need to test new ... If I run the calculation on new hardware to 2 trillion ... and low level firmware, ... nothing requiring advance technology. ...
    (sci.math)
Loading