unknown@ucscb.UCSC.EDU (The Unknown User) (10/09/90)
In article <14053@smoke.BRL.MIL> gwyn@smoke.BRL.MIL (Doug Gwyn) writes: >In article <7587@darkstar.ucsc.edu> unknown@ucscb.UCSC.EDU (The Unknown User) writes: >>1/17000000 on my HP calculator! > >While that is a portion of the requisite calculation, the actual specs >need to be determined by a variety of factors, most of which are contained >in the hardware reference manual (although it doesn't try to teach you >digital circuit design). Don't rely on such a simplistic calculation.. That's a simplistic calculation, yes, but is it not correct that it's the LOWEST (that is, FASTEST) value you'd ever need for a certain speed of a CPU? I meant for it to be a simplistic calculation, I just thought that would give an asymptotic value which would be the minimum ever needed. -- / Apple II(GS) Forever! unknown@ucscb.ucsc.edu \ \"If cartoons were meant for adults, they'd be on in prime time."-Lisa Simpson/
kadickey@phoenix.Princeton.EDU (Kent Andrew Dickey) (10/09/90)
In article <7619@darkstar.ucsc.edu> unknown@ucscb.UCSC.EDU (The Unknown User) writes: >In article <14053@smoke.BRL.MIL> gwyn@smoke.BRL.MIL (Doug Gwyn) writes: >>In article <7587@darkstar.ucsc.edu> unknown@ucscb.UCSC.EDU (The Unknown User) writes: >>>1/17000000 on my HP calculator! >> >>While that is a portion of the requisite calculation, the actual specs >>need to be determined by a variety of factors, most of which are contained >>in the hardware reference manual (although it doesn't try to teach you >>digital circuit design). Don't rely on such a simplistic calculation.. > > That's a simplistic calculation, yes, but is it not correct that it's >the LOWEST (that is, FASTEST) value you'd ever need for a certain speed of >a CPU? I meant for it to be a simplistic calculation, I just thought that >would give an asymptotic value which would be the minimum ever needed. No, it is not necessarily correct. Hardware designers are free to do whatever they like with their system to improve performance. I believe the Transwarp design requires static RAMs that run at 4 times the obvious speed. (That is, 10MHz operation requires 25ns RAM). When you get much below 20ns, the price goes up through the roof. Perhaps ZipGS (if it ever materializes) will not have this requirement (or maybe only a factor of 2 difference). btw, does anyone know if a RamKeeper w/ a GSRAM is fully DMA compatible? What if I plug two cards into the RamKeeper? I need to determine if I should just get a good deal on a cheap Apple 1Meg card and add it to my RamKeeper or if I should just perchase a whole new card that takes 1 Meg chips. Kent Dickey kadickey@phoenix.Princeton.EDU
gwyn@smoke.BRL.MIL (Doug Gwyn) (10/09/90)
In article <7619@darkstar.ucsc.edu> unknown@ucscb.UCSC.EDU (The Unknown User) writes: >That's a simplistic calculation, yes, but is it not correct that it's the >LOWEST (that is, FASTEST) value you'd ever need for a certain speed of a CPU? No, in fact it's an upper bound on the chip response time or a lower bound on speed. Because more than one atomic event occurs within a clock cycle, the actual requirements are "stiffer" than the calculation would indicate.