On Saturday, June 14, 2014 12:13:48 PM UTC+1, ghi...@gmail.com wrote: > > > > On Saturday, June 14, 2014 3:31:12 AM UTC+1, Russell Standish wrote: >> >> On Sat, Jun 14, 2014 at 02:22:56PM +1200, LizR wrote: >> > Oh, OK, obviously I was misinformed. I will smack Charles' bottom >> later. >> > >> > >> > On 14 June 2014 14:27, Russell Standish <li...@hpcoders.com.au> wrote: >> > >> > > On Sat, Jun 14, 2014 at 01:52:01PM +1200, LizR wrote: >> > > > >> > > > Moore's law appears to have stopped working about 10 years ago, >> going by >> > > a >> > > > comparison of modern home computers with old ones. That is, the >> > > processors >> > > > haven't increased much in speed, but they have gained more "cores", >> i.e. >> > > > they've been parallelised, and more memory and more storage. But >> the >> > > > density of the components on the chips hasn't increased by the >> predicted >> > > > amount (or so I'm told). >> > > > >> > > >> > > Moore's law was never about GHz. It was originally about number of >> > > transistors per dollar, and with greater transistor counts per CPU, >> that >> > > has >> > > been turned into bigger caches and multiple cores (with 50+ core >> chips >> > > now on the market). >> > > >> > > But of real interest is processing power per dollar as a function of >> > > time. This has been exponential since the start of the computing age >> > > (perhaps even with a reduction of the time constant sometime in the >> > > '90s), and shows no sign of slowing down. The rate of 1 order of >> > > magnitude of performance improvement at a given price point every 5 >> > > years has held throughout my professional life. In my career, the >> > > following purchases were made*: >> > > >> > > 1992 CM5, 4GFlops $1.5M >> > > 1996 SGI Power Challenge, 8GFlops, $800K >> > > 2000 SGI Origin 56 GFlops $1.2M >> > > 2004 Dell cluster, 1TF, $500K >> > > 2013 HP GPU cluster, 300TF, $500K >> > > >> > > * subject to a certain amount uncertainty due to my recall of the >> > > facts >> > > >> > > Attached is an image of the performance per dollar plotted as a >> > > function of year. >> > > >> >> Incidently, the "kink" at 2000 was caused by the move from proprietry >> systems to commodity systems running Linux. I tried to make the 2000 >> purchase a Linux-based purchase, but was unable to convince my >> colleagues. If I'd been successful, the curve would have been a lot >> flatter! >> >> Cheers >> >> -- >> >> ---------------------------------------------------------------------------- >> >> Prof Russell Standish Phone 0425 253119 (mobile) >> > > it's throwaways like this that say the most if they accumulate in time > which they do with you - my window anyway >
p.s. "say" I am 99% sure is obvious, but due to some minor local self-esteem issues and the other local matter of a one-to-many rearguard action, I shall have to cave in and add "in a very positive direction" -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to everything-list+unsubscr...@googlegroups.com. To post to this group, send email to everything-list@googlegroups.com. Visit this group at http://groups.google.com/group/everything-list. For more options, visit https://groups.google.com/d/optout.