Is it time to stop research in Computer Architecture ?
|
From: Andrew Reilly on 20 Oct 2009 02:40 On Mon, 19 Oct 2009 20:40:39 -0700, Andy \"Krazy\" Glew wrote: > Andrew Reilly wrote: >> Isn't it the case, though, that for most of that "popular software" >> speed is a non-issue? > > I've been manipulating large Excel spreadsheets. well, there's your problem ;-) I've never got the hang of spreadsheets, and never found a problem that didn't look like more of a job for awk or matlab, or even a real program of some sort. I guess that there must be some (or at least users who think differently than I): it's certainly popular. > Minutes-long recalcs. > > Sometimes as long as 15 minutes to open up such a spreadsheet embedded > as an OLE object in a Word document. > > I'm reasonably sure it's computation, and not disk. For small values of "computation", probably: i.e., it's probably wading through vast and convoluted layers of "abstraction", choking on memory latency, rather than running your real "computation". Have you made any estimates of how long your computation job would take when expressed in a more conventional programming language? On the other hand, there clearly are some things that PCs still take a long time to do. (My colleagues recently turned on link-time- optimization on a project that I sometimes compile, and now I have time to make and drink a cup of coffee...) > However, I am also fairly certain, because I have done a few > experiments, that there are a few quadratic algorithms that could be > linear. > > Algorithms trump hardware, nearly every time. Yes, that's true. And one can probably guess that quite a lot of the programs that had their geneses when PCs had floppy disks and no caches also have some corners with hand tuned assembly language kernels to make a simple (poorly scaling) algorithm "fast". How much of that there is in Excel I have no idea, of course. > (Hmm.... in the past I have devised "instruction rewriting" hardware > that converted linear code such as i++;i++;...;i++ into O(lg N) latency, > O(N lg N) size parallel prefix code. I wonder if the problems I am > seeing Excel could be optimized away by a subsystem.) I'm afraid that I don't know enough of Excel to comment. Maybe you'd be able to get a reasonable speed-up on a single core by running the "computation" part of the recalculation process through a compiler: strip away a a few layers of abstraction with some suitable constant propagation. Cheers, -- Andrew
From: nmm1 on 20 Oct 2009 03:25 In article <hbikbc$vr3$1(a)news.eternal-september.org>, Stephen Sprunk <stephen(a)sprunk.org> wrote: >>> >>> Printing in China has been used for about a 1000 years when movable types >>> where invented there - and that invention didn't catch on. It was easy >>> enough to carve the writing into stone or wood and print that way, movable >>> types are only cost effective if you print in low volume. >> >> That's not true. The converse is. Carving a whole block in reverse >> and printing from that is only cost-effective if the number of pages >> printed is small. > >I don't see how the number of original pages is relevant. What matters >is the number of copies of each page, and the more you print, the more >copies you can amortize the typesetting cost across. Yes. I corrected Bernd, and he corrected me back again :-) That is the right calculation. >> You need extremely skilled people to carve them, in order to keep the >> error rate down. > >You'd need skilled artisans for the original molds, just like with Latin >type, but pouring and using each individual piece of type is a >relatively mechanical process that doesn't require nearly as much skill. Precisely. >> There are also problems with movable type and Chinese characters (i.e. >> the number of them!) > >You wouldn't need as many different pieces of type as you'd think; most >of the more complicated characters are composed of various combinations >of simpler sub-characters. Yes, but few of them are simply separable, and making blocks for the sub-characters that fitted together in all the combinations needed. >> Before moveable type, almost all printing was things like pictures, >> prayers, fabric patterns and other uses where the number of different >> pages is small. > >ITYM where the number of original pages was small but the number of >copies was huge. Usually. That's not always been true for etchings and woodblock 'paintings'. Regards, Nick Maclaren.
From: Bernd Paysan on 20 Oct 2009 05:10 kenney(a)cix.compulink.co.uk wrote: > In article <1947351.S7ZIpNfcyU(a)elfi.zetex.de>, bernd.paysan(a)gmx.de > (Bernd Paysan) wrote: > >> Printing in China has been used for about a 1000 years when movable >> types where invented there - > > Movable type really requires an alphabetic script. You are better of > with wood block printing for a written language that has several > thousand characters. The Chinese attempt at movable type was associated > with the Yuan dynasty and a script that Kublai introduced, it failed as > much for political reasons as technical ones. What you describe has nothing to do with history. Look up "Bi Sheng", e.g. on Wikipedia http://en.wikipedia.org/wiki/Bi_Sheng who was the inventor of movable types. His types were made of baked clay, so more fragile than wood. He lived in the Song dynasty, and used the ~3000 characters that were in use back then (today, about 5000-7000 characters are in use). It took several centuries of progress to make movable types competitive with the established woodblock printing, and most of that progress was made in Korea. Note that the woodblock printing principle again has won over movable types in the 20th century, and it has expanded to fields no one could have imagined (from circuit boards to chips, all that is basically derived from woodblock printing technology). > Even given that the major constraint on printing in the West was the > cost of type I think you are wrong. Cheap books required the > introduction of the rotary press and means to convert type to plates > that could be used in them. The rotary press wasn't invented until 1843. Yes, it was again lowering the price of books by another order of magnitude, but affordable books came out of normal printing presses before. Now, printing books is going to end soon, through e-book readers, lowering the price of a book by another order of magnitude. -- Bernd Paysan "If you want it done right, you have to do it yourself" http://www.jwdt.com/~paysan/
From: Noob on 20 Oct 2009 07:20 Terje Mathisen wrote: > The main target of Itanium was to have a "closed-source" cpu that simply > couldn't be cloned at all: > > A new separate (joint venture) company to develop it, making existing > deals with both Intel and HP moot, lots of funky little patented details > that were intentionally exposed to the programmer, making it very much > harder to invent around. > > OTOH, I really do believe Intel intended to start deliver in 1997, in > which case it _would_ have been, by far, the fastest cpu on the planet. > When they finally did deliver, years later, it was still the fastest cpu > for dense fp kernels like SpecFP. > > They delivered too little, too late, but still managed to terminate > several competing architecture development tracks at other vendors. In other words, Itanium is the epitome of FUD. Hail FUDzilla! :-) (RIP Alpha and PA-RISC)
From: nmm1 on 20 Oct 2009 07:50
In article <2813063.ke3PP5UZLK(a)elfi.zetex.de>, Bernd Paysan <bernd.paysan(a)gmx.de> wrote: >kenney(a)cix.compulink.co.uk wrote: > >> Movable type really requires an alphabetic script. You are better of >> with wood block printing for a written language that has several >> thousand characters. The Chinese attempt at movable type was associated >> with the Yuan dynasty and a script that Kublai introduced, it failed as >> much for political reasons as technical ones. > >What you describe has nothing to do with history. Look up "Bi Sheng", e.g. >on Wikipedia > >http://en.wikipedia.org/wiki/Bi_Sheng > >who was the inventor of movable types. His types were made of baked clay, >so more fragile than wood. He lived in the Song dynasty, and used the ~3000 >characters that were in use back then (today, about 5000-7000 characters are >in use). It took several centuries of progress to make movable types >competitive with the established woodblock printing, and most of that >progress was made in Korea. The truth is half-way in between. You need a lot MORE progress to make moveable type viable for ideograms than for alphabets. That is one of the reasons that moveable type spread so fast in the West, but not in the East. >Note that the woodblock printing principle again has won over movable types >in the 20th century, and it has expanded to fields no one could have >imagined (from circuit boards to chips, all that is basically derived from >woodblock printing technology). It never went away. It always was a better principle for pages that aren't feasible to build up out of smaller, replicatable ones, such as images. >Now, printing books is going to end soon, through e-book readers, lowering >the price of a book by another order of magnitude. That's two extreme speculations in one sentence :-) Regards, Nick Maclaren. |