Is it time to stop research in Computer Architecture ?
|
From: nmm1 on 23 Oct 2009 04:17 In article <4AE12FA9.1000706(a)patten-glew.net>, Andy \"Krazy\" Glew <ag-news(a)patten-glew.net> wrote: >Robert Myers wrote: > >I am not aware of an Itanium shipped or proposed that had an "x86 core >on the side". I am. I can't say how far the proposal got - it may never have got beyond the back of the envelope stage, and then got flattened as a project by management. >> That is to say, I find it hard to believe that anyone took Itanium >> seriously as an x86 competitor. > >I can assure you that it was sold that way to Intel senior management. And not just Intel's. It was to most of the Tier 1 vendors and major resellers, too. Your statements aren't in conflict, though. It is hard to believe, but it was true :-) Regards, Nick Maclaren.
From: Anton Ertl on 23 Oct 2009 05:02 Stephen Fuld <SFuld(a)alumni.cmu.edu.invalid> writes: >Anton Ertl wrote: >> Probably not, because for most users Excel is fast enough even with >> slow algorithms. And those where it isn't, they have probably >> invested so much in Excel that most of them would not change to a >> spreadsheet program with better algorithms even if one is available. >> So there is no incentive for other spreadsheet programs to improve >> their algorithms, and therefore also no incentive for Excel. > >I understand that. But isn't say the spreadsheet in Open Office >supposed to be compatible with Excel? Yes. When someone sends me an .xls file, I load it into OpenOffice. I don't know if the OpenOffice spreadsheet is compatible enough for the Excel power users, though (it may or may not be, as an Excel non-user I don't know). >That all makes sense, but I am bothered by what seems two contradictory >ideas here. Andy says the problem is an algorithm Excel uses that is >O(N**2) compared with one he could write himself that is perhaps O(N) or >O(N log N). I presume this algorithm isn't in the parsing but in the >calculation themselves. On the other hand, Paul said the bulk of the >time is probably in the parsing rather than the recalculation of the >parsed results. Not wholy a contradiction: If the implementation uses an O(N^2) algorithm for recalculation and reparses on recalculation, then parsing probably adds a constant factor >2; then the majority of the run-time is spent in parsing; the fact that the algorithm could be optimized by a factor of N or so is mostly orthogonal to that. Actually, it's not quite orthogonal: I think that an implementation that parses at calculation time is unlikely to preprocess the spreadsheet to allow recalculation of the cells in a topologically sorted order. - anton -- M. Anton Ertl Some things have to be seen to be believed anton(a)mips.complang.tuwien.ac.at Most things have to be believed to be seen http://www.complang.tuwien.ac.at/anton/home.html
From: Anton Ertl on 23 Oct 2009 05:37 Robert Myers <rbmyersusa(a)gmail.com> writes: >On Oct 22, 10:42=A0am, an...(a)mips.complang.tuwien.ac.at (Anton Ertl) >wrote: [Code not tuned for IA-64 compiled with gcc] >> The performance per cycle of the Itanium II is not particularly good, >> but also not particularly bad. >> >I'm not sure what you think this proves. It gives an impression of the performance of the Itanium-II compared to others on code not tuned for IA-64 and compiled with gcc. Among other things, it answers your question: |Yes, gcc will run on Itanium, but with what level of performance? And I think it also proves that it is not necessary to rewrite software for IA-64 to get acceptable performance. >I'm assuming that those who used SGI's compiler used it for their own >applications, not for building the system, as I commented elsewhere. With SGI's compiler being free software, I think it would have been used for IA-64 if it was gcc-compatible enough and if the performance benefit it gave was big enough. >Windows Server does run on Itanium. Yes, just as WNT ran on Alpha. If Microsoft really had wanted to run with IA-64, they could have done that by just not supporting AMD64. Then Intel would not have supported AMD64, pardon, Intel 64, either, and AMD64 would likely be as relevant today as 3Dnow; ok there would have been a certain risk involved in that strategy, especially because AMD at the time was in a stronger position than at the introduction of 3Dnow, but given that most Windows and even Linux users still run 32-bit software exclusively, that risk seems to be small. >http://www.microsoft.com/servers/64bit/itanium/overview.mspx > >and it would appear that Microsoft is very happy to have those >customers. My guess is that Microsoft happily takes money from HP, Intel, and/or who else has interest in Itanium, so these companies can market their Itanium systems as Windows-compatible. That has certainly been the story for other non-IA32/AMD64 architectures, e.g., Alpha. - anton -- M. Anton Ertl Some things have to be seen to be believed anton(a)mips.complang.tuwien.ac.at Most things have to be believed to be seen http://www.complang.tuwien.ac.at/anton/home.html
From: Bill Todd on 23 Oct 2009 07:30 nmm1(a)cam.ac.uk wrote: > In article <DcudnTf8XIfmmXzXnZ2dnUVZ_u2dnZ2d(a)metrocastcablevision.com>, > Bill Todd <billtodd(a)metrocast.net> wrote: >> The fact that Itanic came so close to world domination *despite* its >> abject failure to deliver on the promises that had seemed to make that >> domination inevitable tends to prove that the attempt to bluff its way >> to success was a daring and risky move but hardly an insane one. ... > > Not really. It was a lot further from that than the hype indicated. Not really. Consider, for example, just the VMS market (a not-inconsiderable annual $4 billion system market all by itself before the Alphacide). When Alpha got the axe in mid-2001, Compaq mounted a concerted, fairly slick, and thoroughly disgusting effort to present Itanic as an upgrade rather than a regrettable alternative - and the majority of the VMS population seemed quite willing to go along with that view until a small subset of us screamed bloody murder for several years running setting the record straight. Enterprise customers tend to be slaves to their vendors - even more so back then when open source was even less an option than it is today. Being slaves, they seek desperately to believe that their vendors will treat them well rather than screw them royally as Compaq did. And in the absence of strong doses of reality therapy that belief will prevail, simply because the alternative is too uncomfortable to contemplate for the average Joe who just wants things to run smoothly. Had the VMS market remained anywhere nearly as strong as it had been Itanic would have gotten a significant credibility shot in the arm simply by virtue of doubling its market share. That very likely would have happened had the MBAs had their way without opposition. While I wasn't as closely involved with such reality therapy for other platforms (e.g., PA-RISC) I do know anecdotally that at least somewhat similar events unfolded there as well: without the existence of a vocal minority who had no real reason beyond pure outrage to become involved, the sheep would have lined up docilely to be shorn and Itanic would very likely merely have suffered a few years' delay in its quest for world domination. > It made practical headway in two areas, so let's consider them. > > HPC was its most successful area, and something like two sites tried > it and rejected it for every one that delivered a service using it. The importance of HPC to real-world success of a platform is pretty debatable. .... > The other was Mission Critical computers for Big Business. Exactly. I met > people from several of those, and they had all taken the position > that they were going to run it in parallel with their existing > systems for a year or more before making a decision. Which is what they would have done with virtually *any* new system, no matter how wonderful it seemed: these people give new meaning to the word 'cautious'. .... > My point here is that, if the Itanic had started to be pushed much > harder, the real heavyweights would have joined the opposition. What earthly reason do you have to think that? IBM never lifted a finger against Itanic but rather got on board early and shipped product for several years. One might suggest that they were confident that it would sink of its own weight if IBM had not become a supporter (though certainly not an evangelist in the ways that HP and Intel were) at a time when Itanic's perceived future was riding very high indeed. IBM seemed relatively willing to give up its processor business just as several others were if Itanic really could cut the mustard, the main difference being that IBM, unlike they, didn't do so prematurely. > It never had an earthly of doing what it was originally hyped to > do (i.e. entirely replace x86). Save for the grace of AMD it still might have: without a credible, inexpensive, and pervasive 64-bit alternative Intel could have just waited until desktops began to demand 64-bit processors. Without real competition even otherwise ridiculous 'solutions' can become successful (I'm sure you can think of as many examples of this as I could). And when you've got as close to a monopoly on developing solutions as Intel did back in the mid-'90s dividing up and conquering the market in the way that it planned to doesn't take technical excellence, just marginal adequacy. - bill
From: Mayan Moudgill on 23 Oct 2009 07:45
Andy "Krazy" Glew wrote: > E.g. Terje, you're known to be a Larrabee fan. Can you vectorize CABAC? Not a chance. > For example: divide the image up into subblocks, and run CABAC on each > subblock in parallel. Problem is with the standard. H.264 specifies that the frame is CABAC encoded. |