Home |
Search |
Today's Posts |
|
Metalworking (rec.crafts.metalworking) Discuss various aspects of working with metal, such as machining, welding, metal joining, screwing, casting, hardening/tempering, blacksmithing/forging, spinning and hammer work, sheet metal work. |
Reply |
|
|
LinkBack | Thread Tools | Display Modes |
#1
|
|||
|
|||
Need a computer system for your nuclear program?
http://cgi.ebay.com/ws/eBayISAPI.dll...5579 961&rd=1
"If I'm going to reach out to the the Democrats then I need a third hand.There's no way I'm letting go of my wallet or my gun while they're around." "Democrat. In the dictionary it's right after demobilize and right before demode` (out of fashion). -Buddy Jordan 2001 |
#2
|
|||
|
|||
Most excellent!!!
http://www.spikynorman.dsl.pipex.com...NCE.950313.txt According to that article, the Cray J932 can contain up to 32 processors and run 6.4 gflops. That's a whole lotta power for the $99 opening bid... assuming the system is in complete working condition. If it's not, good luck getting parts! --- Cray/Media: Steve Conway, 612/683-7133 Mardi Larson, 612/683-3538 Cray/Financial: Brad Allen, 612/683-7395 CRAY RESEARCH DOUBLES TOP PERFORMANCE OF FAST- SELLING CRAY J90 PRODUCT LINE New CRAY J932 UNIX-Based Systems Scale Low-Cost Supercomputer Line Up To 32 Processors, 6.4 Billion Calculations Per Second DALLAS, March 13, 1995 -- At the UniForum 95 open computing conference, Cray Research, Inc. (NYSE: CYR) announced today its new CRAY J932 supercomputer that doubles the maximum performance of the company's fastest- selling product line ever -- the CRAY J90 series of low-cost, UNIX-based supercomputer systems. The new CRAY J932 products are air-cooled, compact supercomputer systems with 16 to 32 processors that provide up to 6.4 billion calculations per second (gigaflops) of peak computing speed -- twice the maximum offered in the previously announced, binary-compatible CRAY J916 systems, Cray officials said. The CRAY J932 systems, slated to begin shipping in second-quarter of this year, are priced in the U.S. from under $1 million to about $2.6 million. Cray reported it has already received 11 advance orders for the new CRAY J932 systems. The orders are from British Gas, London, England; Howard University, Washington D.C.; Penn State University, University Park, Penn.; University Groningen, The Netherlands; and two German organizations, Konrad-Zuse- Zentrum fuer Informationstechnik Berlin (Konrad-Zuse Center for Information Technology Berlin), and the Christian- Albrechts-Universitaet zu Kiel (University of Kiel). An additional undisclosed customer has ordered five CRAY J932 systems that will be clustered using Cray Research's SuperCluster software environment. Terms were not disclosed. The CRAY J90 Series is aimed at small- to mid-sized organizations, or departments within larger organizations, that do scientific and technical simulation work in the government, industrial and university sectors. These Cray systems are designed to operate as simulation servers for large, complex problems that exhaust the capabilities of workstations and workstation-based server products. A single CRAY J932 system's large, real central memory -- up to eight gigabytes -- can handle heavy workloads that would exhaust the cache memory and memory bandwidths of uniprocessor workstations, Cray said. "Our J90 customers have anywhere from a few users to more than a hundred running a wide range of third-party and proprietary applications," said Robert H. Ewald president and chief operating officer of Cray Research. "Cray Research's complete catalogue of more than 600 third-party supercomputer applications is available on these systems, compared with far fewer application offerings on competing systems. The CRAY J90 systems are the most general-purpose supercomputers in their price class." The company has captured about 100 orders for the CRAY J916 supercomputer, the smaller system in the series that was announced in September 1994, Ewald said. The CRAY J916 systems begin volume shipments this month, at U.S. list prices ranging from $225,000 to $1.5 million for systems with four to 16 processors and up to 3.2 gigaflops of peak performance. He said nearly 40 percent of the 110 CRAY J90 Series orders are from new customers. "With the volume of CRAY J90 Series orders already in hand, we are well on our way to, in 1995, substantially increasing Cray Research's share in the growing open systems market for high-performance simulation servers," Ewald said. According to Gary Smaby, president of Smaby Group, Inc., a high-performance computing (HPC) analyst firm in Minneapolis, "Early order activity leads us to believe that with the J90 Series, Cray has finally hit on a compelling combination of price, performance and application breadth to address the sub- $2 million HPC server market. While Cray historically has dominated the market for high-end enterprise supercomputers, the company has not had a significant presence in this price tier. We expect the J90 Series to finally put Cray on the map in this billion dollar, 30+ percent growth market tier, propelling the company from a mere three percent market share in 1994 to a respectable 11 percent by year end. "In our view, the catalyst will be the ability of CRAY J90 Series users to gain plug and play' access to Cray's impressive third-party applications library -- including the fifty or so key codes which drive the majority of the HPC systems sales in this price range." Ewald said that the CRAY J90 Series systems are scaled down versions of Cray Research's proven supercomputing technology combined with the cost advantages of CMOS. To maintain the product line's unrivaled performance on real-world applications -- which average 60 to 70 percent of peak processor speed -- Cray has doubled the maximum memory bandwidth over the CRAY J916 system to 51.2 billion bytes (gigabytes) per second in the CRAY J932 system. This memory bandwidth is more than 20 times faster than similarly priced competing products and maintains Cray's leadership in high- bandwidth computing, Ewald said. The new systems are based on custom RISC microprocessors that Cray designed and has manufactured by an undisclosed outside vendor, Ewald said. The microprocessor is two CMOS ASICs (Application Specific Integrated Circuit) with approximately 800,000 cells per chip. The microprocessor is an innovative design that reduces the Cray Research central processing unit (CPU) -- previously consisting of hundreds of chips and multiple multi-layered printed circuit boards -- to two powerful processor chips and a handful of supporting chips, he said. "All Cray supercomputers to date have been based on RISC architecture. In fact, in the world of supercomputing, the Cray RISC architecture is the most commonly used," Ewald said. "By applying the price/performance advantages of CMOS to our proven architecture, our new J90 line features RISC microprocessors tuned specifically for hundreds of supercomputing applications." Ewald said the CRAY J932 supercomputers: - Can be installed in an office environment with standard air conditioning. - Offer sustained multiple gigaflops on a wide range of applications for under $1.5 million. - Are compatible with workstations and designed to operate in a client/server environment. As a central simulation server, the CRAY J932 supercomputer is an open system and easily connects to workstation devices made by a variety of suppliers. It also supports the most popular workstation data formats with transparent, automatic data conversion. - Run the same standard UNIX operating system (UNICOS) as larger Cray systems, with multi-user features and other enhancements for the demanding high-end computing market. - Have unmatched, flexible scaling features. Scalability of competing low-cost supercomputers is limited by adding another processor or workstation to the cluster. Many of these products rely on the low speeds of ethernet as their "system" interconnect and offer limited scalability, as they are confined to low-end configurations when total system performance is considered. With the CRAY J932 system's balanced design and leading bandwidth, customers can easily and inexpensively add processors at $30,000 each (U.S. list price) in a single system chassis and gain significantly more from the addition of the processors. - Are based on Cray Research supercomputer technology for industry-leading "delivered" price/performance. Because of the system's leading bandwidth of 51.2 gigabytes per second, the sustained-to-peak ratio for this system on a wide range of application far exceeds the competition. Per processor the system achieves from 60-70 percent of its peak speed of 200 megaflops (million floating-point operations per second). - Can be clustered with Cray's SuperCluster software environment, providing virtually unlimited, versatile capacity upgrades and total peak capabilities as high as 200 gigaflops. With this Cray software, customers can link multiple supercomputers for a tightly integrated distributed memory environment to tackle even larger workloads. Cray today also announced that it has received Advanced Systems Magazine's 1994 Best Products Award for the previous-generation CRAY EL94 system that was introduced last year. Cray officials will be given this award at a luncheon ceremony held here at the UniForum show tomorrow. Cray Research provides the leading supercomputing tools and services to help solve customers' most challenging problems. ### Editor's Note: A CRAY J916 system is being showcased in Cray's booth (#2125) at the UniForum conference and exhibition held here. |
#4
|
|||
|
|||
Lawrence Glickman wrote in message ... On 3 Dec 2004 06:47:49 -0800, (Miki Kanazawa) wrote: Most excellent!!! http://www.spikynorman.dsl.pipex.com...oduct/J932_ANN OUNCE.950313.txt According to that article, the Cray J932 can contain up to 32 processors and run 6.4 gflops. That's a whole lotta power for the $99 opening bid... assuming the system is in complete working condition. If it's not, good luck getting parts! University lab might be the market. I can't imagine running anything my 2.5 giger can't handle. It gets the job done. LOL - the Cray could do jobs in a few hours that your toy computer would not be able to complete in the rest of your lifetime Dale In the meantime, good luck paying the electric bill for that *thing.* Lg misc.survivalism |
#5
|
|||
|
|||
On Fri, 03 Dec 2004 19:40:10 GMT, "dalecue"
wrote: Lawrence Glickman wrote in message ... On 3 Dec 2004 06:47:49 -0800, (Miki Kanazawa) wrote: Most excellent!!! http://www.spikynorman.dsl.pipex.com...oduct/J932_ANN OUNCE.950313.txt According to that article, the Cray J932 can contain up to 32 processors and run 6.4 gflops. That's a whole lotta power for the $99 opening bid... assuming the system is in complete working condition. If it's not, good luck getting parts! University lab might be the market. I can't imagine running anything my 2.5 giger can't handle. It gets the job done. LOL - the Cray could do jobs in a few hours that your toy computer would not be able to complete in the rest of your lifetime I don't need that kind of power for any applications that I can think of. I'm not modeling a virtual nuclear detonation, for example. But I _can_ keep track of every satellite in orbit that is still up there, in Real Time. You need more than that? Buy it. 2.5 gigs is screaming lightening for most *consumer* applications. Furthermore, the fact that IBM is getting out of the PC market should tell you something. Compaq even folded, and was bought out by HP. Get a clue. What people want these days is connectivity, to all the other computers *out there.* So the ceiling on CPU processing speed is a function now of _broadband connectivity_, not isolated Mips. Lg |
#6
|
|||
|
|||
Keywords:
In article , "dalecue" wrote: Lawrence Glickman wrote in message ... On 3 Dec 2004 06:47:49 -0800, (Miki Kanazawa) wrote: Most excellent!!! http://www.spikynorman.dsl.pipex.com...oduct/J932_ANN OUNCE.950313.txt According to that article, the Cray J932 can contain up to 32 processors and run 6.4 gflops. That's a whole lotta power for the $99 opening bid... assuming the system is in complete working condition. If it's not, good luck getting parts! University lab might be the market. I can't imagine running anything my 2.5 giger can't handle. It gets the job done. LOL - the Cray could do jobs in a few hours that your toy computer would not be able to complete in the rest of your lifetime Don't be so sure. The museum at Los Alamos has a Cray 1, sitting next to a SPARC 10, which has roughly the same computing horsepower.... Doug White |
#7
|
|||
|
|||
In article , Lawrence Glickman
says... Furthermore, the fact that IBM is getting out of the PC market should tell you something. Compaq even folded, and was bought out by HP. Get a clue. Gigaflops? What were those? g http://en.wikipedia.org/wiki/Blue_Gene Were they like "albums?" Jim -- ================================================== please reply to: JRR(zero) at pkmfgvm4 (dot) vnet (dot) ibm (dot) com ================================================== |
#8
|
|||
|
|||
Gunner wrote:
http://cgi.ebay.com/ws/eBayISAPI.dll...5579 961&rd=1 We get that much power on a single chip now. http://www.eet.com/semi/news/showArt...cleId=51200997 -- Samiam is Scott A. Moore Personal web site: http:/www.moorecad.com/scott My electronics engineering consulting site: http://www.moorecad.com ISO 7185 Standard Pascal web site: http://www.moorecad.com/standardpascal Classic Basic Games web site: http://www.moorecad.com/classicbasic The IP Pascal web site, a high performance, highly portable ISO 7185 Pascal compiler system: http://www.moorecad.com/ippas Good does not always win. But good is more patient. |
#9
|
|||
|
|||
In rec.crafts.metalworking Lawrence Glickman wrote:
On 3 Dec 2004 06:47:49 -0800, (Miki Kanazawa) wrote: Most excellent!!! http://www.spikynorman.dsl.pipex.com...NCE.950313.txt According to that article, the Cray J932 can contain up to 32 processors and run 6.4 gflops. That's a whole lotta power for the $99 opening bid... assuming the system is in complete working condition. If it's not, good luck getting parts! University lab might be the market. I can't imagine running anything my 2.5 giger can't handle. It gets the job done. A 2.5Ghz pentium may only be a couple of times faster than a P100, for some tasks. If it requires random access to large (greater than the on-chip cache) amounts of memory, random access speed of chips hasn't really changed in the past decade. Some 'supercomputer' type computers took extreme measures to get round this. |
#10
|
|||
|
|||
On 04 Dec 2004 17:10:21 GMT, Ian Stirling
wrote: I can't imagine running anything my 2.5 giger can't handle. It gets the job done. A 2.5Ghz pentium may only be a couple of times faster than a P100, for some tasks. If it requires random access to large (greater than the on-chip cache) amounts of memory, random access speed of chips hasn't really changed in the past decade. Some 'supercomputer' type computers took extreme measures to get round this. OK, I put in 512MB of RAM and this thing is working as advertised. Less than 512 is not a good idea, because I can see from my stats that I only have 170MB of unused physical memory at the moment. Once you get enough RAM so you don't have to do HD read/writes, you're on your way to speed. But M$ has shafted everybody, because M$ loves huge swap files. The more RAM you have the less you will need to read/write to this swap file, and the more smoothly things seem to run. For my own purposes, this is all the computer I need. I suppose if you had to, you could put up to 1 gig of RAM into this thing, but from what I can see, that would be a waste of money. It wouldn't be used. Lg |
#11
|
|||
|
|||
In rec.crafts.metalworking Lawrence Glickman wrote:
On 04 Dec 2004 17:10:21 GMT, Ian Stirling wrote: I can't imagine running anything my 2.5 giger can't handle. It gets the job done. A 2.5Ghz pentium may only be a couple of times faster than a P100, for some tasks. If it requires random access to large (greater than the on-chip cache) amounts of memory, random access speed of chips hasn't really changed in the past decade. Some 'supercomputer' type computers took extreme measures to get round this. OK, I put in 512MB of RAM and this thing is working as advertised. Less than 512 is not a good idea, because I can see from my stats that I only have 170MB of unused physical memory at the moment. Once you get enough RAM so you don't have to do HD read/writes, you're on your way to speed. But M$ has shafted everybody, because M$ loves Not for all tasks. For a few scientific computing tasks (nuclear simulations are a big one), the big bottleneck is how long it takes you to read a dozen bytes from a 'random' location in memory. All the information is in memory - it's just that the memory isn't fast enough. The increase in random access speed has not nearly matched the increase in streaming speed. IIRC, a 20 year old 30 pin 1Mb SIMM would typically take 150ns (billionths of a second) for the first byte, then 100ns for the next one. A modern 1Gb memory DIMM might read out the second (and subsequent) bytes a hundred times faster than the 1Mb SIMM. But the initial access to any location in that DIMM is still only around at best 3-4 times faster. In some tasks, this can be the limiting factor for calculation. |
#12
|
|||
|
|||
In article , Lawrence
Glickman wrote: On 04 Dec 2004 17:10:21 GMT, Ian Stirling wrote: I can't imagine running anything my 2.5 giger can't handle. It gets the job done. A 2.5Ghz pentium may only be a couple of times faster than a P100, for some tasks. If it requires random access to large (greater than the on-chip cache) amounts of memory, random access speed of chips hasn't really changed in the past decade. Some 'supercomputer' type computers took extreme measures to get round this. OK, I put in 512MB of RAM and this thing is working as advertised. Less than 512 is not a good idea, because I can see from my stats that I only have 170MB of unused physical memory at the moment. Once you get enough RAM so you don't have to do HD read/writes, you're on your way to speed. But M$ has shafted everybody, because M$ loves huge swap files. The more RAM you have the less you will need to read/write to this swap file, and the more smoothly things seem to run. For my own purposes, this is all the computer I need. I suppose if you had to, you could put up to 1 gig of RAM into this thing, but from what I can see, that would be a waste of money. It wouldn't be used. The issue isn't about the hard drive...I think we all take it as obvious that if hard disk accesses are frequent, performance suffers. The issue is that there at least 3 different speed regimes for the solid-state memory: -- relatively slow dynamic RAM (DRAM) installed as "main memory"--this is what people talk about when they say they have 512 MB of RAM. Access speeds are much slower than processor cycle speeds. -- static RAM (SRAM), which is several times faster than DRAM because of the way the circuitry works (cross-coupled inverters rather than bit lines and storage in depletion regions). (The reason DRAMs and SRAMs have their niches is because DRAMs can be built with only one transistor per memory cell, whereas SRAMs tend to take 4 transistors per memory cell--the canonical design of cross-couple inverters making a latch takes 4 transistors and 2 resistors. Various clever designs keep reducing the sizes, but never down to what a single-transistor DRAM has been at for nearly 30 years.) The SRAM may be in a separate module, either next to the processor, or even in the same package. Or some of it may be integrated onto the chip, as various levels of cache. The processor looks for references in cache (via things like "TLBs," standing for "translation lookaside buffer"). If it finds what it needs in cache, especially on-chip cache (or even faster, on-chip registers, the fastest access of all), then memory access can happen in just one or a couple of processor cycles. If what it needs is NOT in cache, a "cache miss" has occurred, and the processor gets a block of memory from DRAM and puts it in cache. The analogy is like this: Imagine you have several pieces of things you need--papers, pens, stationery, etc. Some are right on your desk, where they can be accessed immediately. Some are stored in drawers or file cabinets, where they can be gotten to quickly, but not as quickly as the "cached" items. And some are stored in other rooms. So you need a pen. You can't one. You suffer a cache miss, and your work stalls for a while. You root around in your desk and find one. But since you have already paid the price of stalling and going into your desk, you might as well "refill the cache" with several pens, and maybe a stapler, etc. (But you don't want to completely flush your old cache, as you may need some of its items soon. So you only partly flush the cache and replace it with stuff from your desk drawer. Stategies for what to flush depend on things like "oldest gets flushed" and "least recently used" and even some metrics coming from what you expect to be working on.) Worse is having to go out to a storage box in the garage for something. This is like accessing main memory. And even worse is having to drive over to Office Depot or Staples for something...like staples. This is like loading from a hard disk, with access time thousands of times slower than main memory. And even slower is having to access from a tape drive (rare for most home users these days) or even floppies stored in a box somewhere. Hundreds of millions of times slower than memory accesses. And because the cost of this access is so high, you don't just get the actual staples you'll need, you get more. You "swap from disk into main memory" and, when you resume work, you put some of those staples in a stapler kept on your desk....you have swapped into main memory and then loaded cache. Look on your processor and you'll references to how much "L1" and/or "L2" (levels, a measure of "closeness" to the processor core) cache it has. One of the biggest differences between consumer-grade CPUs and server-grade CPUs, like the difference between a Celeron and a Xeon--is the amount of cache. Some processors have up to 8 MB of on-chip cache, assuring higher "hit rates" on the cache. Supercomputers use various strategies for boosting performance. Lots of fast cache memory is one of them. Another is to have lots of processors. The current top-ranked supercomputers use lots of cache (one of them uses the Itanium-2, with, IIRC, 4 MB of cache per processor). One of the interesting trends is to emphasize the memory over the processor and, instead of attaching memory to processors, consider the processors to be embedded in a sea of memory. This is the "processor in memory" (PIM) approach. And it's used in the IBM "Blue Gene" supercomputer, currently the fastest supercomputer in the world (the aforementioned Itanium-2-based machine is the second fastest). --Tim May |
#13
|
|||
|
|||
On 04 Dec 2004 19:00:43 GMT, Ian Stirling
wrote: In rec.crafts.metalworking Lawrence Glickman wrote: On 04 Dec 2004 17:10:21 GMT, Ian Stirling wrote: I can't imagine running anything my 2.5 giger can't handle. It gets the job done. A 2.5Ghz pentium may only be a couple of times faster than a P100, for some tasks. If it requires random access to large (greater than the on-chip cache) amounts of memory, random access speed of chips hasn't really changed in the past decade. Some 'supercomputer' type computers took extreme measures to get round this. OK, I put in 512MB of RAM and this thing is working as advertised. Less than 512 is not a good idea, because I can see from my stats that I only have 170MB of unused physical memory at the moment. Once you get enough RAM so you don't have to do HD read/writes, you're on your way to speed. But M$ has shafted everybody, because M$ loves Not for all tasks. For a few scientific computing tasks (nuclear simulations are a big one), the big bottleneck is how long it takes you to read a dozen bytes from a 'random' location in memory. All the information is in memory - it's just that the memory isn't fast enough. The increase in random access speed has not nearly matched the increase in streaming speed. IIRC, a 20 year old 30 pin 1Mb SIMM would typically take 150ns (billionths of a second) for the first byte, then 100ns for the next one. A modern 1Gb memory DIMM might read out the second (and subsequent) bytes a hundred times faster than the 1Mb SIMM. But the initial access to any location in that DIMM is still only around at best 3-4 times faster. In some tasks, this can be the limiting factor for calculation. No calculation my computer is ever going to need to make. We're talking apples and oranges here. I'm talking about getting a motorcycle up to 200 mph, you're talking about taking a truck up to 200 mph. Yes, they are going to require different engines. For my purposes ( motorcycle ), the load my CPU has to deal with, things appear normal in human time. I am not simulating nuclear detonations in virtual reality. Any nanosecond speed differences are not perceptible to this human unit. I just performed 170! ( factorial of 170 ). That's the limit on my *engine* I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* ??????? Lg |
#14
|
|||
|
|||
In article , Lawrence
Glickman wrote: I just performed 170! ( factorial of 170 ). That's the limit on my *engine* I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* ??????? I despair. A good example of where many consumer PCs are running out of gas is in rendering DVDs for burning: this is the equivalent of "ripping" a CD with music. A friend of mine, who is tech-savvy, is finding a DVD render is taking about 11 hours on his 1 GHz PC. This is with both Pinnacle's software and with Adobe's (which he returned when the claimed "faster than the competition" turned out not to be true, at least not in his application). A 2 GHz machine might cut this time approximately in half. This is a serious rate limiter for many applications, and illustrates why people who are using their PCs for video editing, music work, scene rendering, Photoshop, compilation of programs, etc., are buying the fastest machines they can buy, such as dual-3.4 GHz Pentium 4-based machines, loaded with RAM. --Tim May |
#15
|
|||
|
|||
On Sat, 04 Dec 2004 11:53:14 -0800, Tim May
wrote: In article , Lawrence Glickman wrote: I just performed 170! ( factorial of 170 ). That's the limit on my *engine* I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* ??????? I despair. Don't despair. A good example of where many consumer PCs are running out of gas is in rendering DVDs for burning: this is the equivalent of "ripping" a CD with music. A friend of mine, who is tech-savvy, is finding a DVD render is taking about 11 hours on his 1 GHz PC. This is with both Pinnacle's software and with Adobe's (which he returned when the claimed "faster than the competition" turned out not to be true, at least not in his application). This is pure manure, unlike your other post which I have saved to my archives because of its perfection. DVD Decrypter will make an *image* of a DVD on my computer in about 15 to 20 minutes MAX. I can play that back and not know the difference between the original and the one I wrote to disc at 36 frames / second. OTOH, if I am going to BURN a DVD, well then, that is another story. But, both DVD Decrypter and/or DVD Shrink makes short work out of DVD ripping on my 2.5 gig Celeron, you can believe it. If it is taking 11 hours for your *tech savy* friend to rip a DVD, I don't want to be near this guy. He is doing something terribly wrong. BTW, the cost of the software I mentioned was $0.00 US. Lg A 2 GHz machine might cut this time approximately in half. This is a serious rate limiter for many applications, and illustrates why people who are using their PCs for video editing, music work, scene rendering, Photoshop, compilation of programs, etc., are buying the fastest machines they can buy, such as dual-3.4 GHz Pentium 4-based machines, loaded with RAM. --Tim May |
#16
|
|||
|
|||
In article , Lawrence
Glickman wrote: On Sat, 04 Dec 2004 11:53:14 -0800, Tim May wrote: A good example of where many consumer PCs are running out of gas is in rendering DVDs for burning: this is the equivalent of "ripping" a CD with music. A friend of mine, who is tech-savvy, is finding a DVD render is taking about 11 hours on his 1 GHz PC. This is with both Pinnacle's software and with Adobe's (which he returned when the claimed "faster than the competition" turned out not to be true, at least not in his application). This is pure manure, unlike your other post which I have saved to my archives because of its perfection. DVD Decrypter will make an *image* of a DVD on my computer in about 15 to 20 minutes MAX. I can play that back and not know the difference between the original and the one I wrote to disc at 36 frames / second. OTOH, if I am going to BURN a DVD, well then, that is another story. But, both DVD Decrypter and/or DVD Shrink makes short work out of DVD ripping on my 2.5 gig Celeron, you can believe it. If it is taking 11 hours for your *tech savy* friend to rip a DVD, I don't want to be near this guy. He is doing something terribly wrong. I said _rendering_, as from video tapes: "is finding a DVD render is taking about 11 hours on his 1 GHz PC." I didn't say "make an image of," as in doing a bit-copy. Taking 13 GB (for example) of data from a video tape and making it fit on a 4.7 GB DVD requires processing via a codec. You can educate yourself on this at many sites, including this introductory article: http://www.pcworld.com/howto/article...7,pg,11,00.asp --Tim May |
#17
|
|||
|
|||
In article ,
Martin H. Eastburn wrote: Tim Williams wrote: [ ... ] Should use it to draw these in real time. http://webpages.charter.net/dawill/Images/Mbrot20.gif (With my QBasic program, it took about half an hour to draw the 1281 x 801 RAW format file. D'oh!) [ ... ] I remember doing 100! or factorial 100 in full integer precision printout. The first time 11 1/2 hours then 3 hours. First in Advanced Disk Basic, then Machine language (not assembly). Oh the processor 8080. Does the advanced disk basic support that many digits of integer precision? Or were you processing it in slices of some form? 69! is the maximum for the HP 15C calculator, and that can only be represented in scientific or engineering notation, not in integer format. Scientific notation gives: 1.71122452E98 However, the unix DC command does it rather quickly -- even on my slowest system. -- In order of system speed. (The important time values at the end of each are the ones ending in 'u' (user space time) and those ending in 's' (system space time -- the math part of things.) Sorry that I don't have my old 8MHz v7 unix system still up and running. That might even take enough time to count. :-) But -- it would take most of a day to excavate it, and set it up. It is stored broken down into components. SPARC-10 (quad 35 MHz CPUs), running SunOs 4.1.4 (Solaris 1.2) ================================================== ==================== popocat:csu 15:35 # time dc /tmp/trial 93326215443944152681699238856266700490715968264381 62146859296389521759\ 99932299156089414639761565182862536979208272237582 51185210916864000000\ 000000000000000000 0.10u 0.16s 0:00.00 0.0% ================================================== ==================== Total 0.26 second SS-5 (110 MHz CPU) running OpenBSD 3.4 ================================================== ==================== popocat-2:csu 20:14 # time dc /tmp/trial 93326215443944152681699238856266700490715968264381 6214685929638952175\ 99993229915608941463976156518286253697920827223758 2511852109168640000\ 00000000000000000000 0.050u 0.040s 0:00.09 100.0% 0+0k 2+0io 7pf+0w ================================================== ==================== Total 0.090 second SS-5 (170 MHZ CPU) running Solaris 2.6, ================================================== ==================== izalco:dnichols 15:32 time dc /tmp/trial 93326215443944152681699238856266700490715968264381 621468592963895217\ 59999322991560894146397615651828625369792082722375 825118521091686400\ 0000000000000000000000 0.02u 0.05s 0:00.09 77.7% ================================================== ==================== Total 0.070 second SPARC Ultra-2, Dual 400 MHz CPUs running Solaris 9 ================================================== ==================== Hendrix:csu 0:54 # time dc /tmp/trial 93326215443944152681699238856266700490715968264381 621468592963895217\ 59999322991560894146397615651828625369792082722375 825118521091686400\ 0000000000000000000000 0.00u 0.02s 0:00.04 50.0% ================================================== ==================== Total 0.02 second Celeron 2.3 GHz running OpenBSD 3.5 ================================================== ==================== curlmakr:csu 11:33 # time dc /tmp/trial 93326215443944152681699238856266700490715968264381 6214685929638952175\ 99993229915608941463976156518286253697920827223758 2511852109168640000\ 00000000000000000000 0.000u 0.007s 0:00.00 0.0% 0+0k 2+1io 6pf+0w ================================================== ==================== Anyone feel like comparing the restults to see if any one was wrong? :-) The "program" was: ================================================== ==================== 100 99 * 98 * 97 * 96 * 95 * 94 * 93 * 92 * 91 * 90 * 89 * 88 * 87 * 86 * 85 * 84 * 83 * 82 * 81 * 80 * 79 * 78 * 77 * 76 * 75 * 74 * 73 * 72 * 71 * 70 * 69 * 68 * 67 * 66 * 65 * 64 * 63 * 62 * 61 * 60 * 59 * 58 * 57 * 56 * 55 * 54 * 53 * 52 * 51 * 50 * 49 * 48 * 47 * 46 * 45 * 44 * 43 * 42 * 41 * 40 * 39 * 38 * 37 * 36 * 35 * 34 * 33 * 32 * 31 * 30 * 29 * 28 * 27 * 26 * 25 * 24 * 23 * 22 * 21 * 20 * 19 * 18 * 17 * 16 * 15 * 14 * 13 * 12 * 11 * 10 * 09 * 08 * 07 * 06 * 05 * 04 * 03 * 02 * 01 * p q ================================================== ==================== if anyone cares. There may be a way to let it automatically decrement on its own, but this was quicker to set up. :-) It is a reverse Polish notation, so two numbers are entered into the stack, then they are multiplied, another entered, and multipled, etc. The 'p' is "print the results", and the 'q' is "quit now". This was in 1975 by myself. And this was in 2004 (with systems dating back to perhaps 1998 or so. :-) Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#18
|
|||
|
|||
In article ,
Spehro Pefhany wrote: On Sat, 04 Dec 2004 05:15:10 GMT, the renowned "Martin H. Eastburn" wrote: [ ... ] I remember doing 100! or factorial 100 in full integer precision printout. The first time 11 1/2 hours then 3 hours. First in Advanced Disk Basic, then Machine language (not assembly). Oh the processor 8080. This was in 1975 by myself. Martin Barely noticable (in human terms) time on a modern machine: 9332621544394410218832560610857526724094425485496 0571509166910400 4079950642429371486326940304505128980429892969444 7489825873720431 1236641477561877016501813248. Something is wrong there. The number of digits is right 158 characters but a sanity check says that there should be a *lot* of trailing zeros. After all, it has as factors 100 90 80 70 60 50 40 30 20 and 10, not counting other factors which add up to more zeros, such as 2 and 5. I get: 93326215443944152681699238856266700490715968264381 621468592963895217\ 59999322991560894146397615651828625369792082722375 825118521091686400\ 0000000000000000000000 twenty-four trailing zeros. It looks as though your program (whatever it was) used some form of extended precision floating-point math, with some kind of conversion errors. 9332621544394410 (yours) 9332621544394415 (mine) so -- we start to differ in the 16th digit. Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#19
|
|||
|
|||
In article ,
Lawrence Glickman wrote: On 04 Dec 2004 19:00:43 GMT, Ian Stirling wrote: In rec.crafts.metalworking Lawrence Glickman wrote: On 04 Dec 2004 17:10:21 GMT, Ian Stirling wrote: [ ... ] In some tasks, this can be the limiting factor for calculation. No calculation my computer is ever going to need to make. We're talking apples and oranges here. I'm talking about getting a motorcycle up to 200 mph, you're talking about taking a truck up to 200 mph. Yes, they are going to require different engines. If you do image processing, you will need the big RAM, or a bit swap file. The image processing programs like to have the *whole* image in RAM -- even if it has to use swap space to do it. :-) For my purposes ( motorcycle ), the load my CPU has to deal with, things appear normal in human time. I am not simulating nuclear detonations in virtual reality. Any nanosecond speed differences are not perceptible to this human unit. I just performed 170! ( factorial of 170 ). That's the limit on my *engine* I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* My SS-5 (a mere 170 MHz CPU speed) does that one at: ================================================== ==================== izalco:dnichols 20:09 time dc /tmp/trial-170 72574156153079989673967282111292631147169916812964 513765435777989005\ 61843401706157852350749242617459511490991237838520 776666022565442753\ 02532890077320751090240043028005829560396661259965 825710439855829425\ 75689663134396122625710949468067112055688804571933 402126614528000000\ 00000000000000000000000000000000000 0.05u 0.05s 0:00.16 62.5% ================================================== ==================== A total of 0.1 second between user and system time, and 0:00.16 seconds wall clock time (with the system doing other things at the same time.) 307 digits, which makes your 3.2574E+306 pretty close, but it is missing a lot of digits for true precision. Do you have a way to do it in multi-precision integer math? Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#21
|
|||
|
|||
In article , Lawrence
Glickman wrote: For rendering analog video into digital video, as TM suggested, I haven't explored that yet. Not sure I have enough interest in doing so to justify the investment in time and money. Maybe DVD is more permanent than tape, but it isn't forever. Delamination is a problem, along with the growth of anaerobic organisms that attack the media. I think you will be lucky to get 5 years out of DVD even if it is stored in a controlled environment. The stuff isn't as archival as people would lead you to believe. I have DVDs older than this, with no delamination or "rust." I usually put critical files onto two different brands of DVD, with the assumption that while one may have problems, the chances of both failing in the first N years are low (pace the Poisson distribution). My friend makes DVDs of videos for two reasons: 1. Randomly-accessible viewing, with chapters, titles, etc. A lot faster than fast-forwarding through a tape. (Even more so with DV, as few people have DV decks, just DV camcorders. The wear and tear on a DV camcorder used as a playback deck is an excellent reason to make DVDs! And store the DV as an archival "safe" version.) 2. Mainly he does the rendering (the overnight rendering I mentioned) so that he has the source he can quickly burn onto a DVD+R (or -R, his does both) for a parent of one of the kids in the baseball and football programs he makes videos of. The cost of a blank DVD+R is about 50 cents, a VHS tape is more. And even more importantly, a DVD+R can be burned in about 15-30 minutes, depending on length and burn speed, but a VHS tape take 1-2 hours, depending on the length of the source.\ As I said, and as the article I referenced said, rendering either home-edited video or non-digital-sourced video into a form that can be written to DVD is notoriously time-consuming. Many sites say "Your 15-minute home video may take hours to render, even on a fast PC." Which was my point. Your point that you can do 130 factorial, or whatever, faster than you need to, is not the point. Nor is it that you now have a 2.5 GHz Celeron. My point was to refute your claim that fast PCs are not needed for everyday use. They are. As for my friend, he also worked at Intel, and he has 5 or 6 PCs. That he is rendering on one of his 1 GHz machines is beside the point. The point is that rendering is exorbitantly CPU-intensive, whether it takes 11 hours or 5 hours or 2 hours or even 10 minutes (for some professional folks, even this is too slow, for obvious time value of money reasons). (And, yeah, I can do 130!, in Lisp, Scheme, Mathematica, or Haskell. In full bignum precision, needless to say. Both tail-recursion optimizing for a recursive version, or, of course, the brute force 130 * 120 * 128 * 127... version. Mostly I now favor Haskell. See the Net for details on why.) Personally, I dump my DV videos directly through a Firewire port into my Philips standalone deck burner, whose codec is set up for real time streaming of digital data. This means no fancy editing of titles, jump cuts, or even deletions, but it creates an archival copy of a 1-hour DV tape in 1 hour. No muss or fuss. It's a DVD+R and DVD+RW deck. My iMac has a DVD-R and DVD-RW burner, which is suitable for writing content after editing or digitizing. I never use it, as I haven't used the (admittedly sophisticated) features of iMovie (or its pro-grade versions like FinalCutPro) to edit my home movies, insert fancy titles, insert soundtracks, and then burn to DVD. This would presumably require rendering times comparable to what my friend is seeing, e.g., several times the length of the original per gigahertz of CPU speeed, e.g., overnight for a 2-hour piece of source material. As for your 2.5 GHz Celeron making a DVD bit copy at high speeds, congratulations. Except you should realize that bit copies are basically made at the writing speed of your DVD burner. So if your burner writes at 12x, a 2-hour standard DVD gets copied in 20 minutes. Whether your CPU runs at 2.5 GHz or at 1 GHz. Think about it. Rendering is not the same thing as bit copying. --Tim May |
#22
|
|||
|
|||
In article ,
Lawrence Glickman wrote: On 4 Dec 2004 20:15:09 -0500, (DoN. Nichols) wrote: [ ... ] If you do image processing, you will need the big RAM, or a bit swap file. The image processing programs like to have the *whole* image in RAM -- even if it has to use swap space to do it. :-) Is why I stopped using Photoshop. I have no need for high-resolution photographic illustrations for magazine covers and full-page advertisements. I process hundreds of personal digital images at a time, and with the programs I use, the work is done completely without my involvement in a few minutes. Of course, I had to initially set up the programming to do what I wanted it to do, but now it is automated. Good enough to better for 99.88% of my photos. Understood. I've got a (unix) shell script which takes a directory full of images, and makes a skeleton web page around them, producing both thumbnails and reduced size images (4X for the Nikon D70 at medium resolution, 2x for the Nikon CoolPix 950) for what you get if you click on the thumbnail. Then -- all I need to do is add descriptive text -- or forget it depending on what the subject matter is. A bunch of shots of the 4th of July fireworks doesn't need much in the way of text, but details of a project may need quite a bit more. [ ... ] 307 digits, which makes your 3.2574E+306 pretty close, but it is missing a lot of digits for true precision. Do you have a way to do it in multi-precision integer math? Not at this time, no. Otoh, if I needed to do that for a good reason, I would get hold of the necessary programming. O.K. For unix, the tool of choice is "dc" (desktop calculator) which takes RPN commands, and handles multiple precision as needed. I've only tried up to 500!, and that worked well. If finally got to where I could see the time that it took. (0.50 seconds combined -- and it took longer to print the results on the screen -- all seventeen lines of digits. :-) For rendering analog video into digital video, as TM suggested, I haven't explored that yet. Not sure I have enough interest in doing so to justify the investment in time and money. Maybe DVD is more permanent than tape, but it isn't forever. Delamination is a problem, along with the growth of anaerobic organisms that attack the media. I think you will be lucky to get 5 years out of DVD even if it is stored in a controlled environment. The stuff isn't as archival as people would lead you to believe. Agreed -- but it is quicker access than a 5GB+ 8mm tape. It is useful for short-term backups, and save the tapes for more serious backups. (Perhaps time to move to an Exabyte "Mammoth" tape drive, to get 18GB+, as my disk partitions keep getting bigger. :-) And -- if they would just come out with an affordable DVD writer with a SCSI interface, I could go into that world, too. :-) Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#23
|
|||
|
|||
In article , DoN. Nichols
wrote: Maybe DVD is more permanent than tape, but it isn't forever. Delamination is a problem, along with the growth of anaerobic organisms that attack the media. I think you will be lucky to get 5 years out of DVD even if it is stored in a controlled environment. The stuff isn't as archival as people would lead you to believe. Agreed -- but it is quicker access than a 5GB+ 8mm tape. It is useful for short-term backups, and save the tapes for more serious backups. (Perhaps time to move to an Exabyte "Mammoth" tape drive, to get 18GB+, as my disk partitions keep getting bigger. :-) I disagree in almost every respect. I'll try to politely explain why. First, I have had several tape breakages on my 8mm and Hi-8 tapes in the 17 years I've been using camcorders. I have not yet had a breakage in one of my DV tapes, but I am keeping my fingers crossed--any time a thin Mylar tape is moving over heads, I am worried. And the innards of my camcorders are filled with lots of gears and tension rollers and other moving parts. Second, none of my DVDs have gone bad. And because I might step on them, or whatever, I usually make at least two (2) of each. Third, a friend of mine got a tape drive backup system. From H-P, as I recall. When it crapped out, he faced a $1000 repair bill, with few alternatives. He ended up deciding not to replace the tape drive, and was fortunate to have most of the files still on disk. (By contrast, I have two laptops that can read DVDs, two desktops that can read DVDs, and six machines or readers than can read CDs. And I know I can go into the nearest Circuit City and buy a new DVD+-R+-RW machine for $100. So I know that so long as my DVDs are OK, I can read them. The same cannot be said of the several formats of tape drive that have been sold over the past decade.) Fourth, I routinely give or lend DVDs and CDs to my friends. The same cannot be said of tape from various competing tape drive formats. And -- if they would just come out with an affordable DVD writer with a SCSI interface, I could go into that world, too. :-) Fifth, I very, very, very happily gave up the world of SCSI when I went to Firewire several years ago. I retired my stack of SCSI drives, from Iomega to 80 MB to 200 MB to 1 GB to 5 GB and haven't turned them on since (I transferred the contents to a higher capacity drive, then started backing up onto CD-Rs and DVD-Rs). I started using SCSI in 1986 and, while it was an advance over mostly non-existent alternatives in the PC world, it was a 15-year series of hassles over cable lengths, terminators, ID switches, and diagnostics to track down problems. Firewire, or the PC standard of USB 2.0, solves these problems. (Or internal, ATA or IDE.) I would not consider getting a SCSI DVD writer for all the tea in China. --Tim May |
#24
|
|||
|
|||
On 4 Dec 2004 20:02:11 -0500, the renowned (DoN.
Nichols) wrote: In article , Spehro Pefhany wrote: On Sat, 04 Dec 2004 05:15:10 GMT, the renowned "Martin H. Eastburn" wrote: [ ... ] I remember doing 100! or factorial 100 in full integer precision printout. The first time 11 1/2 hours then 3 hours. First in Advanced Disk Basic, then Machine language (not assembly). Oh the processor 8080. This was in 1975 by myself. Martin Barely noticable (in human terms) time on a modern machine: 933262154439441021883256061085752672409442548549 60571509166910400 407995064242937148632694030450512898042989296944 47489825873720431 1236641477561877016501813248. Something is wrong there. The number of digits is right 158 characters but a sanity check says that there should be a *lot* of trailing zeros. After all, it has as factors 100 90 80 70 60 50 40 30 20 and 10, not counting other factors which add up to more zeros, such as 2 and 5. I get: 9332621544394415268169923885626670049071596826438 1621468592963895217\ 5999932299156089414639761565182862536979208272237 5825118521091686400\ 0000000000000000000000 twenty-four trailing zeros. It looks as though your program (whatever it was) used some form of extended precision floating-point math, with some kind of conversion errors. 9332621544394410 (yours) 9332621544394415 (mine) so -- we start to differ in the 16th digit. Enjoy, DoN. Good point. That explains why the function is limited to 150-odd factorial when I spec'd hundreds of digits. It is using floating point math to calculate some part of it rather than the 300-digit arbitrary precision math that I asked it to use. Best regards, Spehro Pefhany -- "it's the network..." "The Journey is the reward" Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com |
#25
|
|||
|
|||
In article ,
Tim May wrote: In article , DoN. Nichols wrote: [ ... ] Agreed -- but it is quicker access than a 5GB+ 8mm tape. It is useful for short-term backups, and save the tapes for more serious backups. (Perhaps time to move to an Exabyte "Mammoth" tape drive, to get 18GB+, as my disk partitions keep getting bigger. :-) I disagree in almost every respect. I'll try to politely explain why. First, I have had several tape breakages on my 8mm and Hi-8 tapes in the 17 years I've been using camcorders. I have not yet had a breakage in one of my DV tapes, but I am keeping my fingers crossed--any time a thin Mylar tape is moving over heads, I am worried. And the innards of my camcorders are filled with lots of gears and tension rollers and other moving parts. O.K. Now camcorders are subjected to a lot more shocks than Exabyte drives built into drive cases with other drives. I have, beside my desktop machine, a case containing (from top down) One Dual SCSI PCMIA reader Two Exabyte 8505 8mm tape drives. One Yahama CD-R/RW drive. These are connected to one standard (50-pin) SCSI port on the computer. Connected to the other a two internal 18 GB disk drives one internal DVD reader drive and a Multipack housing containing twelve fast-wide SCSI disk drives, ranging in size from 4 GB to 36 GB, with most being either 18 GB or 9 GB. Any one of these latter drives can be unmounted while the system is running, and then unplugged and replaced. True hot-swapping. And I have done it several times, as I replaced smaller drives with larger ones. That 4 GB drive is an endangered species, and will be swapped out as soon as I find another 18 GB or better at a hamfest. Both the computer and the Multipack are Sun products, and the disk drives are SCA interface (fast wide SCSI with the power and SCSI ID selection all brought in through the drive's connector. Second, none of my DVDs have gone bad. And because I might step on them, or whatever, I usually make at least two (2) of each. If I were writing them I might be able to make the same statement. But I can't, at the present. The only failure that I have had in the Exabyte 8505 drives (5 GB with standard tape, 8 GB with XL tapes, all without compression, and more possible with compression built into the drives) was when a nylon gear split on one end of the loading mechanism. This happened when the tape was ejecting after a nightly automatic backup. So, for the short term, I swapped over to the other drive in the housing, and in the long term, I got a replacement drive from eBay (the other drives came from either eBay or hamfest purchases.) And, BTW, I was able to extract the tape without damage -- it is still readable, but it take some careful work. Third, a friend of mine got a tape drive backup system. From H-P, as I recall. When it crapped out, he faced a $1000 repair bill, with few alternatives. He ended up deciding not to replace the tape drive, and was fortunate to have most of the files still on disk. a) HP is overpriced. (And I think that their drives are DAT drives, not 8mm drives though I have seen their expensive backup jukebox systems using drives from other vendors.) b) He apparently did not investigate used equipment for a replacement. (I buy almost exclusively used equipment, as I am retired, and don't have the money which I would if I were earning with the aid of these systems. c) I *always* have more than one drive which can read my backup tapes. For many of them, I have more than two drives. (Again surplus prices -- not new prices.) But if you are going to buy an expensive tape backup system -- buy a backup for that system. d) Know more about your backup system than the vendor's representative does. (By contrast, I have two laptops that can read DVDs, two desktops that can read DVDs, and six machines or readers than can read CDs. And I know I can go into the nearest Circuit City and buy a new DVD+-R+-RW machine for $100. So I know that so long as my DVDs are OK, I can read them. The same cannot be said of the several formats of tape drive that have been sold over the past decade.) Fourth, I routinely give or lend DVDs and CDs to my friends. The same cannot be said of tape from various competing tape drive formats. I have one friend with whom I occasionally exchange 8mm tapes. The format is set by the OS, not the drive, and for exchange purposes, I use the tar format by choice. Granted -- Windows systems tend to have dozens of mutually incompatible backup systems which even may use the same physical tapes, but not be able to read data interchangeably. This is a problem with the combination of the OS and proprietary backup formats, not with tape per se. And -- if they would just come out with an affordable DVD writer with a SCSI interface, I could go into that world, too. :-) Fifth, I very, very, very happily gave up the world of SCSI when I went to Firewire several years ago. I retired my stack of SCSI drives, from Iomega to 80 MB to 200 MB to 1 GB to 5 GB and haven't turned them on since (I transferred the contents to a higher capacity drive, then started backing up onto CD-Rs and DVD-Rs). I started using SCSI in 1986 and, while it was an advance over mostly non-existent alternatives in the PC world, it was a 15-year series of hassles over cable lengths, terminators, ID switches, and diagnostics to track down problems. The SCA forms of the SCSI drives eliminate the ID switches consideration. (The ID is defined by the socket into which the drive plugs.) The termination problem has pretty much been eliminated by the vendors who continue to use SCSI as a standard part of their system. The MultiPack drive housings, and the single-drive Unipack (with switches on the back to define the SCSI ID) are self-terminating, and will automatically switch off or on the termination needed as you plug in something downstream from them. They will even terminate half of the fast-wide SCSI bus automatically, if you plug in a transition from fast-wide (68-pin) to standard (50 pin). *And* -- they will tell you which halves of the bus are terminated by a glance at the LEDs on the back of the drive housing. SCA drives started (with Sun) in the internal drives in the SS-5 and SS-20 machines. Then they appeared in the big RAID systems (30 drives -- Sun Storage Arrays) connected to the system via an optical interface (yet another sBus card). The SCA drives also appeared in some DEC made RAID systems -- truly hot-swappable, including two plug-in power supplies for each horizontal cluster of five disk drives, IIRC. (We had those at work before I retired.) There are also MultiPack housings which accept only six drives, but those are Ultra-SCSI -- an even faster bus than standard fast-wide. (And -- the system will adjust the speed automatically for the slowest drive currently connected, so you *can* mix slower drives in there if you really want to. These 6-slot MultiPacks have only one switch (other than the power switch) -- a switch which selects whether the drives will get the lower half of the address range (0-6), or the upper half (8-14). These are normally used on a second SCSI card in the system, an the twelve-slot one is used on the system's built-in SCSI bus. So it provides SCSI IDs of 2, 3, 4, 5, 8, 9 10, 11, 12, 13, 14 and 15 -- skipping over 0, and 1 (the two internal disk drives), plus 6 (the internal DVD-ROM drive, or the CD-ROM drive, depending on the system. Firewire, or the PC standard of USB 2.0, solves these problems. (Or internal, ATA or IDE.) How many drives can you hang off a single IDE bus? Two. And it is picky about which drive it will allow as a boot drive. Typical systems have two IDE bus connectors, and can only boot from half of the drives -- thanks to some combination of the design of the interface and the design of the boot firmware in the mother board. How many active drives can you hang off USB before they start getting in each other's way? And confusing the keyboard and mouse, which now also use the USB. (I don't know, but I have heard some horror stories. How fast is USB compared to a fast wide Ultra SCSI bus? Serial ATA is one drive per socket on the motherboard. Two maximum in the token Windows box (which is kept away from the outside net, so I don't have to keep patching it. It is a purely internal use system. (Income tax software, transforming weird format images from weird cameras to something which other systems understand, and shoving them through the internal net to the unix boxen where they actually get *used*. The only box which I have with firewire on the Motherboard is running an OS which does not (yet) know how to use it. (But let me tell you OpenBSD really *screams* on a 2.3 GHz Celeron. :-) Could you even manage to run 14 disk drives on a single Windows system using all of those interfaces you have mentioned together? I'm running all of those on the single built-in fast-wide SCSI bus. If I want more, I can plug in extra SCSI cards into the system. (And no, I don't count an external RAID box making a cluster of drives look like a single one as falling within the spirit of that, because there is a limit to the number of partitions you can make on those.) I like partitions to keep related things confined to a single partition -- which *may* be using a single disk drive as one partition, or having up to eight partitions on Sun Solaris (or older Sun SunOs) systems, and up to sixteen partitions on a single drive on OpenBSD systems (on the same hardware). Also -- I *could* plug an external RAID system into each of those SCSI IDs, so I would still be ahead of your disk count -- 14 RAID arrays. I would not consider getting a SCSI DVD writer for all the tea in China. While I would not consider getting one in IDE, except for the token Windows box. (Granted, the Sun does not *have* an IDE interface, but it has an excellent SCSI system. But -- your experience with SCSI has been with systems which use it as an afterthought -- not as a normal part of the OS and hardware, where it has been made to work really well. Obviously, we have different experiences, and thus different perceptions and preferences. Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#26
|
|||
|
|||
Lawrence Glickman wrote:
On 04 Dec 2004 19:00:43 GMT, Ian Stirling wrote: In rec.crafts.metalworking Lawrence Glickman wrote: On 04 Dec 2004 17:10:21 GMT, Ian Stirling wrote: I can't imagine running anything my 2.5 giger can't handle. It gets the job done. A 2.5Ghz pentium may only be a couple of times faster than a P100, for some tasks. If it requires random access to large (greater than the on-chip cache) amounts of memory, random access speed of chips hasn't really changed in the past decade. Some 'supercomputer' type computers took extreme measures to get round this. OK, I put in 512MB of RAM and this thing is working as advertised. Less than 512 is not a good idea, because I can see from my stats that I only have 170MB of unused physical memory at the moment. Once you get enough RAM so you don't have to do HD read/writes, you're on your way to speed. But M$ has shafted everybody, because M$ loves Not for all tasks. For a few scientific computing tasks (nuclear simulations are a big one), the big bottleneck is how long it takes you to read a dozen bytes from a 'random' location in memory. All the information is in memory - it's just that the memory isn't fast enough. The increase in random access speed has not nearly matched the increase in streaming speed. IIRC, a 20 year old 30 pin 1Mb SIMM would typically take 150ns (billionths of a second) for the first byte, then 100ns for the next one. A modern 1Gb memory DIMM might read out the second (and subsequent) bytes a hundred times faster than the 1Mb SIMM. But the initial access to any location in that DIMM is still only around at best 3-4 times faster. In some tasks, this can be the limiting factor for calculation. No calculation my computer is ever going to need to make. We're talking apples and oranges here. I'm talking about getting a motorcycle up to 200 mph, you're talking about taking a truck up to 200 mph. Yes, they are going to require different engines. For my purposes ( motorcycle ), the load my CPU has to deal with, things appear normal in human time. I am not simulating nuclear detonations in virtual reality. Any nanosecond speed differences are not perceptible to this human unit. I just performed 170! ( factorial of 170 ). That's the limit on my *engine* I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* ??????? Lg The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Martin -- Martin Eastburn, Barbara Eastburn @ home at Lion's Lair with our computer NRA LOH, NRA Life NRA Second Amendment Task Force Charter Founder |
#27
|
|||
|
|||
DoN. Nichols wrote:
In article , Lawrence Glickman wrote: On 04 Dec 2004 19:00:43 GMT, Ian Stirling wrote: In rec.crafts.metalworking Lawrence Glickman wrote: On 04 Dec 2004 17:10:21 GMT, Ian Stirling wrote: [ ... ] In some tasks, this can be the limiting factor for calculation. No calculation my computer is ever going to need to make. We're talking apples and oranges here. I'm talking about getting a motorcycle up to 200 mph, you're talking about taking a truck up to 200 mph. Yes, they are going to require different engines. If you do image processing, you will need the big RAM, or a bit swap file. The image processing programs like to have the *whole* image in RAM -- even if it has to use swap space to do it. :-) For my purposes ( motorcycle ), the load my CPU has to deal with, things appear normal in human time. I am not simulating nuclear detonations in virtual reality. Any nanosecond speed differences are not perceptible to this human unit. I just performed 170! ( factorial of 170 ). That's the limit on my *engine* I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* My SS-5 (a mere 170 MHz CPU speed) does that one at: ================================================== ==================== izalco:dnichols 20:09 time dc /tmp/trial-170 72574156153079989673967282111292631147169916812964 513765435777989005\ 61843401706157852350749242617459511490991237838520 776666022565442753\ 02532890077320751090240043028005829560396661259965 825710439855829425\ 75689663134396122625710949468067112055688804571933 402126614528000000\ 00000000000000000000000000000000000 0.05u 0.05s 0:00.16 62.5% ================================================== ==================== A total of 0.1 second between user and system time, and 0:00.16 seconds wall clock time (with the system doing other things at the same time.) 307 digits, which makes your 3.2574E+306 pretty close, but it is missing a lot of digits for true precision. Do you have a way to do it in multi-precision integer math? Enjoy, DoN. I posted a reply earlier tonight - I had used an array and integer math at that! Martin -- Martin Eastburn, Barbara Eastburn @ home at Lion's Lair with our computer NRA LOH, NRA Life NRA Second Amendment Task Force Charter Founder |
#28
|
|||
|
|||
In article , DoN. Nichols
wrote: But -- your experience with SCSI has been with systems which use it as an afterthought -- not as a normal part of the OS and hardware, where it has been made to work really well. A silly comment. My first SCSI machine was a Mac Plus, in 1986, where it was indeed designed in close conjunction with the OS and hardware. It was sensitive to termination, addresses, etc. for very basic reasons. The PC versions arrived years later, through add-on cards, and were so flaky that SCSI was considered a joke in the PC community. --Tim May |
#29
|
|||
|
|||
On Sun, 05 Dec 2004 07:37:22 GMT, "Martin H. Eastburn"
wrote: The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Martin What's that Martin? You got some sort of prejudice against BCD??? G Mark Rand RTFM |
#30
|
|||
|
|||
On Sun, 05 Dec 2004 07:37:22 GMT, "Martin H. Eastburn"
wrote: I got an answer of 7.2574E+306 before I could even take my finger off the enter key. Is that not fast enough for a *consumer computer* ??????? Lg The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Martin so, there are a lot of other factors involved than raw CPU speed. This is what benchmarking measures. Speeds of data bus, memory bus, disc drive I/O, and so forth. There is a cluster of components that need to work together to produce a result, and the slowest component in the chain is always going to be the bottleneck. So to be fair, we do benchmarking using different algorithms, to give us an idea of SYSTEM performance, not raw Mips. But again, I am not going to buy an elephant rifle to kill a mosquito. Maybe scientific laboratories can profit from these high-end machines, but unless I get into analog to digital interpolation, it is highly doubtful I am going to benefit from having a cannon on board where a pellet pistol would do the job just as well. And even then, the throughput of the A/D converter is going to determine to a large extent what limitations my *system* would have. I try to stay digital all the way. Keep analog out of it, and you've solved a lot of problems before they even got off the launch pad. I've tossed all my legacy equipment, just threw it out into the garbage ( after sanitizing it of course, with a flamethrower ). Lg |
#31
|
|||
|
|||
In article , Lawrence
Glickman wrote: So to be fair, we do benchmarking using different algorithms, to give us an idea of SYSTEM performance, not raw Mips. But again, I am not going to buy an elephant rifle to kill a mosquito. Maybe scientific laboratories can profit from these high-end machines, but unless I get into analog to digital interpolation, it is highly doubtful I am going to benefit from having a cannon on board where a pellet pistol would do the job just as well. And even then, the throughput of the A/D converter is going to determine to a large extent what limitations my *system* would have. I try to stay digital all the way. Keep analog out of it, and you've solved a lot of problems before they even got off the launch pad. You're missing the point. The ADC is _not_ what makes rendering take many, many hours. In fact, the output of a 13.7 GB digital video (DV) is fed directly into a computer via Firewire ports. Digital at all times once it's in the camcorder, so digital at all times in the computer. What takes time is the resampling/resizing/reformatting/rewhatevering to "fit" 13.7 GB of digital data, for example, into 4.7 GB of space on the the blank DVD. This is akin to when a printer driver says "The image will not fit on this page...would you like to crop, scale, or abort?" Except with 13.7 GB that's a lot of data, a lot of swapping to disk, a lot of work. And I mentioned rendering only as one frequently-encountered "overnight job." There are a lot of things which actually take measurable time. The point being that your "I can do a factorial in the blink of an eye so I obviously don't need a faster computer!" example is silly. And if you never need a faster computer, why did you even buy a 2.5 GHz machine in the _first_ place? --Tim May |
#32
|
|||
|
|||
On Sun, 05 Dec 2004 10:25:31 -0800, Tim May wrote:
In article , Lawrence Glickman wrote: So to be fair, we do benchmarking using different algorithms, to give us an idea of SYSTEM performance, not raw Mips. But again, I am not going to buy an elephant rifle to kill a mosquito. Maybe scientific laboratories can profit from these high-end machines, but unless I get into analog to digital interpolation, it is highly doubtful I am going to benefit from having a cannon on board where a pellet pistol would do the job just as well. And even then, the throughput of the A/D converter is going to determine to a large extent what limitations my *system* would have. I try to stay digital all the way. Keep analog out of it, and you've solved a lot of problems before they even got off the launch pad. You're missing the point. The ADC is _not_ what makes rendering take many, many hours. In fact, the output of a 13.7 GB digital video (DV) is fed directly into a computer via Firewire ports. Digital at all times once it's in the camcorder, so digital at all times in the computer. What takes time is the resampling/resizing/reformatting/rewhatevering to "fit" 13.7 GB of digital data, for example, into 4.7 GB of space on the the blank DVD. This is akin to when a printer driver says "The image will not fit on this page...would you like to crop, scale, or abort?" Except with 13.7 GB that's a lot of data, a lot of swapping to disk, a lot of work. What the computer is doing is MPEGII compression of the video. That requires having at least 7 frames of video in memory at the same time. The algorithms look at the time domain data of the individual frames, translate it into the frequency domain, ie they do a fast Fourier transform, and then using the preceeding and following frames they find the most compact way to represent the group of frames as a MPEGII compressed data stream, which is what's laid down on the DVD. MPEGII is lossy compression. Some of the data is discarded, repetitive data is compressed. The algorithms do this in such a way so as to not produce visible artifacts (most of the time anyway). That's what makes it difficult. The algorithms are actually making *judgements* about picture content. Moving objects can be rendered with less resolution than stationary ones, so data is discarded for parts of the picture determined to be moving. Fixed areas in the picture don't have to be repeated in each frame, they're flagged for repetition instead. Etc. Very complicated. Decompressing MPEGII is easy. The compressed data stream has flags embedded in it telling the decompressor exactly what to do. There are commodity codec chips to do it on the fly in realtime. Compressing it is hard. Even the professional purpose built massively parallel processing boxes we use to do it for DTV broadcast lag realtime by about 2 seconds. Doing realtime with a single processor PC having limited memory is impractical. You'd just keep losing ground to the realtime signal. You can do offline processing with the PC, but it takes a very long time. 11 hours to do 1 hour of video sounds about right for a 2.5 GHz PC with a gigabyte of memory. And that's just for regular resolution video. Doing HDTV with a PC is out of the question. With only a gig of memory, the PC would be thrashing the disc all the time. HDTV MPEGII converters are very expensive. Only a few production houses, and the major networks, have them. At the local station level, we just pass through the already compressed streams. Uncompressed HDTV is a data stream at 1.2 Gb/s. Compression brings it down to 45 Mb/s. Then we do a complex modulation called 8VSB which allows us to stuff it into a 6 MHz TV channel bandwidth for broadcast. Gary |
#33
|
|||
|
|||
In article ,
Tim May wrote: In article , DoN. Nichols wrote: But -- your experience with SCSI has been with systems which use it as an afterthought -- not as a normal part of the OS and hardware, where it has been made to work really well. A silly comment. My first SCSI machine was a Mac Plus, in 1986, where it was indeed designed in close conjunction with the OS and hardware. It was sensitive to termination, addresses, etc. for very basic reasons. Would one of those reasons perhaps be Mac's use of the DB-25 connector for SCSI, instead of a proper 50-pin connector? The proper design of single-ended SCSI required an individual ground wire paired with *each* signal conductor, and ideally either twisted with that conductor, or at a minimum alternating with the signal conductor in a ribbon cable, so there was a shield between each of the signal conductors and its neighbor. Of *course* you have problems with a DB-25 connector used for that, because it interrupts the proper ground pairing, resulting in ground loops and other undesirable actions. The old Sun systems used a DD-50 connector (three rows of pins) for SCSI, and later used a miniature two-row connector from AMP (smaller than a DB-25), but a full 50 pins with click locks. Other vendors sometimes used the 50-pin Amphenol Blue Ribbon connectors (mis-called "Centronics" connectors, because the printer manufacturer, Centronics, used the 36-pin version of that connector as their standard parallel printer interface. Above I mentioned "single-ended SCSI", because there is also a differential SCSI. In that, the pins previously used for ground are driven in the opposite sense to the signal with which it is paired. There is a differential line receiver at the end (and each intermediate drive) on this. Standard single ended SCSI had a length limitation, and the higher the speed, the shorter the cable must be. This included the length of ribbon cable within the drive housings, which could sometimes add several feet. Differential SCSI can drive lines *much* longer than the single-ended without problems -- but it needs a proper differential terminator, not the normal single-ended one. There are several styles of terminators. The passive ones (just pairs of resistors to ground and VCC), which were adequate for shorter runs, the active ones, which had a voltage regulator from the VCC and ground to provide a proper intermediate voltage to which a single resistor is run from each signal, and the FPT (Forced Perfect Termination), which added to the active sets of diodes to ensure that the lines could not ring beyond the VCC and ground voltages. Now -- I mentioned a relationship between total cable length and speed. Sun systems sensed errors, and at a certain point would back off the speed from the maximum for that interface. Also -- cable quality makes a *big* difference. A proper SCSI cable was twisted pairs, run within a good shield (perhaps two or three layers woven, foil, and then woven again. On the other hand, using cheap cables *did* lead to problems. Using ones marked with the Sun logo did not. So -- I stand by my statement about Suns being designed to properly use SCSI, and Macs and PCs not being so designed. The very use of the DB-25 connector asks for problems before anything else gets involved. The PC versions arrived years later, through add-on cards, and were so flaky that SCSI was considered a joke in the PC community. Most of the early PC versions used the DB-25 connector as well, because that fit their card brackets without problems. The later cards (by Adaptec, for example), use the AMP high-density 50-pin connector which the Suns use, and those are quite well behaved with good cables. (I've used them -- but only with cables from Sun.) So -- most of the problems with home computers and SCSI stems from trying to do it on the cheap (the DB-25 connectors), and secondarily from the lack of an adaptive driver which knew when to drop the speed back to make up for an over-length total cable setup. A proper drive enclosure will have the 50-pin connector for the cable entrance, a loop of wire past the drive, with an IDC connector to plug into the drive, and the loop continues to another 50-pin connector. Use of the DB-25 connector makes this impractical, so there is typically a board with two DB-25s side by side, and a stub going off to the drive. Each stub introduces more reflections, and thus more potential for problems. Apple has, from the very start, depended on making software substitute for hardware, so the original Apple, as well as the long-lived Apple-II had weird artifacts, such as split memory segments in the middle of memory for graphics interfaces, simply to save a very few ICs in address decoding. And they used stripped floppy drives (compared to the rest of the world), substituting software for hardware there, too. A result of that was that they could squeeze more data into a floppy, by changing the data rate depending on which track you were using. This was nice -- but it meant that they were incompatible with the rest of the world. Then -- let us go into the Mac world. The early 68000-based systems could only address 16 MB of space (24 address bits), though the internals of the chip had room for 32-bit addresses, which could address 4 GB. OK, it is perhaps reasonable that they (at that time) figured that 8MB max for RAM, and the other 8MB for memory-mapped I/O was sufficient. But they started a "space-saving" trick which lead to a lot of heartaches later on. they used the upper eight bits of the address variables to store flags, and other things. When they move past the 68010 to the 68020 (which *could* address the full 32-bit address space, and which was *planned* for and announced when the 68000 was introduced, all of a sudden those flags were trying to address other real spaces, so programs from the earlier systems started breaking. Even worse, their software vendors picked up on the flags storage trick, so programs from the vendors also broke when put on the 68020 (and later) based systems. So -- with this record, I am supposed to believe that they got things right with SCSI? Sorry, I don't -- based on the evidence. For Sun, SCSI works very well. Even back in the Sun 2/120 (my first Sun machine), SCSI worked well. Yes -- I eventually stacked up enough drives in separate boxes so things started to fail -- until I subsitued 18" cables between boxes for the 6-foot ones which I had started with. SCSI, properly implemented, works well. (I've used it since its predecessor, SASI, which I used to interface drives to home-made interfaces in a SWTP 6800 running DOS-69 and OS-9 (the *real* OS-9, not what Mac came out with later.) :-) (Granted, I also had to write drivers for OS-9 for my wire-wrapped interface, but it worked, and worked well.) I was driving pairs of MFM drives connected to dual MFM controllers for the SASI bus. Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#34
|
|||
|
|||
In article ,
Martin H. Eastburn wrote: [ ... ] The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Hmm ... are you *sure* about that? I get 2568 characters from calculating 1000! -- after stripping out the '\' at the end of each line and the newline (line feed) which followed it, so there are only digits. (The output from dc is normally formatted for a terminal.) Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... Sure -- basic "bignum" techniques. This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Yep -- want to build one around an Intel 4004? :-) Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#35
|
|||
|
|||
In article ,
Mark Rand wrote: On Sun, 05 Dec 2004 07:37:22 GMT, "Martin H. Eastburn" wrote: The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Martin What's that Martin? You got some sort of prejudice against BCD??? Doe the 4-bit OS (and CPU) have provisions for handling overflow -- e.g. add 9 and 9 and you get 18 (0x12), which overflows the 4-bit values. And if you need to handle *signed* numbers, your maximum value is 7 -- one more becomes a negative number. Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#36
|
|||
|
|||
Mark Rand wrote:
On Sun, 05 Dec 2004 07:37:22 GMT, "Martin H. Eastburn" wrote: The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Martin What's that Martin? You got some sort of prejudice against BCD??? G Mark Rand RTFM No in fact I wrote a 64 bit to BCD converter - both directions many years ago. The nasty trick was to build the hardware display it used. Seven Segments were expensive and so were the BCD to 7-segment decoder chips. But I used TI's instead of the handful of decode logic. Monsanto was the 7-segment maker for me at the time. I later found out that an in-law type relative alt-tree type - was the Fairchild VP for Opto. Such is life. Martin -- Martin Eastburn, Barbara Eastburn @ home at Lion's Lair with our computer NRA LOH, NRA Life NRA Second Amendment Task Force Charter Founder |
#37
|
|||
|
|||
Whoops - wonder what 1745 was - maybe that was e or something.
Color me yellow :-( Yes - 1000! (2567)4.0238 72600 77093 77345 out to a few places - 'compiled from Ballistic Research Laboratory, A table of the factorial numbers and their reciprocals from 1 to 1000! to 20 significant digits, Tech note 381, Aberdeen Proving Ground Md.(1951) (with permission) [to] United States Department of Commerce. National Bureau of Standards "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables" Applied Mathematics Series: 55 Fourth printing December 1965. ~8x10" x 2.5" Library of Congress Catalog Card Number : 64-60036 A very tech table book and was useful since my schooling days! If you need trig tables out or log table out or Gama tables - transcendental functions and other fun things we had to use prior to the little plastic boxes with built in functions. My 4 banner - Nixie tube - 12 digit - had taped to the top the mathematical arguments to generate logs from scratch. That sucker cost us 600 USD and was prior to TI or HP. It was a business machine. Martin DoN. Nichols wrote: In article , Martin H. Eastburn wrote: [ ... ] The method I employed was to set up an array of integers of the F size. e.g. Not so easy to guess when the numbers get large, but 1000! is under 2000 digits. Hmm ... are you *sure* about that? I get 2568 characters from calculating 1000! -- after stripping out the '\' at the end of each line and the newline (line feed) which followed it, so there are only digits. (The output from dc is normally formatted for a terminal.) Then take the first two numbers and multiply, place the results into the array. Multiply that by the next number - and when done, make sure each array element is a single digit. If not - bump it up.... and re-test... Sure -- basic "bignum" techniques. This then takes massive numbers and puts them into very simple baby talk numbers. Any system can handle one digit vs. another. (unless it were a 4 bit o.s. :-) ) Yep -- want to build one around an Intel 4004? :-) Enjoy, DoN. -- Martin Eastburn, Barbara Eastburn @ home at Lion's Lair with our computer NRA LOH, NRA Life NRA Second Amendment Task Force Charter Founder |
#38
|
|||
|
|||
In article ,
(DoN. Nichols) wrote: [...] Then -- let us go into the Mac world. The early 68000-based systems could only address 16 MB of space (24 address bits), though the internals of the chip had room for 32-bit addresses, which could address 4 GB. OK, it is perhaps reasonable that they (at that time) figured that 8MB max for RAM, and the other 8MB for memory-mapped I/O was sufficient. But they started a "space-saving" trick which lead to a lot of heartaches later on. they used the upper eight bits of the address variables to store flags, and other things. When they move past the 68010 to the 68020 (which *could* address the full 32-bit address space, and which was *planned* for and announced when the 68000 was introduced, all of a sudden those flags were trying to address other real spaces, so programs from the earlier systems started breaking. Even worse, their software vendors picked up on the flags storage trick, so programs from the vendors also broke when put on the 68020 (and later) based systems. To clarify: Apple implemented a 32-bit mode on the later 68k machines. If you needed backward compatibility, which I did for MS's ****, you could switch modes and reboot. The OS simply told all of the apps it was using all of the addresses above 16MB, so apps stayed out of there and they could get away with the 24bit memory hackery. You were just short on RAM until you switched back and rebooted. Since Apple was pretty reliable about selling computers with 16M or less RAM up until they switched to Power PC, it's not too much of a rip-off. (: [...] -- B.B. --I am not a goat! thegoat4 at airmail dot net http://web2.airmail.net/thegoat4/ |
#40
|
|||
|
|||
Martin H. Eastburn wrote:
Sorry Spehro - I forgot - it was 1000! it had 1732 ? chars in it. Just fit on my Queme printout. Well, that's: 40238726007709377354370243392300398571937486421071 463254379991042993\ 85123986290205920442084869694048004799886101971960 586316668729948085\ 58901323829669944590997424504087073759918823627727 188732519779505950\ 99527612087497546249704360141827809464649629105639 388743788648733711\ 91810458257836478499770124766328898359557354325131 853239584630755574\ 09114262417474349347553428646576611667797396668820 291207379143853719\ 58824980812686783837455973174613608537953452422158 659320192809087829\ 73084313928444032812315586110369768013573042161687 476096758713483120\ 25478589320767169132448426236131412508780208000261 683151027341827977\ 70478463586817016436502415369139828126481021309276 124489635992870511\ 49649754199093422215668325720808213331861168115536 158365469840467089\ 75602900950537616475847728421889679646244945160765 353408198901385442\ 48798495995331910172335555660213945039973628075013 783761530712776192\ 68490343526252000158885351473316117021039681759215 109077880193931781\ 14194545257223865541461062892187960223838971476088 506276862967146674\ 69756291123408243920816015378088989396451826324367 161676217916890977\ 99119037540312746222899880051954444142820121873617 459926429565817466\ 28302955570299024324153181617210465832036786906117 260158783520751516\ 28422554026517048330422614397428693306169089796848 259012545832716822\ 64580665267699586526822728070757813918581788896522 081643483448259932\ 66043367660176999612831860788386150279465955131156 552036093988180612\ 13855860030143569452722420634463179746059468257310 379008402443243846\ 56572450144028218852524709351906209290231364932734 975655139587205596\ 54228749774011413346962715422845862377387538230483 865688976461927383\ 81490014076731044664025989949022222176590433990188 601856652648506179\ 97023561938970178600408118897299183110211712298459 016419210688843871\ 21855646124960798722908519296819372388642614839657 382291123125024186\ 64935314397013742853192664987533721894069428143411 852015801412334482\ 80150513996942901534830776445690990731524332782882 698646027898643211\ 39083506217095002597389863554277196742822248757586 765752344220207573\ 63056949882508796892816275384886339690995982628095 612145099487170124\ 45164612603790293091208890869420285106401821543994 571568059418727489\ 98094254742173582401063677404595741785160829230135 358081840096996372\ 52423056085590370062427124341690900415369010593398 383577793941097002\ 77534720000000000000000000000000000000000000000000 000000000000000000\ 00000000000000000000000000000000000000000000000000 000000000000000000\ 00000000000000000000000000000000000000000000000000 000000000000000000\ 00000000000000000000000000000000000000000000000000 00 (Took an unnoticiable time to calculate using bc on my home computer. Man, it's been ages since I used bc!) |
Reply |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
DRO System Mill/Drill | Metalworking | |||
Radiators and CH system... | UK diy | |||
mains Hot water, and do I convert open heating to a closed heating system | UK diy |