Nick_Janow@mindlink.bc.ca (Nick Janow) (06/16/91)
If you want a fast machine with a nice interface, take a look at NeXT. I've heard that it's the best platform for Mathematica, especially if you give it enough RAM (>the 8M standard version). If you have several NeXTs networked, you can use the Zilla app to multiprocess. There's talk of a 88110-based NeXT, which would boost performance even more. That configuration would aslo have a Motorola 96001 DSP chip, for fast floating point array processing. Check out the available information at a university that carries NeXTs, and also check out comp.NeXT. Only you can decide which computer is best for you. :) -- Nick_Janow@mindlink.bc.ca
NJ_GOKEMEIJE@FANDM.BITNET (06/17/91)
Ok, Here is the situation. We do a lot of calculations for theoritcal physics, and basically are unsatisfied with the time it takes. What would be the best price/performance numbercruncher?? Rightnow we have a VAX 3800 station running Speakeasy. Nice, virtual memory etc, but slow ( 4 VUPS == VaxMIPS) Bad user interface : Only linediting, no previous commands, no windows How is a MAC SE/30 or II ci running MATHEMATICA? or II fx ( two expensive) With system seven you can have virtual memory, but mac OS is slow, no multitasking etc ( cannot have it crunch while doing other stuff) (no flame intended: the MAC has multitasking, but just not of such quality as would be nice in this case) How about THEORIST on the MAC? Or do I have to purchase a UNIX station running MAPLE, to get good results? how expensive is that and how fast? Where can I put an A3000 or A3000UX? is the UX faster? is AMIGADOS friendlier than UNIX, and still robust enough to numbercrunch? (IS virtual memory available and reliable?) Is MAPLE the only math program for the AMIGA. Any responses are welcome. Nils Gokemeije (NJ_GOKEMEIJE@FANDM.bitnet)
jet@karazm.math.uh.edu (J Eric Townsend) (06/17/91)
In article <71A0D62DC000263C@FANDM> NJ_GOKEMEIJE@FANDM.BITNET writes: >What would be the best price/performance numbercruncher?? Sun SparcStation-2, perhaps? Intel iPSC/860 hypercube? >How is a MAC SE/30 or II ci running MATHEMATICA? >Is MAPLE the only math program for the AMIGA. Probably for the time being. Several people have hassled Wolfram in public regarding Mathematica, but he's got some sort of anti-Amiga religion, it would seem. -- J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126 Skate UNIX! (curb fault: skater dumped) -- If you're hacking PowerGloves and Amigas, drop me a line. --
Nick_Janow@mindlink.bc.ca (Nick Janow) (06/17/91)
jet@karazm.math.uh.edu (J Eric Townsend) writes: > In article <1991Jun17.102210.29176@news.iastate.edu> taab5@isuvax.iastate.edu > writes: > >+ Why should he support the Amiga? I'd like a simple straight answer to that >+ question. Why should he support a machine that has an installed base of >+ fewer than 100,000 systems capable of running Mathematica? He is not > > Gosh, there're *maybe* that many SparcStations up and running, and > Mathematica is available for the Sparc... You also have to consider the percentages that are owned by people who would buy and use Mathematica. I've heard rumours (unconfirmed, mind you) that Amigas tend to be used mainly for games and video art. :) -- Nick_Janow@mindlink.bc.ca
taab5@isuvax.iastate.edu (Marc Barrett) (06/17/91)
In article <1991Jun17.023916.1026@menudo.uh.edu>, jet@karazm.math.uh.edu (J Eric Townsend) writes: >In article <71A0D62DC000263C@FANDM> NJ_GOKEMEIJE@FANDM.BITNET writes: >>What would be the best price/performance numbercruncher?? > >Sun SparcStation-2, perhaps? Intel iPSC/860 hypercube? > >>How is a MAC SE/30 or II ci running MATHEMATICA? > >>Is MAPLE the only math program for the AMIGA. > >Probably for the time being. Several people have hassled Wolfram >in public regarding Mathematica, but he's got some sort of >anti-Amiga religion, it would seem. Why should he support the Amiga? I'd like a simple straight answer to that question. Why should he support a machine that has an installed base of fewer than 100,000 systems capable of running Mathematica? He is not exactly losing much money by not supporting it, as the MAC market alone accounts for far more potential than the Amiga ever will, especially in the U.S. In order for it to be worth Wolfram's while to port Mathematica to AmigaDOS, Commodore will have to generate a large installed base of A3000 systems in the education market, which accounts for the bulk of Wolfram's sales. Needless to say, with the progress that Commodore is making in the education market, I do not see this happening. > >-- >J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126 >Skate UNIX! (curb fault: skater dumped) > > -- If you're hacking PowerGloves and Amigas, drop me a line. -- ------------------------------------------------------------- / Marc Barrett -MB- | BITNET: XGR39@ISUVAX.BITNET / / ISU COM S Student | Internet: XGR39@CCVAX.IASTATE.EDU / ------------------------------------------------------------ \ The great thing about standards is that / \ there are so many of them to choose from. / -------------------------------------------------------
peter@sugar.hackercorp.com (Peter da Silva) (06/17/91)
In article <71A0D62DC000263C@FANDM> NJ_GOKEMEIJE@FANDM.BITNET writes: > Where can I put an A3000 or A3000UX? is the UX faster? Slower. It's the exact same machine, running a bigger O/S. Probably still faster than any Mac, though. Even the IIfx isn't as fast as a 3000/25. And 040 cards are just around the corner. > is AMIGADOS friendlier than UNIX, and still robust enough to numbercrunch? Friendlier to the user. Some of the programming is a bit of a pain. Plenty robust enough... I'm numbercrunching right now and it doesn't even effect my foreground (I/O intensive) term program... and the numbercruncher is still getting near 100% of the CPU time. > (IS virtual memory available and reliable?) No. > Is MAPLE the only math program for the AMIGA. I have no idea what apps there are. The platform, however, is the best of any PC. And the price is way lower than any workstation. -- Peter da Silva. `-_-' <peter@sugar.neosoft.com>. 'U` "Have you hugged your wolf today?"
es1@cunixb.cc.columbia.edu (Ethan Solomita) (06/17/91)
In article <71A0D62DC000263C@FANDM> NJ_GOKEMEIJE@FANDM.BITNET writes: > >Where can I put an A3000 or A3000UX? is the UX faster? >is AMIGADOS friendlier than UNIX, and still robust enough to numbercrunch? >(IS virtual memory available and reliable?) >Is MAPLE the only math program for the AMIGA. > >Any responses are welcome. Nils Gokemeije (NJ_GOKEMEIJE@FANDM.bitnet) I THINK you are better off with the 3000, not the 3000UX. The Amiga OS doesn't have virtual memory, but RAM is cheap and if you put in 9MB you will probably never run out. And yes, it has a nicer interface and the interface is much faster (although number crunching is probably the same). Also, Maple has been out under AmigaDOS for a while, it is new on Unix. -- Ethan "...Know-Nothing-Bozo the Non-Wonder Dog, an animal so stupid that it had been sacked from one of Will's own commercials for being incapable of knowing which dog food it was supposed to prefer, despite the fact that the meat in all the other bowls had engine oil poured all over it."
melling@cs.psu.edu (Michael D Mellinger) (06/17/91)
In article <1991Jun17.142339.21049@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes:
I THINK you are better off with the 3000, not the 3000UX.
The Amiga OS doesn't have virtual memory, but RAM is cheap and if
you put in 9MB you will probably never run out. And yes, it has a
nicer interface and the interface is much faster (although number
crunching is probably the same). Also, Maple has been out under
AmigaDOS for a while, it is new on Unix.
Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a
computer with virtual memory.
-Mikees1@cunixb.cc.columbia.edu (Ethan Solomita) (06/18/91)
In article <fv5H7bfy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991Jun17.142339.21049@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes: > > > I THINK you are better off with the 3000, not the 3000UX. > The Amiga OS doesn't have virtual memory, but RAM is cheap and if > you put in 9MB you will probably never run out. And yes, it has a > nicer interface and the interface is much faster (although number > crunching is probably the same). Also, Maple has been out under > AmigaDOS for a while, it is new on Unix. > >Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a >computer with virtual memory. > >-Mike And wait for it to page in and out? When memory is $60/meg? What's the point? So get 18MB if you need that much. The time it saves you (if you need it) makes it more than worth the cost. -- Ethan "...Know-Nothing-Bozo the Non-Wonder Dog, an animal so stupid that it had been sacked from one of Will's own commercials for being incapable of knowing which dog food it was supposed to prefer, despite the fact that the meat in all the other bowls had engine oil poured all over it."
melling@cs.psu.edu (Michael D Mellinger) (06/18/91)
In article <1991Jun17.172930.13518@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes: >Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a >computer with virtual memory. > >-Mike And wait for it to page in and out? When memory is $60/meg? What's the point? So get 18MB if you need that much. The time it saves you (if you need it) makes it more than worth the cost. The problem is you don't know how much you will need. People run some pretty large Mathematica jobs. I'm willing to bet there are times when 32MB of RAM is not enough. Besides, if a job is running for hours you don't want to find out half way through that it ran out of memory. Let it page to disk. If you are using the machine for something else the CPU will be given to another task while the system is paging. Then there is the case where you actually want to use your system when Mathematica is running. Fire up that word processor and C compiler(well, I wouldn't do that w/o a little memory protection). BTW, RAM is down to $40/meg, and the price of 4MB SIMMS is rumored to be ready to drop again. -Mike
es1@cunixb.cc.columbia.edu (Ethan Solomita) (06/18/91)
In article <nd8H?xjy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991Jun17.172930.13518@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes: > > >Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a > >computer with virtual memory. > > > >-Mike > > And wait for it to page in and out? When memory is > $60/meg? What's the point? So get 18MB if you need that much. The > time it saves you (if you need it) makes it more than worth the > cost. > >The problem is you don't know how much you will need. People run some >pretty large Mathematica jobs. I'm willing to bet there are times >when 32MB of RAM is not enough. Besides, if a job is running for >hours you don't want to find out half way through that it ran out of >memory. Let it page to disk. If you are using the machine for >something else the CPU will be given to another task while the system >is paging. Then there is the case where you actually want to use your >system when Mathematica is running. Fire up that word processor and C >compiler(well, I wouldn't do that w/o a little memory protection). > >BTW, RAM is down to $40/meg, and the price of 4MB SIMMS is rumored to >be ready to drop again. > >-Mike IF the person we're giving advice to may really need more than 18MB of RAM, then probably the Unix machine would be better suited. From what I've seen of the Unix machines, the NeXT in particular, is that the biggest slowdowns occur when paging to/from disk. If there were constant paging the slowdown could be enormous. That's why if it fits in memory, keep it there. As to also using a WP and a compiler, they take up very minimal amounts of memory, especially compared to Mathematica (or Maple in this case). Not particularly relevant. -- Ethan "...Know-Nothing-Bozo the Non-Wonder Dog, an animal so stupid that it had been sacked from one of Will's own commercials for being incapable of knowing which dog food it was supposed to prefer, despite the fact that the meat in all the other bowls had engine oil poured all over it."
peter@sugar.hackercorp.com (Peter da Silva) (06/18/91)
In article <fv5H7bfy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a > computer with virtual memory. You can put 26 MB in an A3000 today, and god knows how much when Zorro-III memory cards come out. Plus, virtual memory means virtual performance. -- Peter da Silva. `-_-' <peter@sugar.neosoft.com>. 'U` "Have you hugged your wolf today?"
jet@karazm.math.uh.edu (J Eric Townsend) (06/18/91)
In article <1991Jun17.102210.29176@news.iastate.edu> taab5@isuvax.iastate.edu writes: > Why should he support the Amiga? I'd like a simple straight answer to >that question. Why should he support a machine that has an installed base >of fewer than 100,000 systems capable of running Mathematica? He is not Gosh, there're *maybe* that many SparcStations up and running, and Mathematica is available for the Sparc... -- J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126 Skate UNIX! (curb fault: skater dumped) -- If you're hacking PowerGloves and Amigas, drop me a line. --
jet@karazm.math.uh.edu (J Eric Townsend) (06/18/91)
In article <fv5H7bfy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991Jun17.142339.21049@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes: >>you put in 9MB you will probably never run out. And yes, it has a >Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a >computer with virtual memory. Or put 18 or so megs in your 'miga. Or get a *real* computer: Sun SparcStation-2. Or an Intel iPSC/860. -- J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126 Skate UNIX! (curb fault: skater dumped) -- If you're hacking PowerGloves and Amigas, drop me a line. --
jet@karazm.math.uh.edu (J Eric Townsend) (06/18/91)
In article <nd8H?xjy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: >The problem is you don't know how much you will need. People run some >pretty large Mathematica jobs. Why are these jobs so "large"? Maybe Mathematica needs tuning? -- J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126 Skate UNIX! (curb fault: skater dumped) -- If you're hacking PowerGloves and Amigas, drop me a line. --
dant@ryptyde.UUCP (Daniel Tracy) (06/18/91)
Responding to the following: "Even the IIfx isn't as fast as a 3000/25." That's bull. The IIfx will blow the 3000 away. It has two dedicated I/O chips (SCSI DMA too, which can be used with Unix), and a killer graphics coprocessed board (the 8-24GC) (which was released with the IIfx). I doubt you've done any kind of serious benchmarks.
jbickers@templar.actrix.gen.nz (John Bickers) (06/18/91)
Quoted from <1991Jun17.102210.29176@news.iastate.edu> by taab5@isuvax.iastate.edu (Marc Barrett): > Why should he support the Amiga? I'd like a simple straight answer to Because it's a better platform. > / Marc Barrett -MB- | BITNET: XGR39@ISUVAX.BITNET / -- *** John Bickers, TAP, NZAmigaUG. jbickers@templar.actrix.gen.nz *** *** "Endless variations, make it all seem new" - Devo. ***
taab5@isuvax.iastate.edu (Marc Barrett) (06/18/91)
In article <1991Jun17.233023.9110@menudo.uh.edu>, jet@karazm.math.uh.edu (J Eric Townsend) writes: >In article <1991Jun17.102210.29176@news.iastate.edu> taab5@isuvax.iastate.edu writes: >> Why should he support the Amiga? I'd like a simple straight answer to >>that question. Why should he support a machine that has an installed base >>of fewer than 100,000 systems capable of running Mathematica? He is not > >Gosh, there're *maybe* that many SparcStations up and running, and >Mathematica is available for the Sparc... The SparcStation != Amiga 3000. For one thing, most A3000 systems are being used by individuals as PCs, while most SparcStations are being used by researchers in offices or labs. Wolfram can also get away with charging more for the Sparc version of Mathematica, because researchers generally have adequate cash to spend on software. Here is the gist of the problem: Wolfram charges low prices for the MAC version of Mathematica because there are so many MACs being used by students who would want Mathematica for their system. Wolfram can also charge high prices for the Sparc version of Mathematica for the reasons I explained above. Wolfram will not sell Mathematica for the A3000, because there are not enough of them being used in research to get away with a high price, and not enough of them being used by students to get away with a low price. > >-- >J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126 >Skate UNIX! (curb fault: skater dumped) > > -- If you're hacking PowerGloves and Amigas, drop me a line. -- ------------------------------------------------------------- / Marc Barrett -MB- | BITNET: XGR39@ISUVAX.BITNET / / ISU COM S Student | Internet: XGR39@CCVAX.IASTATE.EDU / ------------------------------------------------------------ \ The great thing about standards is that / \ there are so many of them to choose from. / -------------------------------------------------------
schweige@aldebaran.cs.nps.navy.mil (Jeffrey M. Schweiger) (06/18/91)
In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: >Responding to the following: > >"Even the IIfx isn't as fast as a 3000/25." > >That's bull. The IIfx will blow the 3000 away. It has two dedicated I/O >chips (SCSI DMA too, which can be used with Unix), and a killer graphics >coprocessed board (the 8-24GC) (which was released with the IIfx). I doubt >you've done any kind of serious benchmarks. Actually, it probably depends on the application. Also, if we're talking about standard configuration machines. If I recall correctly, the IIfx CPU is rated at 40MHz, but the bus speed is limited to 20MHz. This would yield a faster CPU for the Mac IIfx and a faster bus for the Amiga 3000/25. Jeff Schweiger -- ******************************************************************************* Jeff Schweiger Standard Disclaimer CompuServe: 74236,1645 Internet (Milnet): schweige@taurus.cs.nps.navy.mil *******************************************************************************
daveh@cbmvax.commodore.com (Dave Haynie) (06/19/91)
In article <2394@aldebaran.cs.nps.navy.mil> schweige@taurus.cs.nps.navy.mil (Jeffrey M. Schweiger) writes: >In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: >>Responding to the following: >>"Even the IIfx isn't as fast as a 3000/25." >>That's bull. The IIfx will blow the 3000 away. >Actually, it probably depends on the application. I haven't personally run any comparative benchmarks on the two systems, so I don't really know how a IIfx does against an A3000, all told. I imagine the last statement is true and, in that context, so are the others. For example, disk activity. Without the IIfx using DMA (it didn't in System 6, does it in System 7?), it's going to be significantly slower if it's running full speed SCSI and also attempting to deal with a few CPU bound jobs. With DMA, while the SCSI transfers won't be as fast as the A3000's, there could possibly be more available CPU time if the IIfx has a decent DMA efficiency. >Also, if we're talking about standard configuration machines. If I recall >correctly, the IIfx CPU is rated at 40MHz, but the bus speed is limited to >20MHz. The IIfx has a CPU/cache bus which runs at 40MHz. The main motherboard bus runs at 20MHz, and NuBus of course runs at 10MHz. The A3000/25 runs its main motherboard bus at 25MHz and Zorro III based on 25MHz (Zorro III isn't itself actually clocked, but its strobes are derived from two 25MHz clocks when the 68030 acts as bus master). >This would yield a faster CPU for the Mac IIfx and a faster bus for the >Amiga 3000/25. Well, yes and no. Things that run full speed on the bus will go faster on an A3000 than on a IIfx, but most I/O devices don't necessarily go full speed. Our DMAC does, so it's faster than a IIfx DMAC could be. The 8520s, on the other hand, are pretty slow. Then again, AmigaOS is better able to allocate CPU cycles than MacOS due to its preemptive multitasking. So, bottom line, you don't know. I would wager that the Mac is noticably faster doing a single CPU-bound task, and the A3000 under a heavily disk bound task. Anything else is up for grabs. -- Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy "This is my mistake. Let me make it good." -R.E.M.
david@twg.com (David S. Herron) (06/20/91)
In article <nd8H?xjy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991Jun17.172930.13518@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes: >BTW, RAM is down to $40/meg, and the price of 4MB SIMMS is rumored to >be ready to drop again. Really? That's good news. Disk space is still well below $5/meg. I recently picked up a new 600 meg drive for $1200, that's $2/meg. *THIS* is the advantage which virtual memory has -- that it is cheaper. You simply buy enough memory to satisfy normal usage and for abnormal times you have disk to expand into... >-Mike -- <- David Herron, an MMDF & WIN/MHS guy, <david@twg.com> <- <- <- "MS-DOS? Where we're going we don't need MS-DOS." --Back To The Future
murphy@gibbs.physics.purdue.edu (William J. Murphy) (06/20/91)
In article <1991Jun18.122614.28459@news.iastate.edu> taab5@isuvax.iastate.edu writes: >In article <1991Jun17.233023.9110@menudo.uh.edu>, jet@karazm.math.uh.edu (J Eric Townsend) writes: >>In article <1991Jun17.102210.29176@news.iastate.edu> taab5@isuvax.iastate.edu writes: >>> Why should he support the Amiga? I'd like a simple straight answer to >> >>Gosh, there're *maybe* that many SparcStations up and running, and >>Mathematica is available for the Sparc... > > Here is the gist of the problem: Wolfram charges low prices for the >MAC version of Mathematica because there are so many MACs being used by >students who would want Mathematica for their system. Wolfram can also >charge high prices for the Sparc version of Mathematica for the reasons Although it has been a few years since I talked to people at WRI, I recall that they said that someone (an independent developer) would have to do the front-end development for the Amiga. Only when that is done will ther be much of a consideration for porting the MMA kernel to the Amiga. Now consider the other Unix boxes. The kernel is probably quite portable across Unix systems. WRI has had to generate an X-windows interface and a NeXTStep interface plus some other odd Unix window interfaces. It doesn't cost WRI that much to generate variants of the front-end and almost nothing to service the kernels. I agree that the Unix versions do cost too much, but as was pointed out earlier, corporations and universities have deeper pockets than the average home user. Why does MAC get MMA, because they have more '020 and '030 systems with nice windows I suppose. PCs got MMA about 2 years after MACs did. The other element that hasn't been considered is that maybe some corporation said to WRI, "We need MMA for MAC, here's $$$$$$ to port it." MatLab got ported to HP workstations because NASA Langley wanted it and paid for it. Bill Murphy murphy@physics.purdue.edu
melling@cs.psu.edu (Michael D Mellinger) (06/20/91)
In article <1991Jun17.203338.23645@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes:
IF the person we're giving advice to may really need more
than 18MB of RAM, then probably the Unix machine would be better
suited. From what I've seen of the Unix machines, the NeXT in
particular, is that the biggest slowdowns occur when paging
to/from disk. If there were constant paging the slowdown could be
enormous. That's why if it fits in memory, keep it there.
As to also using a WP and a compiler, they take up very
minimal amounts of memory, especially compared to Mathematica (or
Maple in this case). Not particularly relevant.
-- Ethan
I was referring to the fact that playing with C on a computer w/o
memory protection isn't a good thing if you are going to run a
program in the background for a few hours.
-Mikemelling@cs.psu.edu (Michael D Mellinger) (06/20/91)
In article <1991Jun17.233312.9357@menudo.uh.edu> jet@karazm.math.uh.edu (J Eric Townsend) writes:
Or put 18 or so megs in your 'miga. Or get a *real* computer:
Sun SparcStation-2. Or an Intel iPSC/860.
A SSII is about 50% faster than a NeXTstation and quite a bit more
expensive. And we can't forget that Mathematica is free on the NeXT
and it costs several hundred dollars for the SUN.
-Mikemelling@cs.psu.edu (Michael D Mellinger) (06/20/91)
In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes:
"Even the IIfx isn't as fast as a 3000/25."
That's bull. The IIfx will blow the 3000 away. It has two dedicated I/O
chips (SCSI DMA too, which can be used with Unix), and a killer graphics
coprocessed board (the 8-24GC) (which was released with the IIfx). I doubt
you've done any kind of serious benchmarks.
I think you must be under Unix(SYSVR2) to take advantage of the I/O
coprocessors in the IIfx. System 7 doesn't even utilize them. That
killer graphics coprocessor board won't work with System 7.0 for a few
more months. However, if you want a real killer graphics board, I
would recommend the NeXTDimension board for the NeXT Cube. I haven't
seen any benchmarks, but it is using an i860, so I'm willing to be
that it's damn fast!
-Mikepeter@sugar.hackercorp.com (Peter da Silva) (06/20/91)
In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: > "Even the IIfx isn't as fast as a 3000/25." > That's bull. No, it's an assertion. A bull is a male herbivore, cetacean, or pinniped. > The IIfx will blow the 3000 away. That's not what posted benchmarks say. > It has two dedicated I/O > chips (SCSI DMA too, which can be used with Unix), That's really going to speed up numbercrunching now, isn't it? In any case, the Amiga has dedicated I/O chips, DMA, dedicated SCSI, and an expansion bus that's far better than NuBus (which is pretty much purely a peripheral bus). > and a killer graphics > coprocessed board (the 8-24GC) (which was released with the IIfx). That's really going to speed up number crunching, now, isn't it. > I doubt you've done any kind of serious benchmarks. No, I haven't. Have you? But at least I know what sort of hardware we're looking at. Do you? -- Peter da Silva. `-_-' <peter@sugar.neosoft.com>. 'U` "Have you hugged your wolf today?"
torrie@cs.stanford.edu (Evan Torrie) (06/20/91)
peter@sugar.hackercorp.com (Peter da Silva) writes: >In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: >> The IIfx will blow the 3000 away. >That's not what posted benchmarks say. Huh? What posted benchmarks? I've never seen any. -- ------------------------------------------------------------------------------ Evan Torrie. Stanford University, Class of 199? torrie@cs.stanford.edu "And in the death, as the last few corpses lay rotting in the slimy thoroughfare, the shutters lifted in inches, high on Poacher's Hill..."
jbickers@templar.actrix.gen.nz (John Bickers) (06/20/91)
Quoted from <69cH+x7?@cs.psu.edu> by melling@cs.psu.edu (Michael D Mellinger): > I was referring to the fact that playing with C on a computer w/o > memory protection isn't a good thing if you are going to run a > program in the background for a few hours. It depends a bit on what you're doing... using something like Enforcer while writing your C (or being a reasonably competent programmer :) could save you a lot of grief should something unfortunate happen. > -Mike -- *** John Bickers, TAP, NZAmigaUG. jbickers@templar.actrix.gen.nz *** *** "Endless variations, make it all seem new" - Devo. ***
dant@ryptyde.UUCP (Daniel Tracy) (06/20/91)
Responding to the following: "I think you must be under Unix(SYSVR2) to take advantage of the I/O coprocessors in the IIfx. System 7 doesn't even utilize them. That killer graphics coprocessor board won't work with System 7.0 for a few more months. However, if you want a real killer graphics board, I would recommend the NeXTDimension board for the NeXT Cube. I haven't seen any benchmarks, but it is using an i860, so I'm willing to be that it's damn fast!" No, the Mac OS takes advantage of the IIfx's I/O chips, just doesn't use the DMA of the SCSI controller. I agree about the NeXTDimension board. Too bad it isn't available for the Mac (there are i860 "cocomputers" available, but no graphics boards yet, I think). I haven't seen any benchmarks done either, but I've read a visual analogy of someone who was working with the board and comparing it with the IIfx. He said with multiple 24-bit image windows up, moving one window from over the other gives the impression that the other image is "just there", with no visual redraw whatsoever. The technical specs of this board are just short of incredible (to quote MacUser). 8 megs (expandable to 32megs) just for the board?! A 33MHz i860, JPEG compression, NTSC conversion with genlock.
dant@ryptyde.UUCP (Daniel Tracy) (06/20/91)
Responding to the following:
"> It has two dedicated I/O
> chips (SCSI DMA too, which can be used with Unix),
That's really going to speed up numbercrunching now, isn't it?"
No, but that 40MHz 68552 FPU that's on the motherboard just might. :-)rjc@wookumz.gnu.ai.mit.edu (Ray Cromwell) (06/20/91)
In article <77@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: >Responding to the following: > >"I think you must be under Unix(SYSVR2) to take advantage of the I/O >coprocessors in the IIfx. System 7 doesn't even utilize them. That >killer graphics coprocessor board won't work with System 7.0 for a few >more months. However, if you want a real killer graphics board, I >would recommend the NeXTDimension board for the NeXT Cube. I haven't >seen any benchmarks, but it is using an i860, so I'm willing to be >that it's damn fast!" > >No, the Mac OS takes advantage of the IIfx's I/O chips, just doesn't use >the DMA of the SCSI controller. I agree about the NeXTDimension board. >Too bad it isn't available for the Mac (there are i860 "cocomputers" >available, but no graphics boards yet, I think). I haven't seen any >benchmarks done either, but I've read a visual analogy of someone who >was working with the board and comparing it with the IIfx. He said >with multiple 24-bit image windows up, moving one window from over the >other gives the impression that the other image is "just there", with >no visual redraw whatsoever. The technical specs of this board are just >short of incredible (to quote MacUser). 8 megs (expandable to 32megs) >just for the board?! A 33MHz i860, JPEG compression, NTSC conversion >with genlock. From what I've heard, the JPEG compression chip has been removed from the design. The board is also incredibly expensive, and it's not out yet. I don't understand its purpose, if it's for live video NTSC is adequate, if it's for rendering/virtual reality why not just buy a SGI? Something like this isn't going to take off in the broadcast video/multimedia market when you can buy a Toaster for $1400. The price of the ND is going to have to come down to compete with the other integrated workstations out there. -- / INET:rjc@gnu.ai.mit.edu * // The opinions expressed here do not \ | INET:r_cromwe@upr2.clu.net | \X/ in any way reflect the views of my self.| \ UUCP:uunet!tnc!m0023 * /
rjc@wookumz.gnu.ai.mit.edu (Ray Cromwell) (06/20/91)
In article <78@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: >Responding to the following: > >"> It has two dedicated I/O >> chips (SCSI DMA too, which can be used with Unix), > >That's really going to speed up numbercrunching now, isn't it?" > >No, but that 40MHz 68552 FPU that's on the motherboard just might. :-) That's 68882. Could you use a different quoting scheme, it's hard to tell sometimes which text is yours and which isn't. Are you using a Mac news reader or something? -- / INET:rjc@gnu.ai.mit.edu * // The opinions expressed here do not \ | INET:r_cromwe@upr2.clu.net | \X/ in any way reflect the views of my self.| \ UUCP:uunet!tnc!m0023 * /
dant@ryptyde.UUCP (Daniel Tracy) (06/21/91)
Responding to the following: "That's 68882." Oops, sorry! :/ "Could you use a different quoting scheme, it's hard to tell sometimes which text is yours and which isn't. Are you using a Mac news reader or something?" I'm sorry, but I'm new to Usenet and Unix. I use a simple Cut & Paste scheme and since I DON'T want to have to Paste in each individual line, the best I can do is to quote it. I'm sure there's a better way to do this (the F command?). I'll try something new. However, it has recently come to my attention during a recent post (trying to guess who is who) that I am recognized by my infamous "Responding to the following" header, and now that I have an identity here, I'm not sure I want to give it up! :-)
kls30@duts.ccc.amdahl.com (Kent L Shephard) (06/21/91)
In article <6ecHr*7?@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991Jun17.233312.9357@menudo.uh.edu> jet@karazm.math.uh.edu (J Eric Townsend) writes: > > > Or put 18 or so megs in your 'miga. Or get a *real* computer: > Sun SparcStation-2. Or an Intel iPSC/860. > >A SSII is about 50% faster than a NeXTstation and quite a bit more SPEC benchmarks have the SSII at about 20-25%. Some things faster some slower. >expensive. And we can't forget that Mathematica is free on the NeXT >and it costs several hundred dollars for the SUN. For education Mathematica is free on the NeXT, others pay. It is somwhere in the $1000 range on a SSII. > >-Mike -- /* -The opinions expressed are my own, not my employers. */ /* For I can only express my own opinions. */ /* */ /* Kent L. Shephard : email - kls30@DUTS.ccc.amdahl.com */
torrie@cs.stanford.edu (Evan Torrie) (06/22/91)
rjc@wookumz.gnu.ai.mit.edu (Ray Cromwell) writes: > From the way I interpret this, the Mac has some form garbage collection >to prevent memory defragmentation. I wouldn't call it garbage collection in the sense of a language like LISP. It's more a memory compactor. Generally invoked either explicitly (_CompactMem), or when a memory request fails because of fragmentation. >I get the suspicous feeling that >the MacOS was developed originally in a language other than C, something >with a run-time memory manager and garbage collector. Am I right? The MacOS was originally developed with Pascal in mind (as the Lisa had been before it). This means it has the general structure of a heap growing upwards (where dynamic memory is allocated), and a stack growing downwards. Heap space can be allocated using a NewPtr() call, which corresponds to a Pascal new() call. That is, you end up with a non-relocatable block in the middle of your heap. Because you can make copies of this pointer (e.g. myptr = thatptr), and pass them around (in fact, you may have 100s of pointers pointing to this block), it's infeasible to allow such a block to move, since to do so would require updating all these pointers throughout memory. The MacOS adds the concept of handles, a ptr to a ptr. The actual ptr to the block is called the master pointer, and there's only ever one of these. You can make as many copies of the handle as you want, and pass it around, since they all point to the master pointer. Such a block can be easily relocated because the OS has to change only the master pointer, and need not concern itself with all the handles. Unlike LISP, the memory manager doesn't have to go around finding unreferenced blocks of memory when it wants more free space. Instead, the program explicitly tells the OS that the block is unreferenced, via a DisposeHandle call. [There are a few more subtleties than this, like making a block able to be purged, or making it unable to be moved [using HLock], but that's the general idea]. So the idea of handles is to allow blocks of memory to be relocated by the OS at any time, yet still let the program (which may have multiple references to this block of memory) be insulated from the actual location of the block. > 1) Double indirection of pointers is slower > (yes you can lock a memory block, but why use handles in the first place? > I know I would be irritated if I had to constantly do a > ptr1->ptr2->mystuff when coding) How about (*ptr)->mystuff, or (**ptr).mystuff? > 2) garbage collection is expensive and it gets worse depending on > the amount of memory you have and the amount being used The Mac memory compaction depends not on the amount being used, but on the number of different blocks being used. If you allocate little blocks for every single object, and then run into low-memory problems, you can see a slow-down in some applications. What you can do of course, is just allocate a big block for your own private use, and then use your own mem. routines to allocate from within that block. > On modern computers an MMU can be used to defrag memory and private >memory pool managing can reduce it greatly. Can someone tell me >why Apple uses this type of memory management? The 68000 didn't have an MMU. How does the Amiga defrag memory? > It still seems strange that you have to 'limit' or tell the OS >the maximum memory your app will need. This is an artifact of the single application model of 1984. The application started its heap in low memory, its stack in high memory, and they grew together until they collided. With multiple applications, the OS has to start the base of the stack somewhere [at less than high memory], and grow downwards from there, to allow other applications to reside in the same address space. The 'limit' is essentially the application writer's best guess as to how much memory his heap and stack need together before they would collide. As Sho says, this limit can be got around by using temporary memory in MultiFinder's unused memory pool. [In System 7, this is better supported and made more robust than in 6]. -- ------------------------------------------------------------------------------ Evan Torrie. Stanford University, Class of 199? torrie@cs.stanford.edu "Lay me place and bake me pie, I'm starving for me gravy... Leave my shoes and door unlocked, I might just slip away - hey - just for the day."
frank@hfsi.UUCP (Frank McPherson) (06/22/91)
In article <83@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: > >I'm sorry, but I'm new to Usenet and Unix. I use a simple Cut & Paste scheme >and since I DON'T want to have to Paste in each individual line, the best >I can do is to quote it. I'm sure there's a better way to do this (the F >command?). I'll try something new. However, it has recently come to my >attention during a recent post (trying to guess who is who) that I am >recognized by my infamous "Responding to the following" header, and now >that I have an identity here, I'm not sure I want to give it up! :-) Which newsreader are you using? I've also had problems following what you're trying to say because of your quoting scheme. All you do to get the quotes like I have them in this article when using the "rn" newsreader is use an uppercase 'F' to followup to it. The command for the nn newsreader may be similar. Please try it sometime. It's much easier to read this way. When you've done that, you just delete the lines you don't feel the need to respond to, and away you go. (Deleting lines is good. It's annoying to have to read an article fifty bizillion times because people include the whole thing when they're replying to it. -- Frank McPherson INTERNET: emcphers@manu.cs.vt.edu --
hunter@phoenix.pub.uu.oz.au (James Gardiner [hunter]) (06/23/91)
In <1991Jun17.233312.9357@menudo.uh.edu> jet@karazm.math.uh.edu (J Eric Townsend) writes: >In article <fv5H7bfy@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: >> >>In article <1991Jun17.142339.21049@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes: >>>you put in 9MB you will probably never run out. And yes, it has a >>Memory is cheap, but Mathematica can use more than 9MB of RAM. Get a >>computer with virtual memory. >Or put 18 or so megs in your 'miga. Or get a *real* computer: >Sun SparcStation-2. Or an Intel iPSC/860. Or even better a new HP 720 snake.. Same price as SS2 and 2.5 times faster in real time testing. ie side by side running same code with full optimization on both. James -- James Gardiner [Hunter]. System Admin, Public Access UNIX Melbourne, Australia PubNet: phoenix!hunter | (voice)+613-532-8030 (data)+613-523-9865&+613-532-8029 Internet: hunter@phoenix.pub.uu.oz.au | PO BOX 54 Chadstone Centre UUCP:..!uunet!munnari!labtam!eyrie!phoenix!hunter | Melbourne Australia 3148
dant@ryptyde.UUCP (Daniel Tracy) (06/24/91)
Responding to the following: "Which newsreader are you using? I've also had problems following what you're trying to say because of your quoting scheme. All you do to get the quotes like I have them in this article when using the "rn" newsreader is use an uppercase 'F' to followup to it. The command for the nn newsreader may be similar. Please try it sometime. It's much easier to read this way. When you've done that, you just delete the lines you don't feel the need to respond to, and away you go. (Deleting lines is good. It's annoying to have to read an article fifty bizillion times because people include the whole thing when they're replying to it." That's precisely why I usually don't use the 'F' command to follow it up. I would, but I'm new to these editors and don't have a list of commands anywhere (I'll get them), so I don't know how to delete lines.
rhealey@kas.helios.mn.org (Rob Healey) (06/26/91)
In article <4ocHh98?@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: >In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: >>>>> "Even the IIfx isn't as fast as a 3000/25." > That's bull. The IIfx will blow the 3000 away. It has two dedicated I/O > chips (SCSI DMA too, which can be used with Unix), and a killer graphics > coprocessed board (the 8-24GC) (which was released with the IIfx). I doubt > you've done any kind of serious benchmarks. > Huh? The speeds that count are to memory and disk. The 3000 has high speed 32 bit paths to both. The SCSI controller is seperate from the other main busses so other I/O should take place at the same time. Better yet, BOTH OS's can take advantage of the speed. From what I've heard, the SCSI controller and SCSI bus on the 3000 can easily handle the highest performance SCSI peripherals available today and still have spare bandwidth. If a graphics board can't currently be used with System 7.0 than it doesn't count, any more than me tring to use the nonexistant A2410 board as an excuse to say other video boards are pathetic. When the 2410 IS finally released, it's TI chipset should provide a VERY impressive video subsystem! -Rob
judge@alchemy.ithaca.ny.us (rory toma) (06/27/91)
rhealey@kas.helios.mn.org (Rob Healey) writes: > In article <4ocHh98?@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) wri > >In article <63@ryptyde.UUCP> dant@ryptyde.UUCP (Daniel Tracy) writes: > >>>>> "Even the IIfx isn't as fast as a 3000/25." > > That's bull. The IIfx will blow the 3000 away. It has two dedicated I/O > > chips (SCSI DMA too, which can be used with Unix), and a killer graphics > > coprocessed board (the 8-24GC) (which was released with the IIfx). I doub > > you've done any kind of serious benchmarks. > > > Huh? The speeds that count are to memory and disk. The 3000 has > high speed 32 bit paths to both. The SCSI controller is > seperate from the other main busses so other I/O should take > place at the same time. Better yet, BOTH OS's can take advantage > of the speed. From what I've heard, the SCSI controller and > SCSI bus on the 3000 can easily handle the highest performance SCSI > peripherals available today and still have spare bandwidth. > > If a graphics board can't currently be used with System 7.0 than > it doesn't count, any more than me tring to use the nonexistant > A2410 board as an excuse to say other video boards are pathetic. > When the 2410 IS finally released, it's TI chipset should > provide a VERY impressive video subsystem! > > -Rob I've also read in MacWeek, that the biggest complaint about the FX was the bottleneck at the I/O and disk, particularly the disk. Theyuse 10 MHz 6510's, I believe. It further stated that people were buying II ci's because of this... rory