dvljhg@cs.umu.se (J|rgen Holmberg) (12/08/90)
I just thought this could be interesting for the comp.sys.next group. In article <673@storm.UUCP> bostrov@storm.UUCP (Vareck Bostrom) writes: >In article <9012040152.AA09539@ucbvax.Berkeley.EDU> 91_bickingd@GAR.UNION.EDU ("Bicking, David") writes: >>>WROONG. SLC is much slower than a 68040-NeXT. NeXTstation is supposed to >>>be even faster than a SparcStation 1+. >> >>I wanna see specs before I believe that, especially considering the slowness of >Ok, I am assuming that the SPARCstation-1 and the SLC are the same speed. >I have tested a SS-1 but not a SLC, and I have also tested a SS-1+ and an >SS-2, as well as 4/490's 4/390's, and 4/330's. In Floating point, the >NeXT easily beats the Sparcstation-1, the 1+, and the 4/3xx and 4/4xx. >In integer (almost useless, but anyway) the NeXT manages to tie almost >exactly the ss-1+. Here are some marks: > >Machine Dhrystones MFlops >NeXT 030/25mhz 7204 >Sun 4/60 (ss-1) 19672.1 (max opt) >Sun 4/65 (ss1+) 22140.2 (max opt) >NeXT 040/25mhz 23661.0 (max opt) >Amiga (no sure) 3350.7 (opt???) >IBM 3090/200 54894.8 Expected at 800 with vector. >Amiga (030/25) 6990.3 >Sun 3/60 3790.3 >rs6000/520 28222.0 >Sun 4/330 25000.0 (max opt) >Ardent titan-3 22831.1 >HP 9000/845 20833.3 >Gould np-1 13730.8 (max opt) >symmetry-81k 5696.9 >Hp 9000/310 948.0 (max opt -- took an hour to compile) > >Anyway, those are figures that I (or a friend of mine) have tested >ourselves. I am quite sure that a Cube-040 can match a ss-1+ in integer >performance, and because of it's cisc nature, I believe that the 040 >cube will handle load better than the ss-1+ will also. THESE ARE JUST >GUESSES. I havent been on an 040 next much, but I have been on 1+'s, >and I will admit that they FEEL fast. The NeXT has to work on that. >The only experience I've really had with the ami is games and a bit >of programming, but those machines crash more often than...anything. >I would love to try an 040/25 amiga running under sys-vr4, but from >what I have seen, they seem to buckle under load much more than the >NeXT or the ss-1. Of course, the ss-2 will blow away anything except >the rios-risc 6000. I hate AIX, though. > >Note I say (max opt) on a lot of them, on the suns that means -O4 or WHATEVEr >we got the best time out of. The SPARCstations would drop to 9000 dhrys with >no optimization. Under ami/os, I had major problems getting ANYTHING >to compile, and it reminded me much of a macintosh in that respect. >Almost all of the ami's I have seen or delt with had poor res screens, >major flicker, or other problems. Going from a 1+GX to a 68000 ami is a >major headache, even if i go to the ami just to play games. Graphics >on the ami do seem very fast for such a cheap machine, though I have seen >suns with graphics accellerators that will blow an ami away (admittidly >those accellerators are $6k plus). > >Anyway, the 040 NeXT is as fast as a SS-1+. I expect that an 040 ami >will be close, also. >>the '030 NeXT (as opposed to the speed of the A3000). >> >>>Note: When comparing the A3000UX and the NeXTstation you should remember >>>that the latter has a 68040 and a motorola DSP. The latter doesn't have >>>color in it's cheapest configuration, though. >> >>When comparing the two you also have to realize the A3000UX can be easilly >>expanded in MANY ways, especially the CPU. More color, more sound, higher >>resolution, all can be added easilly. Soon, when C= finishes their device >>independant libraries, all those expansions will be universally accessible. > >True. I have seen a VERY expanded amiga, and it was very impressive. It >did seem to lack the cpu uumph of the ss-2 or a mips magnum, but it was >moving along. > >The NeXTcube with NeXTdimention and monitor should be impressive also, >68040 CPU, 80860/33 graphics chip (gods, thats 66 MFLOPS (acc. to intel)). > >I couldn't see actual heavy duty sci work being done on an ami/040, as >that is getting close to your maximum cpu power, and you have to expand the >graphics by a power of 10 to get workstation quality (IM TALKING res, not >color. 4096 isn't bad). Also I haven't found any good optimization compilers >for the Ami, and this leaves many of the programs that scientists run >dead in the water. (when i say scientists, I mean scientists/engineers, etc.). >Also, everywhere I go where there are workstations installed, I see Suns, >NeXT's (on occasion), MIPS, DECs, IBM rs/6000's, but NEVER any amigas. I >wonder why? Well, as of now, there is no Ami/Unix (though I have heard and >seen samples of r4 on an ami) there is no 040 ami (once again, ive seen >samples), mid to poor system through put, etc. An 030 machine just isn't >a serious workstation, paticulary without unix. The NeXT IS a unix workstation, >though up till receintly there has been no 040 for it, all along it has >run unix. > >Anyway, both have their uses. I LOVE the amiga to play games on. It is without >a doubt the most devoloped games machine ever. It is fast (for a game machine), >and can be used for other things besides games. > >So...the end. >> >>> >>> Jouni >>-- >>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=- >>Dave Bicking Single Tasking????? Just say NO!!!! >>Union College Box 152 91_bickingd@union.bitnet // >>Schenectady, NY 12308 91_bickingd@gar.union.edu \X/ Amiga >>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- > >Vareck Bostrom >Reply to: bostrov@mist.cs.orst.edu > >Flames to: erica@poux.cs.uoregon.edu You have seen amigas and played games on them. Fine, now use one before you come back to comp.sys.amiga to give us facts that are irrelevant to the amiga and that a great many people on comp.sys.amiga has been asking for days now not to have to expand their kill file for. If we want to find out more about the next I am sure we can g comp.sys.next. OK? -- ******************************************************************************* email dvljhg@cs.umu.se - other ways to communicate are a waste of time. Everything I say is always true, just apply it to the right reality. "Credo, quia absurdum est."