melling@cs.psu.edu (Michael D Mellinger) (06/13/91)
In article <1991Jun12.152204.23497@usenet.ins.cwru.edu> pab@po.CWRU.Edu (Pete Babic) writes:
So will a Mac IIfx, it'll probably look toy-like long before the
A3000. No the A3000 will not be the last computer we will own, but
also, the A3000 will not be the last computer Commodore will ever
sell. I'm sure the next generation Amiga will be just as impressive
when it comes out as the A3000 was during its introduction.
Besides, the reason someone posted that the A3000 can be expanded
to 1.8GB is because a Macoid didn't even believe the A3000 could be
expanded to 128MB. If its usefull or not was not relevant to the
discussion.
The IIfx is already a toy :-). Actually, read comp.sys.next. It
seems that the new NeXT due out early next year will use the new 88K
and run at around 50 to 60 mips(the current machine is 12-15 mips).
To put this in perspective, that's around 10 times faster than the
68030. It looks like NeXT will have it's graphic's coprocessor and
CPU on one chip. Anyone still unconvinced that NeXT isn't on the
move!?!
-Mike
rjc@geech.gnu.ai.mit.edu (Ray Cromwell) (06/13/91)
In article <mx5H+2&t@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >The IIfx is already a toy :-). Actually, read comp.sys.next. It >seems that the new NeXT due out early next year will use the new 88K >and run at around 50 to 60 mips(the current machine is 12-15 mips). >To put this in perspective, that's around 10 times faster than the >68030. It looks like NeXT will have it's graphic's coprocessor and >CPU on one chip. Anyone still unconvinced that NeXT isn't on the >move!?! Mike, don't compare CISC MIPS to RISC MIPS. Let's say RISC processor X is rated at 10x the MIPS of CISC processor Y, but a divide instruction takes 20 instructions on X and one instruction on Y. It seems to me that division is twice as slow on X. Show me the SpecMarks. BTW, isn't Motorola already wrking on the next generation 88k's ? >-Mike > -- / 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 * /
dkoski@cs.arizona.edu (David A. Koski) (06/13/91)
From article <1991Jun12.190315.18622@mintaka.lcs.mit.edu>, by rjc@geech.gnu.ai.mit.edu (Ray Cromwell): > Mike, don't compare CISC MIPS to RISC MIPS. Let's say RISC processor > X is rated at 10x the MIPS of CISC processor Y, but a divide instruction > takes 20 instructions on X and one instruction on Y. It seems to me > that division is twice as slow on X. Show me the SpecMarks. Dumm dumm dumm. There are no such things as CISC MIPS and RISC MIPS. Although the definition of MIPS is millions of instructions per second, no one (except maybe those IBM dorks) uses them that way. For any decent comparison you have to compare the work done, not the instructions executed. Many times dhrystone will be used to get mips ratings by comparing times to the vax 11/780 (a 'one mips machine'). I have seen specmarks for the 88000 and it was faster than the 040, but I don't think anyone has a 88110 yet. Sorry. > BTW, isn't Motorola already wrking on the next generation 88k's ? Yes, that is what Next is 'planning to use' (the plans to use the 88k chip are just rumors) David Koski
rjc@geech.gnu.ai.mit.edu (Ray Cromwell) (06/13/91)
In article <1566@caslon.cs.arizona.edu> dkoski@cs.arizona.edu (David A. Koski) writes: >From article <1991Jun12.190315.18622@mintaka.lcs.mit.edu>, by rjc@geech.gnu.ai.mit.edu (Ray Cromwell): >> Mike, don't compare CISC MIPS to RISC MIPS. Let's say RISC processor >> X is rated at 10x the MIPS of CISC processor Y, but a divide instruction >> takes 20 instructions on X and one instruction on Y. It seems to me >> that division is twice as slow on X. Show me the SpecMarks. > >Dumm dumm dumm. There are no such things as CISC MIPS and RISC MIPS. >Although the definition of MIPS is millions of instructions per second, >no one (except maybe those IBM dorks) uses them that way. For any >decent comparison you have to compare the work done, not the instructions >executed. Many times dhrystone will be used to get mips ratings by >comparing times to the vax 11/780 (a 'one mips machine'). Dumm? Perhaps you're too dumm to get the point. Don't compare MIPS, they are meaningless especially when trying to compare different archtectures. The phrase "don't compare cisc mips to risc mips" means "don't compare mips ratings of cisc processors to mips ratings of risc processors" not that there is seperate definitions for each architecture. >I have seen specmarks for the 88000 and it was faster than the 040, but I >don't think anyone has a 88110 yet. Sorry. But how much faster? I hardly think it is 10x faster as the original NeXToid tried to state. -- / 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 * /
torrie@cs.stanford.edu (Evan Torrie) (06/13/91)
rjc@geech.gnu.ai.mit.edu (Ray Cromwell) writes: >In article <mx5H+2&t@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: >> >>The IIfx is already a toy :-). Actually, read comp.sys.next. It >>seems that the new NeXT due out early next year will use the new 88K >>and run at around 50 to 60 mips(the current machine is 12-15 mips). > Mike, don't compare CISC MIPS to RISC MIPS. He's not. He's comparing VAX MIPS to VAX MIPS. >Show me the SpecMarks. Actually, as people have found out, SpecMarks have a surprisingly close correlation to VAX MIPS. Let's just say that a current 25MHz 88K machine has a 17.3 SpecMark, and the 88110 has been promised [by Motorola] to be at least 3 times as fast. >BTW, isn't Motorola already wrking on the next generation 88k's ? Exactly, the 88110 which is what we're talking about. -- ------------------------------------------------------------------------------ Evan Torrie. Stanford University, Class of 199? torrie@cs.stanford.edu "If it weren't for your gumboots, where would you be? You'd be in the hospital, or in-firm-ary..." F. Dagg
melling@cs.psu.edu (Michael D Mellinger) (06/13/91)
In article <1991Jun12.202001.22613@mintaka.lcs.mit.edu> rjc@geech.gnu.ai.mit.edu (Ray Cromwell) writes:
But how much faster? I hardly think it is 10x faster as the original
NeXToid tried to state.
No, the NeXToid said that the new 88K(50-60mips) is 10 times faster
than a 68030 25MHz 68030(if we are assuming that it's 5 mips. Someone
said that the Amiga's does 7 mips.).
-Mikechucks@pnet51.orb.mn.org (Erik Funkenbusch) (06/15/91)
melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991Jun12.202001.22613@mintaka.lcs.mit.edu> rjc@geech.gnu.ai.mit.edu (Ray Cromwell) writes: > > > But how much faster? I hardly think it is 10x faster as the original > NeXToid tried to state. > >No, the NeXToid said that the new 88K(50-60mips) is 10 times faster >than a 68030 25MHz 68030(if we are assuming that it's 5 mips. Someone >said that the Amiga's does 7 mips.). > >-Mike Actually those mip ratings are very misleading. sure, a RISC processor will get more mips, but then it will take more clock cycles to do a complex instruction than the current 040, because there is a smaller base set of instructions to work on. in it's best form (i'm assuming simple block moves) a RISC chip blows the pants off a CISC, but when you get down to real processing, the effective mips drops. .--------------------------------------------------------------------------. | UUCP: {amdahl!tcnet, crash}!orbit!pnet51!chucks | "I know he's come back | | ARPA: crash!orbit!pnet51!chucks@nosc.mil | from the dead, but do | | INET: chucks@pnet51.orb.mn.org | you really think he's | |-------------------------------------------------| moved back in?" | | Amiga programmer at large, employment options | Lou Diamond Philips in | | welcome, inquire within. | "The First Power". | `--------------------------------------------------------------------------'