#No 68060?
21 messages in this thread
Speaking of the '060, have you heard a rumor that they(motorola) won't
be releasing it after all because it could outperform a powerMac w/ native
apps, when put in an old quadra running 680×0 native software.
I heard it and am hoping it is not true. But I am afraid it is. 🙁
Matt Straney, 3 hours late, without your luggage
from Grand Ledge — on AutoPilot ——–>
Lemme see…
Motorola has a new chip that is really a killer, but the won't be
releasing it because it is too good?
Nope, ain't buying that one. Sounds like the ol' 200 mpg carburator
story to me.
Capitalists being what they are, the chip would definetly see the
light of day. About the only reason Motorola would not release the
chip, is that they have decided that the market just isn't there.
– Alex… Celebrating 10 years on Compuserve
I doubt they would not release it for that reason, though I wouldn't be
suprised that an 060 could run 680×0 native apps faster than a PowerPC can.
"I wouldn't be suprised that an 060 could run 680×0 native apps faster
than a PowerPC can"
Neither would I. A PowerPC would need an emulator to run the 680×0
apps – and that's an anchor. As it is, even the 80Mhz PowerPC is
terribly slow at running 680×0 apps in emulation.
But well-written native PPC code, I think, would beat the pants off
the 68060.
Jerry
> Speaking of the '060, have you heard a rumor that they(motorola)
> won't be releasing it after all because it could outperform a
> powerMac w/ native apps, when put in an old quadra running 680×0
> native software. I heard it and am hoping it is not true. But
> I am afraid it is. 🙁
That is very untrue.
The '060 is shipping in limited quanities at 50Mhz. It is due to ship at 66Mhz
in Q1 95. The delays in shipping come from two reasons:
1) Moto has had a lot of trouble getting the clocks up on the 68K
family since the '040 days. Remember how late the '040 was? Remember
how long the clock was stuck at 25Mhz? There was no PPC chip then to
explain it all away with. 🙂
2) Moto has probably shifted a lot of R&D money/people into PPC.
That's were the money is. Not politics, just money.
What the PPC offered to Moto was the chance to start over with a clean slate.
It let them dump a lot of baggage that the 68K family brings. Things like:
1) Addressing modes which no one uses, but consume chip resources.
2) Variable length instructions, which slow decode and make the instruction
fetch more complex.
3) Primitive condition code handling (ALU flags), which provide bottle necks to
super scalar execution.
These things are not to say the 68060 is not a really good chip, but it's too
little, too late for it.
As for the claim that the 68060 runs software faster than the PPC runs it's
native software, that is simply not true. At the same clock (50Mhz) both chips
are similar for integer, with the PPC slightly ahead. The PPC _smokes_ the
'060 for floating point work.
But here's the kicker…. Nobody uses 50Mhz PPC's!!!!! While the chips are
comparable at the same clock, the PPC601 is shipping at 100Mhz today. The
_low_ end in Mac's is 60Mhz, next month the low end goes to 66Mhz, in May it
goes to 80Mhz. Today the high end is 110Mhz! Bu may that will be 120 or
130Mhz.
It's just easier to get the clock frequencys up on the PPC then the 68K, so
this disparity will continue to grow.
As a further example, Moto started shipping the next generation PPC two months
ago. The PowerPC 604 is 60% faster than the 601 at the same clock. This means
that it's lowest speed of 100Mhz it is as fast as a 160Mhz 601.
That would mean that a '060 would have to be about 160Mhz also to be
comparable. I doubt you will ever see one that fast. That puts that chip at
about 3 times faster than the shipping '060.
Now the same failings that apply to the '060 apply to the PPC. In ten years
(about the life of an architecture), PPC supports will disbelive that the "new"
generations of chips are better and it will take two or three years to prove
that the PPC is just old news.
Oh well. Could be worse, could be raining….
—– Andrew
Well, like I said rumor. I didn't think the numbers were right when it
was being explained to me. He was telling me that the `060 was 4x as fast
as an `040, which is true at 50mhz vs 25mhz respectivly; but high end
mac's use 40mhz `040s so that kinda throws it.
I am still hoping to buy an 060 for my amiga if it isn't a crusher, it
will be awhile I'm sure. I'll have to let the prices come down.
Matt Straney, 3 hours late, without your luggage
from Grand Ledge — on AutoPilot ——–>
> I didn't think the numbers were right when it
> was being explained to me. He was telling me that the `060 was
> 4x as fast as an `040, which is true at 50mhz vs 25mhz
> respectivly; but high end mac's use 40mhz `040s so that kinda
> throws it.
A 50Mhz '060 ought to be about 4x a 25Mhz '040 in a very good
case. 3-4x sounds good.
I don't think there are any '040 Macs shipped higher than 33Mhz.
> I am still hoping to buy an 060 for my amiga if it isn't a
> crusher, it will be awhile I'm sure. I'll have to let the
> prices come down.
The prices are still a bit high, but it is a good chip. I just don't think
there is a long term future in it.
But it is a godsend to the Amiga world because it can double the speed
of peoples machines at a decent price.
—— Andrew
Doesn't the 840AV use a 40Mhz 040 ???
-Jonathan
// Amiga online via Autopilot. MAC Quadra 900 and
\X/ 486 emulation. E-Mail: jonathanb@delrina.com
>Doesn't the 840AV use a 40Mhz 040 ???
I don't know for sure, but you may just be right. I work on a 33Mhz Mac at
work, and come to think of it, I've known in the back of my mind that the 840AV
was faster.
I'd have to check my reference chart, which is at work.
Either way it's faster than the long discarded, 16Mhz, 68000, 128K Mac I bought
for an obscene amount of money in 1984!
—– Andrew
Going to the Warp Engine (40 mhz '040) doubled the speed of my 4000 in most
cases and quadrupled it in some others.
-sja
Interesting thing about the Warp, Steve. It seems a lot of the Warp's
advantage in speed is that is has its own memory bus and beats the Amiga 4000's
usual slow memory access by a lot. I checked FractalPro on the Warp40mhz and
found it was exactly 1.6 times as fast as a stock A4000/040 (exactly the ratio
of the clock speeds). Interestingly, FractalPro is assembly language with all
the fractal calculation routines tighly coded to run entirely within the cache
so the Warp's apparent advantage on other software boils down to only a clock
speed advantage for efficient software. The Warp40mhz is a great product but
it doesn't do anything magical except reveal how much faster most software
could be if the critical calculation parts were written in assembler instead of
high level languages
If programmers took the trouble to optimize and tighten up their
calculation-intensive programs, we'd get more of a performance boost from a
software revision than from a Warp! I'm also a speed demon so every increase
in clock speed is important to me, too. But I'm not relying on faster
processors or better memory access to cover a multitude of programming sins.
Better programming would give everyone a big speed boost without having to rely
on expensive new hardware. The 060 will bring a real new performance increase
in excess of the clock speed increase to 50mhz so I'm definitely looking
forward to seeing FractalPro on the 060!
You'd be suprised how much optimization an intellegent 'C' programmer (and a
optimizing compiler) can squeeze into a critical section. The problem is m re
a lack of profiling time-important code than one of which language is use
From what I saw of Amiga C compiler output a while back I'd have to
doubt your optimism on 'fixing' C programs, but I've heard the SAS
compiler has come a long way so I hope you're right. Profiling
should be fairly obvious to many programs, but isn't. I'd suggest
just examining loops with lots of computation in 'em! At least
experimenting with some table lookups, etc. is worth a try and a few
experiments might reveal big speedups without much work. I think many
programmers 'quit' when the code works at all. Any program which
does computation-intensive stuff like rendering is a worthy subject
for profiling and optimization. It would save Amiga owners time and
money spent on faster clock-speed accelerators. I do think much is to
be gained by improving the programs. I know specific programs, the
programmers of which have told me they'd 'like to get a look back at
that loop….' or they 'wish they had done that loop in assembly' or
some such thing. Sad thing is they never do. More features get
added and those who need performance have to go buy accelerators, but
the original program could easily have been 50% to 200% faster with a
bit more work.
Most of the AmigaOS is written in SAS C. The OS engineers both helped
improve the compiler with their recommendations but also optimized or
downcoded whereever appropriate. Since 1.3, while moving more of the
OS to native 'C', they also made it siginificantly faster.
Your point that most programmers aren't as well motivated re:
performance is well taken.
Dan,
The statement, referring to the '060, I responded to was …
>> But it is a godsend to the Amiga world because it can double the speed
>> of peoples machines at a decent price.
You can do that now with the Warp Engine … and yes, it is because it has t's
own memory buss and corrects the bugs Commodore left in the A4000 desig
If you've got a stock A4000/040, you can, right now, increase its speed b
75%-400% (according to AIBB and other performance checkers) by purchasing t e
Warp Engine 4040 … you don't need to wait for an '060 board. An '060 bo rd
will undoubtably give you even more speed. Again, it's a trade off … n w vs
later.
-sja
I've got an A4000/030 with 18 megs and a Toaster 4000 in it. It's
the old Buster, and sometimes it takes several cold boots to get it
to start. Sadly, this system has been used mostly by the
receptionist as a database computer for the Syndesis mailing list,
but now that task has been moved to the PC, and I'd hate to see this
system go to waste.
My every-day development system is an A2000 with 18 megs, a GVP
040/33 with SCSI and a Toaster 2000.
What are my options for upgrading this A4000 to an '040/33 or better,
with a newer Buster, and very reliable SCSI (for Exabyte, DAT and
CDROM)? Are new Busters quite rare? Which 040 card makers are still
standing? How much might this cost, and can anyone recommend a
dealer? I'll need one of those little adapters to make the A40000
video work with a multi-sync and a Toaster, too.
Eventually, I'd like to add a Flyer to the system, too. That leaves me one
free slot for an Ethernet card, right?
Personally, I'd do exactly what I suggested … the Warp Engine. You get a 40
mhz '040, SCSI II, who cares about Buster (you don't need it), and no slowness
due to Commodore's poor memory management on the 4000.
>> … but now the task has been moved to the PC
John,
just another one bites the dust, and as in most cases, I suspect because
there is still no real quality office software for the Amiga.
Or may I be wrong ?
– wkc – … via AP from Hamburg, Germany
>Sadly, this system has been used mostly by the receptionist as a
>database computer for the Syndesis mailing list
Trying to beat the postal increases, I finally got around to using
your lists of user's groups and dealers. I will let you know about
bounces. You can eliminate PAUG/Joe Molica in Philly — the group
disolved a few months ago.
I have AmigaWorld's list of User's groups if you want that, though I
didn't really merge it with your list (there is overlap I tried to
catch by hand).
PMFJI –
Just want to make sure a list of dealers includes Videology, Inc., 72-6
Newtown Road, Danbury, CT. 06810. (203) 270-9000, and Fax (203) 426-9888.
Hope you had a Happy New Year.
Hank
Videology, Inc – Where you get all the P.O.M.P. under all
Circumstances
About the '060/PPC deal- it seems that everyone is missing an
obvious(although expensive) alternative. Why not redesign the Amiga
to use a PPC rather than 68's ?? The increase in speed would be
outstanding!! I realize that the OS would have to be rewritten, but
hey, it would be worth it. Think of motherboard with two PPC chips,
one running Amiga OS, the other running whatever. The Amiga
chipset(with the dedicated Amiga PPC) could imitate whatever the
'other' OS needed…