Note that both PSX and N64 are MIPS based consoles,
with PSX using a MIPS III processor(R3000A) and
Nintendo using a MIPS IV processor(R4300i). The next
MIPS architecture is MIPS V, on which PSX2 and N128
will be built up on.
MIPS V accelerates geometry processing performence
two fold by allowing floating point unit to handle
two 32bit floating point per cycle. That is too bad
since Hitachi's SH-4 handles four, not two, 32bit
floating point prt cycle.
So what does this mean? This means MIPS V processors
will not be able to match SH-4's 32bit floating point
processing performence, crucial in geometry processing
and light sourcing calcultations. Simply, MIPS V SUX,
AND SO DO MACHINES BASED ON IT, NAMELY PSX2 AND
NINTENDO128.
I can't wait to see Saturn2 crushing PSX2 and N128
like empty Coke cans......
Nice. Can you get me winning lottery numbers from the future as well?
I'm getting bored of my job.
> Note that both PSX and N64 are MIPS based consoles,
> with PSX using a MIPS III processor(R3000A) and
> Nintendo using a MIPS IV processor(R4300i). The next
> MIPS architecture is MIPS V, on which PSX2 and N128
> will be built up on.
Interesting. You've talked to Sony and Nintendo executives about what
they're putting in their next generation machines? Have you thought
about selling this information to the magazines for money?
SNES to N64 was a complete change of processor. There is absolutely
no reason for Nintendo to stick with MIPS if they think they can get
a better processor for less money elsewhere. Its not like there is
this backwards compatibility issue.
Likewise, PSX is in the same boat. Indeed, Sony could easily make the
leap to some other chip, if Nintendo continues to stay in bed with
SGI so closely. But all of this is besides the point.
> MIPS V accelerates geometry processing performence
> two fold by allowing floating point unit to handle
> two 32bit floating point per cycle. That is too bad
> since Hitachi's SH-4 handles four, not two, 32bit
> floating point prt cycle.
Assuming your technical specs are accurate, its still besides the
point. SPECmarks are not based on instructions per clock cycle or
instructions per clock cycle times Mhz, they are based on the total
system performance of a machine.
How much cache is used? What sort of prediction algorithms are used
for loading it and how well do they deal with computer games? What
is the speed of the motherboard? What sort of pipeline is there
between RAM and CPU?
The thing is, those extra instructions per clock cycle make the
concept of bandwidth between CPU and memory much more critical. If
the CPU can't get enough instructions per clock cycle, that extra
chip space is wasted.
For that matter, we'll assume the chips aren't working by themselves.
What other chips are working and how well do they perform? These days
dedicated graphics chips are doing the lion's share of work. I would
be more worried about specs behind the graphics processors.
> So what does this mean? This means MIPS V processors
> will not be able to match SH-4's 32bit floating point
> processing performence, crucial in geometry processing
> and light sourcing calcultations. Simply, MIPS V SUX,
> AND SO DO MACHINES BASED ON IT, NAMELY PSX2 AND
> NINTENDO128.
Interesting. You're speaking of these machines in the present tense.
Do you have more specs on the PSX2 and Nintendo128? I'd like to see
them, thank you.
> I can't wait to see Saturn2 crushing PSX2 and N128
> like empty Coke cans......
Well, at least you're not frothing at the keyboard over Saturn crushing
PSX and N64. "Just wait till next generation," can be your battle cry
at this point.
--
Martin Terman, Therapy and Behavioral Counseling for Troubled Computers.
Disclaimer: Sticks and stones may break my bones, but flames are just ignored
email: mfte...@access.digex.com home page: http://access.digex.net/~mfterman/
"Sig quotes are like bumper stickers, only without the same sense of relevance"
>You should worry that the next Sony and Nintendo
> machines won't be able to compete with Sega's next
> machine, because their CPU sucks hard.
>
> Note that both PSX and N64 are MIPS based consoles,
> with PSX using a MIPS III processor(R3000A) and
> Nintendo using a MIPS IV processor(R4300i). The next
> MIPS architecture is MIPS V, on which PSX2 and N128
> will be built up on.
>
> MIPS V accelerates geometry processing performence
> two fold by allowing floating point unit to handle
> two 32bit floating point per cycle. That is too bad
> since Hitachi's SH-4 handles four, not two, 32bit
> floating point prt cycle.
>
> So what does this mean? This means MIPS V processors
> will not be able to match SH-4's 32bit floating point
> processing performence, crucial in geometry processing
> and light sourcing calcultations. Simply, MIPS V SUX,
> AND SO DO MACHINES BASED ON IT, NAMELY PSX2 AND
> NINTENDO128.
>
> I can't wait to see Saturn2 crushing PSX2 and N128
> like empty Coke cans......
uhhh, where do I start? Have you lost it. Enjoy playing the systems that
are out now. Stop looking at the future so much. Or are you still sore
that the saturn was crushed by the n64 and psx, that you have nothing better
to post.
> You should worry that the next Sony and Nintendo
> machines won't be able to compete with Sega's next
> machine, because their CPU sucks hard.
>
> Note that both PSX and N64 are MIPS based consoles,
> with PSX using a MIPS III processor(R3000A) and
> Nintendo using a MIPS IV processor(R4300i). The next
> MIPS architecture is MIPS V, on which PSX2 and N128
> will be built up on.
>
> MIPS V accelerates geometry processing performence
> two fold by allowing floating point unit to handle
> two 32bit floating point per cycle. That is too bad
> since Hitachi's SH-4 handles four, not two, 32bit
> floating point prt cycle.
>
> So what does this mean? This means MIPS V processors
> will not be able to match SH-4's 32bit floating point
> processing performence, crucial in geometry processing
> and light sourcing calcultations. Simply, MIPS V SUX,
> AND SO DO MACHINES BASED ON IT, NAMELY PSX2 AND
> NINTENDO128.
>
> I can't wait to see Saturn2 crushing PSX2 and N128
> like empty Coke cans......
u cant wait cause your sick of saturn getting dogged right now
Wrong, the R3000 has the MIPS II ISA, and the 4300i has the MIPS III
ISA. See http://www.mips.com for details.
> MIPS V accelerates geometry processing performence
> two fold by allowing floating point unit to handle
> two 32bit floating point per cycle. That is too bad
> since Hitachi's SH-4 handles four, not two, 32bit
> floating point prt cycle.
This is a pretty hilarious post. I suggest you take a class in
Computer Architecture before talking about microprocessor
performance. MIPS V provides two 32-bits float operation per
INSTRUCTION, not per cycle. In other words, this is an ISA feature,
whereas what you're mentioning about the SH-4 is an implementation
detail. A MIPS V chip could very well have 4 execution units as well,
if which case it could run 8 fps per cycle. Besides, Hitachi's chips
run at much lower clock speeds.
Next time, come up with SPEC numbers and we'll have a good laugh.
-denis
leroy @ cs.stanford.edu
-------------------------------------------------------
disclaimer: not the opinions of my employer, etc...
Nice theory, but it doesn't hold water. Saturn 2 will never be since
Segabandai will abandon all hardware development after the 64X flops
badly due to no third party support. Expect an arcade perfect
conversions of VF4 on the PSX2 to be released 11-27-98. My contacts
within AM2 already have confirmed this. I can hardly wait....
>You should worry that the next Sony and Nintendo
> machines won't be able to compete with Sega's next
> machine, because their CPU sucks hard.
All well and good but there are alot of other things that are
important. Like 3rd party support.
Sure both the N64 and PSX could use additional processing power but
I've heard they are more in need ofl RAM. 2M is just not enough to
store all the textures and polygons and is a major limitation.
I'd love to see consoles keep moving forward. But for now I'm
enjoying my PSX and N64.
Nobody's Perfect <nob...@nowhere.com> wrote in article
<334226...@nowhere.com>...
> So what does this mean? This means MIPS V processors
> will not be able to match SH-4's 32bit floating point
> processing performence, crucial in geometry processing
> and light sourcing calcultations. Simply, MIPS V SUX,
> AND SO DO MACHINES BASED ON IT, NAMELY PSX2 AND
> NINTENDO128.
>
> I can't wait to see Saturn2 crushing PSX2 and N128
> like empty Coke cans......
>
You are really reaching...... Even if the SH4 FP unit truley does out
perform the MIPS V, 2 to 1, your argument overlooks one major thing. The
bottle neck in the N64 and the PSX is not in the floating point unit of
the CPU, but in the rasterization hardware (display processor). This
hardware is custom hardware, and no one has any way of knowing who will
come up with the fastest polygon renderer.
>Nobody's Perfect wrote:
>>
>> You should worry that the next Sony and Nintendo
>> machines won't be able to compete with Sega's next
>> machine, because their CPU sucks hard.
>>
>> Note that both PSX and N64 are MIPS based consoles,
>> with PSX using a MIPS III processor(R3000A) and
>> Nintendo using a MIPS IV processor(R4300i). The next
>> MIPS architecture is MIPS V, on which PSX2 and N128
>> will be built up on.
>Wrong, the R3000 has the MIPS II ISA, and the 4300i has the MIPS III
>ISA. See http://www.mips.com for details.
Wrong. The MIPS R3000 is the 2nd generation MIPS I processor.
The R6000 is the only MIPS II processor.
>> Note that both PSX and N64 are MIPS based consoles,
>> with PSX using a MIPS III processor(R3000A) and
>> Nintendo using a MIPS IV processor(R4300i). The next
>> MIPS architecture is MIPS V, on which PSX2 and N128
>> will be built up on.
>>
R3000 is a MIPS I processor.
R4300 is a MIPS III processor.
Nobody's Perfect <nob...@nowhere.com> wrote in article
<334226...@nowhere.com>...
> MIPS V accelerates geometry processing performence
> two fold by allowing floating point unit to handle
> two 32bit floating point per cycle. That is too bad
> since Hitachi's SH-4 handles four, not two, 32bit
> floating point prt cycle.
>
> So what does this mean? This means MIPS V processors
> will not be able to match SH-4's 32bit floating point
> processing performence, crucial in geometry processing
> and light sourcing calcultations. Simply, MIPS V SUX,
> AND SO DO MACHINES BASED ON IT, NAMELY PSX2 AND
> NINTENDO128.
Yes, but you're forgetting several important things. Firstly, in the next
generation (presuming 32 & 64bits to be the current generation, as they now
are), geometry processing and light sourcing will be added to the list
already consisting of texture mapping, gouroud shading and in most cases
(Jaguar, 3D0, PSX, N64 but not Saturn) translucency, that is the list of
things which are done regardless of the CPU. Therefore, the 3d world is
calculated and drawn entirely separate of the CPU and all the CPU has to do
is update game elements - that is where all the objects are in the 3d world
and what they are doing. Some say the N64 already works this way.
Therefore, the speed of the CPU will have no bearing on the speed of the
geometry and so on because it will be calculated by custom chips. If anyone
leaves the CPU to take the strain, it will be Sega who are notoriously bad
at producing a good custom chip set up (in their case having so many
processors is bad, no matter what they tell you).
Note though, what I am saying above is my interpretation of the rumours
and so on that have come so far. Until a proper announcement is made,
conversations such as this can only be based purely on speculation, and so
if my sources disagree with your sources (none of which are actually any of
the people who would really know at this point), then that would make us
disagree. In other words, don't (as so many do) take it personally that I'm
disagreeing.
-Thomas
Oops sorry, yes you're right. Anyways, people curious about the
new MIPS V extension should go to
http://www.sgi.com/MIPS/arch/ISA5/
-------------------
leroy @ mti.sgi.com
Where did you get that info, Nobody? www.sega.com? Whatever.
Regardless of it's truth or your perceptions thereof, small details like
that are toatally irrelevent. Nobody cares if the BB(Sat2) is a bit
more powerful, they care about the Game quality. Personally, I like
Nintendo's games the best, but that doesn't mean "N128" will crush the
others, eiher. BTW, it's the N64II, as yet, NOT N128. A 128-bit
processor would be STUPID anytime soon. The next Nintendo system is
going to be 64 Bit.
--
Cryptic Bug
Cr...@vcn.bc.ca
Who cares? Buy whichever next-next-generation system has the best
games when it comes out. Your obsession with the future tells me
that you aren't too happy with the present-- may I suggest you pick
up another system, now, while you're young enough to enjoy it?
--
Ken Small
kens...@mcs.com
================
Magic 8-BALL sez: Signs point to yes
> Oops sorry, yes you're right. Anyways, people curious about the
> new MIPS V extension should go to
>
> http://www.sgi.com/MIPS/arch/ISA5/
This doesn't change the fact that MIPS V will be significantly
inferior to SH-4 in 4 by 1 floating point MAC operation. SH-4
is so fast in this operation(7 cycles to transform one polygon. Wow!
And SH-4 is clocked at 200 MHz.) that Hitachi is worried about
finding a memory solution that can keep up with SH-4. Hitachi
said that they had looked into Mitsubishi's processor+DRAM
combo and may do the same with SH-4, combining SH-4 core and
2MB of DRAM into one die to solve memory bandwidth problem. Sega
is going to kick Sony and Nintendo's ass out of videogame market
thanks to extreme performence of SH-4+Dreal3D combo.
>
> -------------------
> leroy @ mti.sgi.com
<Sigh> .. You do love to dwell on the future, eh? Must be because the
present is so depressing for you.. hehe.
Tell ya what.. you send me a free Saturn, and I'll become a loyal Sega
customer just like you want. Deal?
While you are predicting the future (official announcements of the specs
on the N128 and PSX2 have not came out yet) could you tell me who will
win the World Series this fall? What about the U.S. presidential
elections in 2000? Didn't think so. Until you can begin reasoning like
a rational person, please remain silent while those of us with a grasp
on reality comment on what is going on.
Go to arcade and see who has the fastest rendering hardware; Lockheed
Martin.
1998.
> BTW, the
> systems that come out in 2020 will make your new Sega system totally and
> completely obsolete.
God, will you be playing videogames 33 years from now?
: > two 32bit floating point per cycle. That is too bad
: > since Hitachi's SH-4 handles four, not two, 32bit
: > floating point prt cycle.
: >
: > So what does this mean?
It means that the SH-4 will run matlab real good. Tob bad that's not
very important for games.
Oh well. Have fun playing double precision math programs.
Pasquale DeMaio
> Denis Leroy wrote:
>
> > Oops sorry, yes you're right. Anyways, people curious about the
> > new MIPS V extension should go to
> >
> > http://www.sgi.com/MIPS/arch/ISA5/
>
> This doesn't change the fact that MIPS V will be significantly
> inferior to SH-4 in 4 by 1 floating point MAC operation. SH-4
> is so fast in this operation(7 cycles to transform one polygon. Wow!
> And SH-4 is clocked at 200 MHz.) that Hitachi is worried about
> finding a memory solution that can keep up with SH-4. Hitachi
> said that they had looked into Mitsubishi's processor+DRAM
> combo and may do the same with SH-4, combining SH-4 core and
> 2MB of DRAM into one die to solve memory bandwidth problem. Sega
> is going to kick Sony and Nintendo's ass out of videogame market
> thanks to extreme performence of SH-4+Dreal3D combo.
>
>
ROFL. Can you talk about something that is fact and not fiction and/or
twisted facts? Because if you want hot air here's some. The N128 will use
a dual DEC Alpha 500's sport 32 MB RAM, a 12X CD drive and a 1MB flash
RAM card for saving games, etc. Did I mention it will have an included
21" monitor?
What you read above was facts disclosed by Hitachi at
microprocessor engineering forums. Check out microprocessor
related jounals like IEEE Micro and Microprocessor Report
to see for yourself.
> William Kenneth Thomas wrote:
>
> > Oh no mommy, the big mean Sega machine is going to make my systems obsolete.
> > Well WHERE IS IT THEN?
>
> 1998.
>
Any concrete evidence or is it just more of your bullshit? You never list
any sources because all of the stuff you post is pure bullshit.
Go to the arcade and see who has the most expensive hardware; Lockheed
Martin.
The price might explain something about why the Lockheed Martin
developed M3 is so powerful don't you think?
> a dual DEC Alpha 500's sport 32 MB RAM, a 12X CD drive and a 1MB flash
> RAM card for saving games, etc. Did I mention it will have an included
> 21" monitor?
Hey while we are on the subject, I happen to know that the Nintendo 256 is
going to use some sort of alien technology that the American and Russian
scientists are trying to decipher at the moment. :)
I'm afraid that there is one very important thing that you have seemingly
forgotten about.... Sega as we now know it will no longer exist in 2 1/2
years from now. They will likely be delinated into two different and
distinct companies with two very different functions, none of which will
have anything to do with console hardware development.
--------------------------------------------
"Change is the history of man, sir." - Commander Data