367 lines
16 KiB
Plaintext
367 lines
16 KiB
Plaintext
|
|
Ultra 64 Hardware Specifications.
|
|
[[*]SGI Index Page] [[*]Ultra 64 Index]
|
|
---------------------------------------------------------------------------
|
|
|
|
Nintendo Ultra 64 Hardware Specs List
|
|
|
|
Obviously, the specs of the NU64 are what many people are interested
|
|
in, but Nintendo have not yet released very many officially confirmed
|
|
specs. The PR Manager for the Nintendo Ultra 64 at Nintendo told
|
|
me that they would be making some announcements during the E3 Conference,
|
|
but in the event very little was said because of the delay in the
|
|
release date. The PR Manager has told me since that the NU64 will
|
|
be unveiled, "...at the Shoshinki Show in Japan over the Thanksgiving
|
|
weekend (November 23).".
|
|
|
|
It's pretty easy to find rumour, hearsay and speculation as to what
|
|
the NU64 will be able to do, but at the end of the day the magazines
|
|
can quote very few finalised and confirmed facts. My main complaint
|
|
is that many magazines print far too much complete rubbish; sometimes
|
|
it's under-hype and sometimes it's over-hype.
|
|
|
|
As a result, this page has confirmed hardware specifications only
|
|
(I used to have an 'unconfirmed specs' section, but I decided to
|
|
remove this).
|
|
|
|
In other words, what I will have here are specs which are all direct
|
|
from Nintendo and/or SGI sources.
|
|
|
|
What I have added is a section designed to dispell some myths which
|
|
have been printed in various magazines, on the grounds that over-
|
|
hype is probably worse than under-hype in the long-term.
|
|
---------------------------------------------------------------------------
|
|
|
|
Confirmed Hardware Specifications
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
Approximate System Power: Reality Engine quality.
|
|
(source: George Zachary of SGI. See
|
|
his quote in the [*]New Scientist
|
|
article from the [*]articles page).
|
|
|
|
Main Controller Processor: 64 bit MIPS RISC CPU (probably the
|
|
[*]R4300i but this isn't known for sure.
|
|
Iam sure it isn't the R4200).
|
|
100 MHz (minimum)
|
|
100 MIPS
|
|
|
|
My main reason for believing the controller to be the R4300i
|
|
is the following extract from the R4300i home page:
|
|
|
|
"The MIPS R4300i microprocessor extends the benefits of
|
|
RISC's performance to consumer electronics. The R4300i
|
|
microprocessor also delivers high performance to existing
|
|
embedded and computing applications at a low cost. The
|
|
low cost and high performance provided by the R4300i are
|
|
needed for the latest consumer applications such as
|
|
interactive television and games."
|
|
|
|
Graphics Processor: 100 MFLOPS? (source: Nintendo magazine)
|
|
|
|
Iam waiting to see if www.nintendo.com
|
|
will provide a complete spec list. Nothing
|
|
so far though, which is daft. If they're
|
|
not going to say anything until November,
|
|
then their specs page should be removed.
|
|
|
|
Graphics Hardware Features: Load Management,
|
|
Tri Linear MIPmapped Interpolation,
|
|
Anti-Aliasing.
|
|
|
|
([*]here is an explanation of these terms)
|
|
|
|
System BUS: Speed: 500MHz
|
|
Bandwidth: 500MB/second transfer rate.
|
|
|
|
Colour: 24 bit, 16.8 million
|
|
|
|
Audio: CD quality
|
|
|
|
Supported Resolutions: PAL (European TV Standard: 768x576),
|
|
NTSC (USA TV Standard: 645x486),
|
|
SVGA (1024x768).
|
|
|
|
(HDTV resolutions are suspected, but
|
|
are as yet unconfirmed and unlikely)
|
|
|
|
Silicon Cartridge ROM size: From: 32 megabits (4MB)
|
|
To: 800 megabits (100MB)
|
|
|
|
Silicon Cartridge: Access Time: 2 million times faster than a CDROM drive.
|
|
Bandwidth: 30MB/sec transfer rate (I previously said
|
|
500MB per sec but I now think this is
|
|
wrong. At least, I'll leave it at 30MB/sec
|
|
until I can find again whatever piece of
|
|
info it was that made me think it was
|
|
500MB/sec).
|
|
|
|
System Cost: US Retail Price of below $250.
|
|
|
|
Cartridge Cost: $50 upwards.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
'Edge' magazine said the controller processor was clocked at 105.58MHz.
|
|
From the above data, it seems very likely that the processor is the
|
|
[*]R4300i, made by [*]MIPS Technologies, a subsidiary of [*]Silicon
|
|
Graphics Inc [[*]Euro-mirror]. However, no one knows for sure yet
|
|
whether this is the case. See below for my reasons as to why I think
|
|
the main controller is not the R4200.
|
|
---------------------------------------------------------------------------
|
|
More info as I get it.
|
|
---------------------------------------------------------------------------
|
|
|
|
Myths, Rumours and Chinese Whispers...
|
|
|
|
Yep, the magazines have been having a field day! Never have I seen
|
|
so much unconfirmed guesswork and rubbish in all my life. Here I
|
|
will dispell some of the untrue things which have been said about
|
|
the Ultra 64.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen, Audio Visuality and Edge said the U64's textured-polygon
|
|
rate is 100000/sec! That ain't much.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
Well, after much badgering and hassling of people at SGI, I still
|
|
can't get any reliable statement on this figure. George Zachary's
|
|
New Scientist quote implied a figure of 600000/sec, but I have been
|
|
told by several people involved with the NU64 that it is not as high
|
|
as this (though it is higher than 100000/sec), which is a shame to
|
|
say the least as this means that (I have to say) George's quote in
|
|
the New Scientist article is misleading. Overall, George would not
|
|
give any precise details because he is not allowed to, but here is
|
|
what he did say:
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
"There is a HUGE difference in the realism between "over 100,000
|
|
real-time, anti-aliased, tri-linear mip-mapped interpolated,
|
|
gouraud-shaded, perspective-correct, z-buffered, lit, blended
|
|
[and the rest of the poly features] polygons" and Sony's 360K
|
|
quote. They compare absolutely different numbers."
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
and:
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
"If you look at NU64 press release back in 1993, we stated over
|
|
100,000 full featured polygons per second. Not 100,000 exactly.
|
|
But at least 100,000."
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
George said that full details would be released when the NU64 is
|
|
unveiled in Japan on November 23rd. It's too long a time to wait,
|
|
IMHO. However, since the 3DOMk2 (the NU64's main rival) will not
|
|
be out until 1996 as well, at least the public will have an opportunity
|
|
to judge properly beforehand. I still say that there should be a
|
|
full disclosure now though because the M2 is looking very good at
|
|
the moment. How can I rant and rave on rec.games.video.advocacy without
|
|
full specs? :D More info as I get it.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
Edge (and others) said the controller is an R4200.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
The U64's main controller is very unlikely to be the R4200; the target
|
|
market of the R4200 is portable notebooks and low-cost desktop systems
|
|
- a very different target market to the R4300i (see above). Further,
|
|
if there was a 105MHz R4200 in existence, I would have thought that
|
|
the [*]relevant page at MIPS Technologies would have mentioned it
|
|
by now. :)
|
|
|
|
I must emphasise again, however, the no official info has yet been
|
|
released as to what the NU64's main controller actually is, so don't
|
|
go believing the magazines' info as if it's fact, because it isn't.
|
|
I currently believe that it's the R4300i because that is the most
|
|
likely candidate based upon the info given at the [*]R4300i Home
|
|
Page as well as the fact that the oft-quoted 100MHz clock speed matches
|
|
that of the R4300i.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen and GameFan said it will have 64bit sound!
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
This is complete rubbish. There is no such thing as 64bit sound.
|
|
The best that exists is 20bit which is used in recording studios.
|
|
The human ear cannot detect better than this. Some people have emailed
|
|
me saying that mags actually said it would have a 64bit sound chip.
|
|
This would be fine, but I am going by what I see in terms of rumour
|
|
and hype on the newsgroups and several people definitely quoted mags
|
|
as saying that the NU64 has 64bit sound which is a very silly statement
|
|
for a mag to make. I've been told that EGM says this in it's June
|
|
issue.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen said the U64 will have 64 sound channels.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
This depends on what they mean by sound channels. If they mean physical
|
|
output channels, then no way can 64 be correct as 7 are needed at
|
|
most for good surround-sound, 2 for standard stereo. If they mean
|
|
internal mixable channels (voices), then maybe, but I'll be surprised
|
|
if it's as high as 64. Mind you, the PlayStation has 24 channels
|
|
and the Saturn has 32, so maybe 64 channels for the U64 is correct.
|
|
We'll see.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen said it can do real-time ray tracing!
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
This statement is wrong mainly because it's meaningless. Ray-tracing
|
|
is a technique which relies heavily on raw floating point computing
|
|
power. It is not something which the U64 is designed to do (the U64,
|
|
just like the other new consoles, excels at [*]hardware texture mapping).
|
|
'real-time ray tracing' is also meaningless because any machine can
|
|
do 'real-time' ray-tracing (eg. my old 8 bit, 4MHz, 64K RAM, BBC
|
|
Micro model B can do real-time ray-tracing... in a 4 by 4 pixel window
|
|
using only two colours!). See what I mean? The compuational work
|
|
needed for ray-tracing varies greatly with the desired level of detail
|
|
and complexity (eg. number of depth levels). To give you some idea:
|
|
my colleague at work renders many ray traced images every week for
|
|
his PhD related sonar work. On the 486DX33 PC in the lab here, a
|
|
640x480 image at a moderate level of detail takes several hours to
|
|
render. On the Sun Sparc 4 which he uses most of the time, it takes
|
|
some 35 minutes. On the 134MHz 64MB RAM SGI Indy I use it takes about
|
|
3 to 5 minutes depending on whether I'm using the machine at the
|
|
same time or not. In other words, there's just no way that the NU64,
|
|
or any of the consoles, will be able to do high quality (eg. Babylon
|
|
5) 'real-time ray-tracing'. You need something like an [*]SGI Power
|
|
Challenge [[*]Euro-mirror] to do that.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen said the Reality Immersion Co-processor goes at 80MHz.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
This is completely unconfirmed. As I write this, SGI have released
|
|
no precise specs on the R.I. chip at all. The same goes for the DSP.
|
|
We will just have to wait and see.
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen said the NU64 would be able to output at 1200x1200!
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
This is a very peculiar figure and one which I find very unlikely
|
|
(if only because 1200x1200 is a square 1:1 ratio and will not fit
|
|
correctly onto any future/existing TV/monitor). I regret to say that
|
|
EGM has now apparently used this figure as well (probably nicked
|
|
it from NextGen. *sigh*).
|
|
|
|
It is known that the NU64 can output in 1024x768 SVGA, and that it
|
|
supports PAL (770x576 European TV standard) and NTSC (645x486 USA
|
|
TV standard), but nothing is currently known about support for HDTV-
|
|
quality resolutions. There are three HDTV standards: 1920x1035, 1920x1152
|
|
and 1600x1200. It is not currently known whether the NU64 will be
|
|
able to output any of these resolutions (all the [*]SGI article actually
|
|
says is that the console will ensure 'future compatibility with HDTV'.
|
|
ie. it doesn't explicity say that the machine can output in either
|
|
HDTV standard; ie. we don't know whether the HDTV output will just
|
|
be the 1024x768 output scaled to fit, or genuine HDTV output). The
|
|
high-res TV standard of 1600x1200 is very new, so I would be surprised
|
|
if the NU64 supports that since there are so few TV's and monitors
|
|
in existence which can display it. Basically, I'm trying to point
|
|
out that, as far as high-res modes go, we just don't know yet for
|
|
sure. SVGA yes. HDTV? I doubt it, but wait and see.
|
|
|
|
Incidentally, I've seen a lot of posts on the rgv* groups (mostly
|
|
from people in the USA) saying that HDTV resolutions are pointless
|
|
because most TVs can't display more than 400 lines. To this I would
|
|
respond with these points:
|
|
|
|
- this 400 line figure refers to the US NTSC TV Standard of 645x486.
|
|
This format is, as far as I know, the smallest anywhere. There are
|
|
several TV standards around the world (NTSC, PAL, HDMAC, HDTV, SECAM).
|
|
|
|
- in Japan, HDTV is very common indeed and laserdisc is practically
|
|
standard there (as opposed to VHS elsewhere). The NU64 is a Japanese
|
|
product first and foremost.
|
|
|
|
- HDTV is now becoming quite common in parts of Europe (the EEC is
|
|
currently running its '16:9 Programme') and HDTV sets have come down
|
|
in price quite sharply in the last 2 years in the UK alone (from
|
|
$3000 to $1200. Still too much, but until there are digital HDTV
|
|
channels and/or laserdisc become even cheaper, the prices will remain
|
|
high); laserdiscs have come down in price quite a lot too.
|
|
|
|
- one way of gently forcing nations into better TV standards is by
|
|
creating a demand for them. If the public start owning game consoles
|
|
that can output in SVGA and upwards, the public are going to start
|
|
wanting HDTV. Good thing too, IMHO. :)
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
NextGen said the U64 has a 64bit RISC data path!
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
A data path is just a data path and has nothing to do with RISC,
|
|
CISC or CRISC at all. Too many mags spout figures and words without
|
|
even knowing what they mean (Audio Visuality provides a good example;
|
|
in their January 1995 issue they state that MFLOPS stands for 'millions
|
|
of floating points per second'. Argh!).
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
Iheard that the U64 will come ready-boxed with a snazzy UNIX Operating
|
|
System and a 200MB hard drive!
|
|
|
|
|
|
-=- [PREFORMATTED] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|
|
The part about the operating system is is complete and utter rubbish.
|
|
The NU64 doesn't even have an operating system (see the [*]New Scientist
|
|
article for more info). As for the harddrive, I would be very surprised
|
|
if this were true. I'd have thought it more likely that flash memory
|
|
cards would be used for position-saving, etc., although New Scientist
|
|
reports that Nintendo is developing a 1.2GB writable CD system for
|
|
the NU64, to be made available in 1997 (ie. not available at launch).
|
|
|
|
---------------------------------------------------------------------------
|
|
More rumours, mis-information and total tosh dispelled as I see it!
|
|
:)
|
|
---------------------------------------------------------------------------
|
|
[[*]SGI Index Page] [[*]Ultra 64 Index]
|
|
|
|
|
|
|
|
---------------------------------------------------------------------------
|
|
HTMLCon Conversion Statistics and Information
|
|
---------------------------------------------------------------------------
|
|
Total bytes processed .......... 15726
|
|
Total lines processed .......... 351
|
|
Total links processed .......... 19
|
|
Total images processed ......... 0
|
|
Total strings replaced ......... 0
|
|
Total filters used ............. 0
|
|
Were references preserved ...... FALSE
|
|
Rough formatting preserved ..... FALSE
|
|
Space compressed used .......... TRUE
|
|
Rough line break used .......... 72
|
|
|