19-Apr-2024 05:23 GMT.		 UNDER CONSTRUCTION
Anonymous, there are 33 items in your selection
[News] Transmeta don't plan 68k supportANN.lu
Posted on 14-Feb-2000 20:02 GMT by Christian Kemp33 comments
View flat
View list
Black Hand says he repeatedly mailed Transmeta to enquire about 68k emulation support, or the posibility to access and modifiy the morphing code. Here's what Frank Priscaro replied: Supporting the Amiga was something that we considered early on, but had to drop because all of our resources needed to be focused on the Linux and x86 markets. We have no plans at this time to support the 68K family of processors.
Transmeta don't plan 68k support : Comment 1 of 33ANN.lu
Posted by Anonymous on 13-Feb-2000 23:00 GMT
Good. Now lets drop this and move on ...
Transmeta don't plan 68k support : Comment 2 of 33ANN.lu
Posted by Fabrice Jogand-Coulomb on 13-Feb-2000 23:00 GMT
I think we need to work with Transmeta but Amiga Inc has to implement the emulation.
Amiga has always been a machine with a lot of emulators and transmeta chip and code morphing could be a good way to improve Amiga OS. We can imagine to have an OS that support any type of application including x86 and run it transparently in the same way Mac OS is doing with 68k software.
Transmeta don't plan 68k support : Comment 3 of 33ANN.lu
Posted by John Waller on 13-Feb-2000 23:00 GMT
In reply to Comment 2 (Fabrice Jogand-Coulomb):
Why bother? If Amiga Corp succeeds with their plans, all this discussion of CPU's will be irrelevant. Besides, except for low power uses, the Transmeta chip offers no inherent advantage over other CPU's where the new Amiga is concerned.
Transmeta don't plan 68k support : Comment 4 of 33ANN.lu
Posted by Gringo^mF on 13-Feb-2000 23:00 GMT
I emailed Transmedia the day the site had news and they said
"no 68K support planned at this time".
For what it's worth.
Gringo^mF
Transmeta don't plan 68k support : Comment 5 of 33ANN.lu
Posted by Coz on 13-Feb-2000 23:00 GMT
When was the last 68K Amiga manufactured. I'm not surprised in the least about no support for 68K. If TM is a big success then maybe somebody will make a new layer supporting 68K for a niche market. Yeah and right about the same time Z-80 and 6502. 8^)
Transmeta don't plan 68k support : Comment 6 of 33ANN.lu
Posted by Marek Pampuch on 13-Feb-2000 23:00 GMT
In reply to Comment 5 (Coz):
I think that TM will not be a great success. The technical data of TM 3120 & 5400 are much more worse then was forecasted. Only ONE advantage in the moment is "cool" (max. temperature of procesor is 47 Celsius compared to 105 in Pentium).
There is NO software yet, even mythic Mobile Linux, and Linus has considerable problems with Mobile Linux coding (i`ve heard). And lack of software killed
some better things ;(
so - nothing to bother
Transmeta don't plan 68k support : Comment 7 of 33ANN.lu
Posted by sutro on 14-Feb-2000 23:00 GMT
Who bothers anyway ? Transmeta will NOT be a factor in Computing.They were not interested in any other CPU except intel in the first place , nor they were able to work with another CPU. A lame x86 emulator and a company ran by intel maniacs , that is what Transmeta stands for.
Baah , bring G4 to miggy now !!!
(I hope H&P's AmigaOS-PPC progresses well)
Transmeta don't plan 68k support : Comment 8 of 33ANN.lu
Posted by XDelusion on 14-Feb-2000 23:00 GMT
I thought Amiga's best possible future would stem from Transmeta, but then I realized that Amiga is in bed with the company who offers us this!!!!
http://www.tao-group.com/2/tao/index.html
Besides maybe Transmeta will change there mind...
Time will tell, though I really want Amiga OS on my PeeSea RIGHT NOW!!!
Sorry, but hey guys, were kinda in shit creek, pardon the fench. Something really imaginary, needed, and inovative needs to come from Amiga/Transmeta if there is ever going to be a hope in hell.
If all else fails, I will stick with A dual (or more) processor based IBM, or standard PeeSea mother board, running Linux, with VMWare assigning independent CPU's for the OS's I will be running in the X enviroment.
Like Amiga OS, it has a GUI, it has CLI, it has Multi-Screens, It has a form of MUI, ect. ect. For me, it is the closest thing to Amiga I have got. And if this amiga business fails, I hope someone makes an Amiga OS based X enviroment for me.
Transmeta don't plan 68k support : Comment 9 of 33ANN.lu
Posted by JW Olson on 14-Feb-2000 23:00 GMT
I too asked about a 68k Amiga emulation and recieved a very nice e-mail back stating I was one of the few actual people who asked this question. They went on to state that most were from obvious fake names and so were ignored to begin with.
Still a software production company can write software for and use any harware they so desire and it really is not up to a hardware company to fix the use.
Transmeta don't plan 68k support : Comment 10 of 33ANN.lu
Posted by Mario Saitti... on 14-Feb-2000 23:00 GMT
In reply to Comment 7 (sutro):
> Who bothers anyway ? Transmeta will NOT be a factor in Computing.They were
Why do you belittle technology you do not understand?
> not interested in any other CPU except intel in the first place , nor they
> were able to work with another CPU. A lame x86 emulator and a company ran by > intel maniacs , that is what Transmeta stands for.
That lame x86 emulator is the world's first self abstracting VLIW processor, and you think that is lame? Why haven't you been able to do it?
> Baah , bring G4 to miggy now !!!
Perhaps you need to look at IBM's work on the DAISY and BOA systems? Take a look at the G4's life span...
Mario.
Transmeta don't plan 68k support : Comment 11 of 33ANN.lu
Posted by Hasse on 14-Feb-2000 23:00 GMT
I think it's very surprising that they even considered Amiga (or M68k) in the first place. I really like both of these, m68k coding is the best, but let's face it. The m68k series and the Amiga in particular are very outdated pieces of technology. For a new major league company like Transmeta to support any of these would be nothing less than madness. And how much good would hardware 68k emulation really do for the Amiga community anyway? Yes, there are quite a lot of good apps for the "classic" Amiga, but what we need is some real progress. With the speed Amiga has changed hardware vise the last decade, we really need a leap in technology now. We're not going to get that if we keep clinging on to obsolete technology like the m68k series of processors.
Transmeta don't plan 68k support : Comment 12 of 33ANN.lu
Posted by Paul Laycock on 14-Feb-2000 23:00 GMT
In reply to Comment 11 (Hasse):
...which can only happen when (if :( ) AmigaOS goes PowerPC. This in itself could take perhaps a year. The Amiga/Tao partnership could work, though, allowing the user to choose as many CPUs as they want, and of whatever type (be it PPC, x86, Transmeta...)
Transmeta don't plan 68k support : Comment 13 of 33ANN.lu
Posted by Neko on 14-Feb-2000 23:00 GMT
Well *DUH*
Neko
Transmeta don't plan 68k support : Comment 14 of 33ANN.lu
Posted by Thomas Palestig on 14-Feb-2000 23:00 GMT
I din't expected that they should make a M68K emulation, but it's strange that
they did emulation for the x86 series, do they realy need that, what i know
is all x86 series compatilble with older x86 and i think it's enough with Intel
and AMD that makes processors that kills other CPU market.
And what I read is the speed of the Tramsmeta that same as "a standard pc" of today..
It's nice to see a company look back with nostalgi but why should a PC user
buy a processor that doesn't run at full speed (even if they say so) as it's
not native code, and the problem in the computer industries isn't fast processors,
it's a OS with good hardware that works in a good way without collide and crash
with each others, and the new version of the Standard OS today does only have
63.000 known buggs that's shouldn't be fixed (what i have read, but havn't asked
billy) so i think the whole IT idustry is a bit crazy right now..
And the Trasmeta cpu does nearly the same as TAO's Elate, and both work in the
software way but TAO's doeas it on all types of CPU (/FPU?) after some coding..
Atleast the old Motorola PPC CPU's din't need extra cooling, like the Transmeta.
Transmeta don't plan 68k support : Comment 15 of 33ANN.lu
Posted by John Block on 14-Feb-2000 23:00 GMT
My understanding: Amiga to run on PPC and PC hardware before migrating to a new chip to be decided. Based on Elate/Intent, Amiga supposedly can be tweaked to run wherever Elate runs. Transmeta emulates an intel 86x so Amiga can run on Transmeta as if it were running on a 86x. Next question, will Transmeta emulate PPC? Don't think they said anything, so probably not.
Transmeta don't plan 68k support : Comment 16 of 33ANN.lu
Posted by Mario Saitti on 14-Feb-2000 23:00 GMT
In reply to Comment 14 (Thomas Palestig):
> I din't expected that they should make a M68K emulation, but it's strange
> that they did emulation for the x86 series, do they realy need that, what i
Well not really. The x86 line is without a doubt the most succesful microprocessor ever. The problem with the x86 is that it's current technology is reaching it's limits. Both AMD and Intel are going to have serious problems keeping the line competitive in the near future. The same goes for the PPC.
Transmeta chose the x86 simply because the market is the largest and best suited to a take over by a new paradigm. Intel and AMD may not be able to continue the x86 line, but Transmeta will, and should be able to outpace any AMD or Intel proessor once their compiler is refined.
> know is all x86 series compatilble with older x86 and i think it's enough
> with Intel and AMD that makes processors that kills other CPU market.
VLIW (the model used in Crusoe) works by using software to replace large and complex amounts of hardware with even complex software. The flaw to this is that since the software is so directly tied to the processor, there is no compatibility whatsoever between generations, meaning all programs must be recompiled. Advantages are that simplified hardware means that frequency can be increased in huge leaps compared to either RISC or CISC.
Transmeta circumvents the compatibility issue in Crusoe by having the processor abstract itself using emulation. This meant either a new Instruction Set Architecture (ISA) was required, or the use of a 3rd party architecture. As you know they settled on X86.
So what you should understand is that the X86 ISA is primarily for one purpose, the generation compatibility between future Transmeta processors. In using VLIW, Transmeta have effectively eliminated most issues with the X86 line.
> And what I read is the speed of the Tramsmeta that same as "a standard pc" of > today..
That will not last. Neither AMD or Intel will be able to match a VLIW processor's clock speed in the long run now that advanced compilers like IBM's DAISY exist, unless they too adopt the technology.
> It's nice to see a company look back with nostalgi but why should a PC user
There is no nostalgia here. Transmeta have set the future...

> And the Trasmeta cpu does nearly the same as TAO's Elate, and both work in
> the software way but TAO's doeas it on all types of CPU (/FPU?) after some
> coding.. Atleast the old Motorola PPC CPU's din't need extra cooling, like
> the Transmeta.
The PPC line is as dead as the X86. IBM has already created the Crusoe equivalent for the PPC. It is only a matter of time before the technology begins to mature and is truly useful.
Finally, elate still needs a fast CPU. In 5 years the VLIW processors will be hitting at least 3 GHz. Can the X86 line also boast this?
Mario
Transmeta don't plan 68k support : Comment 17 of 33ANN.lu
Posted by Plain English on 14-Feb-2000 23:00 GMT
How much will this cost?
These much powerful chips than can beat the pants off todays stuff? Great. What about these points...
AMD and Intel have already reached the limits of the x86 architecture but still come out with faster and cheaper chips. There has been a steady move away from the original chip since the 386 through each generation (ie P, PII, PIII, Athlon, Xeon, etc.), just altering a bit of code here, 64bitting that... The point is that I could see the x86 being around for a good time yet due to mass-production making it cost less and getting more reliable, it may be 50% or 60% slower than a better alternative but it's also about a quarter of the price and just about ALL software works on it.
Will they even try to compete in the Desktop market? There has been a popular move since the Internet to centralised distribution systems, ie GUI based terminals. I know that having a personal computer offers more in the areas of graphics, sound, local processing but for how long? This could be the advent of things like Internet based hand held organizers, the Internet toaster (had to be mentioned).
Just a thought.
Transmeta don't plan 68k support : Comment 18 of 33ANN.lu
Posted by John Waller on 14-Feb-2000 23:00 GMT
In reply to Comment 8 (XDelusion):
>If all else fails, I will stick with A dual (or more) processor based IBM,
>or standard PeeSea mother board, running Linux, with VMWare assigning
>independent CPU's for the OS's I will be running in the X enviroment.
Forget the dual processors, that is old hat. The Amiga/Elate system is multi-processor,
as in Transputer. The more CPU's you tie in the more powerful the Amiga becomes. In
theory, you could have the equivalent of a 10 Ghz. or even 100 Ghz., or more CPU running
in your Amiga. Assuming Amiga corp succeeds with their plans :)
Transmeta don't plan 68k support : Comment 19 of 33ANN.lu
Posted by sutro on 14-Feb-2000 23:00 GMT
In reply to Comment 10 (Mario Saitti...):
Transmeta aims at embedded systems ("Mobile Internet Computing" , strangely it reminds me of "Multimedia PC").
In this section intel's x86 technology , logic and experience are no factors (compared to Motorola).
How will Transmeta compete to intel's and motorola's GHzs , copper technology and 0.1micros process ?
By designing more advanced s/w ?
As for IBM , if you saw the "Transmeta announcement" fiasco on 19th of Jan you should remember they had
a hard time persuading reporters that they had a contract with IBM or even that IBM knew anything about them.
I believe IBM had been already working at these things way before them.
I don't discard the idea (VLIW) but the company (Transmeta).
As for PPC , they are here to stay for many years (un)fortunately to some ....
I bet you 've been a Cyrix pc user...
Transmeta don't plan 68k support : Comment 20 of 33ANN.lu
Posted by thedoctor on 14-Feb-2000 23:00 GMT
In reply to Comment 17 (Plain English):
You are way out in left field in even starting to compare prices, on TM versus the X86 clone chips, the top of the line 700mhz athlon and P3 are OEM at
$500 plus, and the announced price of the 700mhx TM is up to $150...and 700mhz on 128 bit instruction is still 2 times speed of 70mhz on 64 bit instruction...
Thats all simple math...
Nay sayers should get better facts.....
And transmetta never said that what the processor was doing was "emulating", thats is being used as the term when a real terms for this new process, has yet
to be established.....What the CEO said in the presentation was something
like "a new process of translating on the fly, for lack of a better term", and the moderators all injected Emulation to discribe..
Transmeta don't plan 68k support : Comment 21 of 33ANN.lu
Posted by Mario Saitti on 14-Feb-2000 23:00 GMT
In reply to Comment 17 (Plain English):
> How much will this cost?
No idea;)
> These much powerful chips than can beat the pants off todays stuff? Great.
> What about these points...
> AMD and Intel have already reached the limits of the x86 architecture but
> still come out with faster and cheaper chips. There has been a steady move
Sooner or later it will not longer be commercially viable to upgrade the line. Intel and AMD suggest this is rapidly approaching. However there is no chance of these chips vanishing any time soon. Look how the 68k is still being produced in various forms.
> away from the original chip since the 386 through each generation (ie P, PII, > PIII, Athlon, Xeon, etc.), just altering a bit of code here, 64bitting
> that... The point is that I could see the x86 being around for a good time
> yet due to mass-production making it cost less and getting more reliable, it > may be 50% or 60% slower than a better alternative but it's also about a
> quarter of the price and just about ALL software works on it.
Is it going to be a quarter of the price? One thing you have to understand is that VLIW is a solution which eliminates one hell of a lot of hardware from the the equation. The costly part of VLIW production is actually the design of the software element. Once that is done things get cheaper way faster than for a RISC or CISC chip. For starters because there is a lot less chip to actually produce. Think about which model has the potential to be a quarter of the price under these conditions.
Intel has the Itanium/Merced processor for the VLIW front underway. they know this is one of the only avenues for continued domination of the Miroprocessor industry.
> Will they even try to compete in the Desktop market? There has been a popular > move since the Internet to centralised distribution systems, ie GUI based
> terminals. I know that having a personal computer offers more in the areas of > graphics, sound, local processing but for how long? This could be the advent > of things like Internet based hand held organizers, the Internet toaster (had > to be mentioned).
Perhaps and perhaps not. I guess it is way too early to tell where things are heading.
Mario.
Transmeta don't plan 68k support : Comment 22 of 33ANN.lu
Posted by Marek Pampuch on 14-Feb-2000 23:00 GMT
In reply to Comment 10 (Mario Saitti...):
Why do you "bebig" :)) The Transmeta?
I`ve read almost all about TM (on Crusoe page, Slashdot, and some good newsserwers.
Maybe VLIW will be the great thing, but NOW there is any VLIW
(even theory is not reworked :))
Maybe Mobile Linux will be the great thing, but NOW there is any ML
Maybe TM 5400 will the great thing, but NOW there is ANY TM 5400
If anybody concerned with Amiga would have Soros money and 4 years of calm
- all of us shoud have a much more Amiga technology now.
I am VERY dissapointed with Transmeta, and happy that it is not "Amiga inside"
and if you still think that I`m not understanding the Crusoe technology (promises, promises, promises, a lot of words and a lot of water) we can continue the dialog on private
Transmeta don't plan 68k support : Comment 23 of 33ANN.lu
Posted by Plain English on 15-Feb-2000 23:00 GMT
In reply to Comment 21 (Mario Saitti):
Do you think that was the problem with Apple (amongst other problems), that they charged too high for what should've been a lot less?
Transmeta don't plan 68k support : Comment 24 of 33ANN.lu
Posted by Plain English on 15-Feb-2000 23:00 GMT
In reply to Comment 20 (thedoctor):
Good, I could see cheaper chips running faster being a real benefit, as long as they are cheaper. In mass production something may be better but more expensive in the long run. If AMD could produce 10 of its chips at say, $100 a piece and Transmeta produce ten of theirs at $50, when you start talking about millions the trend of prices should reduce. The AMD Athlon chip should cost a couple of $'s and the Transmeta should follow suit.
I know this point is not very clear but you see where I'm getting at, how is Transmeta going to start the mass-production ball rolling to compete with Intel and AMD?
Questions.
When is this going to happen?
700MHz Athlons and faster are already a reality. There is no waiting about.
Is this chip going to run a new Amiga OS / OE?
I'd be quite content with Linux, and maybe BeOS (never used it but looks pretty). I'm an NT / Unix engineer at work so I know what I would want out of a new chip or OS / OE.
Is it going to be 128bit?
It's about time they came into the mainstream.
"That's all simple math"?
Do you mean maths, I'm not American.
"Nay sayers should get better facts"?
Don't be rude, I wasn't rude to you.
Emulation, running, translating on the fly?
Could you clarify this please? In a corporate business I don't want to have to deal with emulation, that's for working with yesterday's computers like Amiga and SNES which isn't going to happen EVER. Does this mean that the only OS available for it will be in emulation, you will never get the performance out of emulation you can out of native code. Does this mean that software for this chip will come in a non-compiled state like Linux/Unix code?
What does CEO stand for? (Always wanted to know that)
Transmeta don't plan 68k support : Comment 25 of 33ANN.lu
Posted by Anonymous on 15-Feb-2000 23:00 GMT
In reply to Comment 20 (thedoctor):
> You are way out in left field in even starting to compare prices, on TM versus
> the X86 clone chips, the top of the line 700mhz athlon and P3 are OEM at $500 plus,
> and the announced price of the 700mhx TM is up to $150...and 700mhz on 128 bit
> instruction is still 2 times speed of 70mhz on 64 bit instruction...
You can't just compare 'MHz' and 'bits' like that. Remember that these are two completely
different processors. For the same reason you can't compare the MHz of a PPC G4 and a x86.
OK, the TM can emulate the instructionset of a x86, but that doesn't mean that the TM
IS a x86.
> Thats all simple math...
No, not the way you are counting ...
> Nay sayers should get better facts.....
alleluia
> And transmetta never said that what the processor was doing was "emulating",
Well have a look at it again, and you will see that it IS a kind of emulation. You see that,
don't you?
> thats is being used as the term when a real terms for this new process, has yet to be
> established.....What the CEO said in the presentation was something like "a new
> process of translating on the fly, for lack of a better term", and the moderators
> all injected Emulation to discribe..
But what is emulation? It is "a process of translating on the fly". Haage&Partners 68k emulator
for PPC processors are doing just that. And you should always take what CEO's are saying in
presentations whit a grin of salt. After all, they are just salesmen selling their products ...
Transmeta don't plan 68k support : Comment 26 of 33ANN.lu
Posted by Plain English on 15-Feb-2000 23:00 GMT
In reply to Comment 25 (Anonymous):
Well said, you'll never get the performance in emulation you can in native code.
Transmeta don't plan 68k support : Comment 27 of 33ANN.lu
Posted by Mario Saitti on 15-Feb-2000 23:00 GMT
In reply to Comment 19 (sutro):
> Transmeta aims at embedded systems ("Mobile Internet Computing" , strangely
> it reminds me of "Multimedia PC").
Today, what about in a year? If it isn't TM it will be someone using similair technology.
> In this section intel's x86 technology , logic and experience are no factors > (compared to Motorola).
> How will Transmeta compete to intel's and motorola's GHzs , copper technology > and 0.1micros process ?
I think you have that backwards. How will Intel and Motorolla match the more capable clockspeeds of a VLIW system? Transmeta is facing increased latencies as a result of using VLIW, that is predominently a software issue for VLIW. Hardware wise Crusoe can already operate at speeds way out of an Athlon's league. AMD, Intel and Motorola on the other hand need a lot of effort to catch up with even the basic Crusoe techniques. Merced is not generation compatible in the same sense as Crusoe, that leaves one hell of a lot of work for Intel to catch up on. Remember this has been researched by companies for almost 25 years to no avail. It is not something even Intel can just pull out of a hat.
> By designing more advanced s/w ?
Both software and hardware. Currently Crusoe uses a single execution unit for all tasks. That will change in time as the compiler/shecduler becomes more refined. The more refined the algorithm, the lower the latencies-> increase in ILP -> faster the clock speeds. IBM's BOA is already comparable to any Superscalar processor on the market.
> As for IBM , if you saw the "Transmeta announcement" fiasco on 19th of Jan
> you should remember they had a hard time persuading reporters that they had a > contract with IBM or even that IBM knew anything about them.
> I believe IBM had been already working at these things way before them.
Way before them. IBM's VLIW lab has been operational since the mid 80's.
> I don't discard the idea (VLIW) but the company (Transmeta).
> As for PPC , they are here to stay for many years (un)fortunately to some
Not in it's current form it isn't. Why make a RISC PPC when one can make the cheaper, faster VLIW equivalent?
> I bet you 've been a Cyrix pc user...
No, just a student of engineering enjoying new uses of technology.
Mario.
Transmeta don't plan 68k support : Comment 28 of 33ANN.lu
Posted by Mario Saitti on 15-Feb-2000 23:00 GMT
In reply to Comment 26 (Plain English):
> Well said, you'll never get the performance in emulation you can in native
> code.
You are correct, with a good emulation platform you can wipe the floor with "native code"... The trick to emulation being applied in the latest models is to break down the barrier between native code and emulated code so they become one.
In computing the word never is a very dangerous word. Use it with extreme caution. Your use of it is very reckless.
Mario.
Transmeta don't plan 68k support : Comment 29 of 33ANN.lu
Posted by Mario Saitti on 15-Feb-2000 23:00 GMT
In reply to Comment 24 (Plain English):
> Good, I could see cheaper chips running faster being a real benefit, as long > as they are cheaper. In mass production something may be better but more
> expensive in the long run. If AMD could produce 10 of its chips at say, $100 > a piece and Transmeta produce ten of theirs at $50, when you start talking
> about millions the trend of prices should reduce. The AMD Athlon chip should > cost a couple of $'s and the Transmeta should follow suit.
Economies of scale play a aignificant role in any business model. You are correct that if TM does not hit "commodity" levels they will suffer pricewise.
However TM's processors being VLIW do not require the costly fabrication that an Athlon would need. So they have advantage there and also the fact that there is less to produce. Less costs less correct?
> I know this point is not very clear but you see where I'm getting at, how is > Transmeta going to start the mass-production ball rolling to compete with
> Intel and AMD?
By offering a viable alternative to the x86 line. Only that way.
> Questions.
> When is this going to happen?
> 700MHz Athlons and faster are already a reality. There is no waiting about.
This is a real issue for TM, they need income to finance their research. It appears that they will get it from the number of partners revealing products running Crusoe chips. Currently noone knows just how fast the Crusoe can really run before speed becomes an issue. I do suspect that the initial speeds are left that way for viability purposes, who would believe a 1.2 GHz processor that runs x86 from an unknown? People in general do not believe an Athlon is faster than a PIII and they run at similair speeds. Transmeta must first become an established player before they can even dream of outpacing other companies and expect market confidence in their products.
> Is this chip going to run a new Amiga OS / OE?
> I'd be quite content with Linux, and maybe BeOS (never used it but looks
> pretty). I'm an NT / Unix engineer at work so I know what I would want out of > a new chip or OS / OE.
If they run on x86 then yes.
> Is it going to be 128bit?
> It's about time they came into the mainstream.
Not yet.
> Emulation, running, translating on the fly?
> Could you clarify this please? In a corporate business I don't want to have
> to deal with emulation, that's for working with yesterday's computers like
> Amiga and SNES which isn't going to happen EVER. Does this mean that the only > OS available for it will be in emulation, you will never get the performance > out of emulation you can out of native code. Does this mean that software for > this chip will come in a non-compiled state like Linux/Unix code?
Read my posts, I give a layman explanation of how it works. Mail me in private at saittim@cytantet.com.cy if you want the engineer version. But basically it means you can run any software designed for the VLIW abstraction system's chosen ISA. In Crusoe's case X86 or IA32.
As for people who use other words for the abstraction. It IS emulation. There is no other word for it. The native VLIW ISA is abstracted with an emulation layer which provides a stable permanent ISA. In Crusoe's case x86.
As for working with emulation, you have been working with emulation transparently for years on Amiga. Emulation does not have to be in your face.
The 68k series uses emulation to implement missing instructions on Amiga. Would you not use an Amiga because it uses emulation? The issue is to make it transparent so people do not notice. As you did not realise this was the case on Amiga.
> What does CEO stand for? (Always wanted to know that)
Chief executive officer:)
Hope this helps.
Mario
Transmeta don't plan 68k support : Comment 30 of 33ANN.lu
Posted by Anonymously Named Novice on 16-Feb-2000 23:00 GMT
In reply to Comment 24 (Plain English):
> Emulation, running, translating on the fly?
> Could you clarify this please? In a corporate business I don't want to have
> to deal with emulation, that's for working with yesterday's computers like
> Amiga and SNES which isn't going to happen EVER. Does this mean that the only
> OS available for it will be in emulation, you will never get the performance
> out of emulation you can out of native code. Does this mean that software for
> this chip will come in a non-compiled state like Linux/Unix code?
This is the exact point of using Transmeta's technology. It'll be TRANSPARENT to
the user even though the emulation/translation is taking place, whereas current
PPC solution for Amiga will always need to run emulators (in a separate process)
to be compatible with older softwares and even AmigaOS. From looking at the
benchmarks for Crusoe processors, they may not be faster than the top of x86
lines but not as slow as 68040 speed that H&P's emulator will be (not that I'm
saying that H&P did a poor job).
I'm disappointed that so many people have downplayed the importance of
Transmeta technology or not even understood a single bit about it.
Transmeta don't plan 68k support : Comment 31 of 33ANN.lu
Posted by Plain English on 16-Feb-2000 23:00 GMT
NT on Alpha.
I know this was never a popular platform, as far as I know the only reason it came about was because of Microsoft's ties with IBM about 5 years ago. There was never much software that came out for this platform, a few server-side pieces of software and that was about it. It could however run software written specifically for the x86, software was emulated on the fly but in a different way.
As the software ran it was emulated in the standard fashion but instead of running that process every time it would use the translation process applied to that application and make a new born application. It wouldn't translate the entire application at once, only when new parts of the application were accessed by the user it would translate them parts too. e.g. Accessing the clipart application in Word but not straight away.
Something like the following process.
Windows NT (Alpha) running
MS Word 97 (x86)
MS Word 97 executed -> emulation process -> application information saved to a new .exe and .dll's and from then on those would be the ones executed.
This was done because RE-emulating it all the time would take unnecessary processor power.
My question is, would this be the same for Transmeta or would they be simply emulating all the time.
PS
The alpha chip was powerful enough to do this sort of work, so even in emulation it was comparable to a 200MHz Pentium which is a very usable machine.
Transmeta don't plan 68k support : Comment 32 of 33ANN.lu
Posted by Shaun Murray on 16-Feb-2000 23:00 GMT
three points.
1) Everyone is missing the point of the crusoe technology. It doesn't just emulate a CPU architecture, it can emulate the whole of the PC's architecture (or any other for that matter). Ie. do what UAE does in hardware.
2) I'm surprised they haven't got 68K on the horizon as crusoe would maybe make a good processor to get the PalmOS people off low powered Dragonball chips which are pretty much 68K cores with added memory controllers, LCD panel controllers etc. The problem is that so far the power requirements of the chip are still a little on the high side for PDA computing.
3) As the crusoe is able to emulate most of a PC's architecture you can replace more than just the CPU. This is an obvious gain in mobile computing. Battery life on laptops will start getting to useful levels and the less components and cheaper battery technology required is going to bring down costs.
Shaun
Transmeta don't plan 68k support : Comment 33 of 33ANN.lu
Posted by Mario Saitti on 16-Feb-2000 23:00 GMT
In reply to Comment 32 (Shaun Murray):
> 1) Everyone is missing the point of the crusoe technology. It doesn't just
> emulate a CPU architecture, it can emulate the whole of the PC's architecture > (or any other for that matter). Ie. do what UAE does in hardware.
Not missed, just not mentioned. Also it does not stop at UAE's limitations, it is closer to FX32 than UAE. For the benefit of others here, (I will try to make this as simple as possible) there are numerous reasons for using this combined hard/soft methodology. Rather than just use hardware assists to speed up emulation(The easiest to grasp example is Crusoe's translation cache which stores the already translated "commands"), it also uses software trickery to alter the program running itself optimising it as it runs, this is done on numerous levels one such level being dynamic recompilation(Basically rewriting/optimising the code) as FX32 on the Alpha.
> 2) I'm surprised they haven't got 68K on the horizon as crusoe would maybe
> make a good processor to get the PalmOS people off low powered Dragonball
> chips which are pretty much 68K cores with added memory controllers, LCD
> panel controllers etc. The problem is that so far the power requirements of
> the chip are still a little on the high side for PDA computing.
There are numerous reasons why it might not be on the horizon. Firstly the complexity. They can't just swap architecture. Secondly the time frame. The very fact that it took so long to produce an X86 emulation suggests it could be at least 2 years for 68k.
VLIW compilers are one the most complex aspects of computing I can think of, it is the slowest aspect of computing in regards to progress iirc. Now knowing that, imagine the burden of emulation piled atop of an already complex task, then add dynamic recompilation to the equation and look at the task before them.
They have to evaluate whether the development costs can be recouped. It appears that they already have.
Anonymous, there are 33 items in your selection
Back to Top