http://www-128.ibm.com/developerworks/power/pa-slof/

The problem for me is that i don't think the S1 Core is powerful enough.

Edit:

Have a look at this:
http://www.opensparc.net/component/option,com_mtree/Itemid,106/task,viewlink/link_id,9/

I'm sending an email to them now.

Yep, that's true. I remember reading that a while ago. Anyways, the FPU performance is not or shouldn't be a problem with the S1 Core since it only consists of 1 Core.

I just got a reaction to my question from the Simply Risc Team (very fast ).

Fabrizio

---

the S1 for sure is powerful enough to run a complete OS
on a desktop machine.

What is lacking now is a complete functional verification
since we are not getting much free support from Sun anymore.

---

Edit:

http://forum.java.sun.com/thread.jspa?threadID=5106875
http://forum.java.sun.com/thread.jspa?threadID=5107353

i'm seeing a revolution in front of my eyes:

Sun Sparc and Open Solaris GPLv3.

This is the future

Here are the current results of the CPU poll which we have been running since the start of the project.
http://www.innovationstage.com/openoem/polls.php?id=2
x86 i.e. Intel/ AMD
65.4% (170)
PPC i.e. IBM/Motorola
5.4% (14)
Cell i.e. IBM/Sony/Toshiba
9.2% (24)
Sparc i.e. Sun
7.3% (19)
Other (Open Cores?)
12.7% (33)

260 voters since 2007-01-09 00:06

As you can see it clearly indicates a preference for the x86 family of processors.

Here are the results of the poll for what type of Computer people want:
http://www.innovationstage.com/openoem/polls.php?id=3
Open Laptop: 15.4" Screen
64.5% (129)
Open Desktop: 19" or 20" Screen.
30% (60)
I don't want a Open Computer.
5.5% (11)

200 voters since 2007-01-12 18:01

and here are the results of the poll to decide upon an operating system.
http://www.innovationstage.com/openoem/polls.php?id=1&comments=1
GNU/ Linux: Debian
13.8% (9)
GNU/ Linux: Ubuntu/Kubuntu
43.1% (2
Fedora Core
4.6% (3)
OpenSuSE
1.5% (1)
FreeBSD
4.6% (3)
Mandriva
0% (0)
NetBSD
0% (0)
OpenBSD
1.5% (1)
Another GNU/Linux Based Distro
7.7% (5)
Another BSD Based Distro
3.1% (2)
Microsoft Windows
4.6% (3)
Haiku or Zeta
10.8% (7)
None of the Above
4.6% (3)

65 voters since 2007-01-09 00:01

I don't think we can be any fairer than this and these polls have been running since the start of the project.
And the preferences stack up like this:
a 15.4" widescreen laptop using an x86 family processor running some form of (K)Ubuntu.

Dale, I did not mean to accuse you or anybody in charge of being unfair; but you where bringing the subjectc out of order and declaring the vote closed without a warning. Those things are what I beleive to be the incorect thing to do in a democratic government.

Let us take the OS vote for example. You opened the discussion and the vote both at the same time. Some voted straight away, and some discused for a while. After some time, the vote was declared effectively closed without warning with only one choice left (gNewSense).

dale

---

I don't think we can be any fairer than this and these polls have been running since the start of the project.

---

guyjohnson

---

A lot of votes were made before a lot of the issues were discussed. For example, I voted for x86 for the processor, before I realised there weren't any available with open specs.

---

See what I mean? Opening the voting too soon, but stopping people from changing their votes, is a trap. Guyjohnson fell in it this time, I might fall the next.

In my opinion, this is what should be done: A subject is called for a vote with the oppening of a new forum. In the call for a vote document (the forum opener), it should be clearly especified the date of the closing of discussions (and the opening of votes) and the date of closing of the vote.

Once the day of the voting arrives, the discussion forum should be closed. Someone with admin powers should post one message calling for the vote. Email could be used for voting, or identified HTTP (by means of a browser session or cookie).

Once the day of the closing of votation arribes, another post should be made to the discussion thread declaring the voting finished and the posting the results. One item on the voting report should be "did not vote".

Once more, I didn't mean to sound harsh. Sorry for not making myself clearer on my last message. I do hope this one gets a better fate.

Can I change a vote I previously issued? Or more more to the point, can guyjohnson do it?

If he can, I shut my mouth.

Now, I did not propose you to hiper extend the voting period. My proposal is completly differently. Look again at what I have previously said in this thread:

vrrivaro

---

In my opinion, this is what should be done: A subject is called for a vote with the oppening of a new forum. In the call for a vote document (the forum opener), it should be clearly especified the date of the closing of discussions (and the opening of votes) and the date of closing of the vote.

Once the day of the voting arrives, the discussion forum should be closed. Someone with admin powers should post one message calling for the vote. Email could be used for voting, or identified HTTP (by means of a browser session or cookie).

Once the day of the closing of votation arribes, another post should be made to the discussion thread declaring the voting finished and the posting the results. One item on the voting report should be "did not vote".

---

I am willing and able to make all the required modifications to your PHP, as you possiby need.

I think just CPU and OS but this has to be it. We can't keep going over the same old subjects. I would like to propose a period of 30 days discussion to be followed by a 1 day voting period.
Discussions to close on the day of voting and final tally to be the decision going forwards.
The CPU candidates are to be decided in the next couple of days as are the Operating System Candidates.
The Only criteria I would like to see is that any candidate CPU's or OS's chosen for the vote are actually available and usable as we set out.
Everybody who is interested should get involved in the discussion and vote.
Dale.

These are big decisions and I don't see a problem with opening up discussions for 30 days. I want this to be the final discussion and the decision to be final. I think if we rush it we are just going to end up in the same situation again where people are not happy with the way the decisions have been made.
This timeframe will give everybody a chance to weigh in on all the options.
I would like us to now assemble a shortlist of OS candidates and CPU choices (with the prime consideration that they are deliverable products)
Linux/BSD/BeOS/(open)Solaris et al.
x86/x64/PPc/Power et al

Dale.

Ok. So I'll create two new forum topics just to talk about the CPU and the OS. They will be active for 30 days and will be closed after that.

Sounds like a great idea!
Who do you have in mind?

I'm going to try and make a list and see if all of you agree with me. I'll put that list soon. Thanks!

I can tell you Richard Stallman's email address, it's rms@gnu.org .

Nikolay Dimitrov
26.02.2007

First of all, remember these words: "Whatever they talk to you, it's always a question of money!". Second, this might look like out-of-order text lines, I just dumped my thoughts. Third, I'd like to discuss technical questions, not my spelling or grammar. Fourth, I wish to say that the following text is only a friendly, constructive criticism, exposing my own opinion as a FLOSS fan, IT specialist, gamer and HDL newcomer (see terms section).

I read the OpenOEM forum thread "CPU Choices for an Open Design", and I like to share my point of view regarding this question (I think it can be too late for the choice, but it's never too late to share an opinion at such complicated question with other people).

Reading through the whole OpenOEM site, I see the words "open" and "free" used interchangeably by site's authors and members. I really think that it's important to differentiate these terms, and to use them in a proper way.

I personally understand the term "open" as a specification / plans / schematic / sources which are accessible by everyone, but its general direction of development is guided by a small group of developers or at least independent organization body, interested in the subject developed. I think a vivid example of such design is (my favourite) OpenRISC cpu (the project's url is in the links part). The project is driven by a small team of educated specialists and it's design docs are freely available at no cost. Another example is the SPARC architecture, with the official SPARC v8, v9 (a.k.a. UltraSPARC) docs, OpenSPARC, SimpleRISC derivative. The common thing between them is that the design is made by a small group, self-elected of by the community, and not by the whole community. Decisions are made mostly by a technical experts, and not by the community as a whole (mostly and users).

The term “free”. I understand it like “GNU/Linux”. Everyone can work on it, everyone can contribute, of course after the kernel-maintainers agreement and control But mostly, it is a result of a community effort. Example is already given ;) . The documents and information produced by such projects is available under the same terms if use as the “open” projects.

So, what is "open", what is "free" and what the proposed hardware / software design must be guided from.

The question remains – will the OpenOEM be an “open” or “free” project? Remember that we are discussing the "open" or "free" aspect of cpu architecture. If we are to be building an “open” project from the grounds up, it is more likely to be able to use an already existent cpu architecture, because it's:
- cheaper (not have to order custom chips, thus having to pay at least 1,000,000$)
- easier (not have to develop own ISA and cpu) it's faster (time-to-market factor)
- it's safer (already deployed and have a criticall mass of software / end users)
- it's legally safer (IMPORTANT! Buying chips is not a concurrent business so this avoids threats from big companies!)

So, here is a list of some architectures, currently available as an already factored chips:
- x86-compatible (the technically worse architecture of all available, the only reason to survive last years is Intel's demonic efforts and money, a few sources: Intel, AMD, VIA, others?, has most software and easier for end-users, awful for OS/driver developers, emulation on RISC will be slower than native execution, if we develop own cpu to run x86 Intel will go after us (AMD/VIA cases))
- PowerPC (technically better x86, sources are IBM, has smaller software share than x86, but several important devices are made with it – Gamecube, Wii, PS3 (Cell), because of cheaper IP core, IBM will go after us for every IP core that looks to be compatible with their instruction set)
- ARM (technically better than x86, market leader in embedded systems, has good performance, affordable chips, easy to make custom boards, if we make our own, ARM will go after us for every ISA compatible with theirs)
- MIPS (technically better than x86, one of the best RISCs ever made, don't remember who is producing them now, MIPS threated anyone who designed MIPS-compatible IP cores, just look at opencores.org, BUT THEIR PATENT FOR UNALIGNED LOAD/STORE EXPIRED IN 2006Q4!!!)
SPARC/UltraSPARC/OpenSPARC/SimplyRISC(technically better than x86, very well designed RISC, freely available docs, Sun greets anyone who wishes to make a compatible chips, has available Verilog sources for several chips: microSPARC, SPARC v8 (LEON, LEON2, LEON3), OpenSPARC, legally safer for anyone who wishes to built his cpu/board without paying lots of money and/or going to jail...)

So, looks like all architectures are better than x86, but it's the only one that occupies our laptops/desktops (98-99% in my opinion). It's again all about the money...

So, lets put the things straight:
Which one we choose: OPEN or FREE?
What will we do: DESIGN NEW CPU or USE EXISTING CPU ARCHITECTURE?
Which is more important: USING NEW FREE HARDWARE or USING OLD SOFTWARE?
Which is more wanted: DESKTOP or LAPTOP?

Hints:
GNU/Linux and the *BSD group operating systems are available for all major cpu architectures, thus making the choice of end-user easier – just run your software on which architecture you have at hand.
Open and Free Projects are about innovation, and innovation means that old things belong to the past. Backware compatibility is good, but cannot long forever without new ideas to suffer from this.
I use x86, but don't like it, because there is no viable (cheap and supported) alternative solution to work with AND to use multimedia/games.
People at forums are just dreaming that the OpenOEM project will give results in a month or so, and they will play their games/run their software on the new system. Most of them even do not realise how big task is it, and how these big problems are to be solved in a complex ways, to bring a viable solutions as good platforms for them. I just cannot imagine someone says “I need a x86 compatible solution for a playing windows games”, and someone to answer “Ok, another thing?”.
Don't get me wrong. I've used/am using assembler, Delphi, C, C++, Java programming languages, listen music, watch movies, play games (currently World of Warcraft), but when things are getting to choose the RIGHT(tm) thing, x86 is not the ultimate way to go, and buying chips from vendors will make us dependent of them. And this is not good.

Thank you for reading all these emotional words!

P.S.: After reading the forum threads, I tried to think about a generic solution for the cpu architecture. So, there will never be “a proper way” to go, every decision made will be partly right and partly wrong, and may be a new VLIW cpu, microcoded for one/several architectures can do the work ;)

Glossary of terms used:
FLOSS – Free/Libre Open Source Software
IT – Information Technology
HDL – Hardware Description Language
ISA – Instruction Set Architecture
RISC – Reduced Instruction Set Computing
VLIW – Very Large Instruction Word

Links:
http://www.sparc.org/standards.html
http://www.opensparc.net
http://www.srisc.com/?s1
http://www.arm.com
http://www.mips.com
http://www-03.ibm.com/chips/power/powerpc/
http://www.opencores.org/projects.cgi/web/or1k/architecture
http://www.opencores.org/browse.cgi/by_category

All mentioned trademarks are copyrighted by their respective owners.

elan

---

I think nobody will interest in x86 maschine.Everybody can buy it in nearest shop and install linux there...

---

Is this the only reason people buy x86 hardware? I don't think so.

elan

---

I think OpenSparc or MultiCore Xscale will be cool.

---

Yep. Cute blonde will be cool too . Coolness is an interesting way to choose things, but only for geeks. We are in a position to choose hardware platform for our machines, and maybe our children's machines. So let's be practical.

P.S.: I don't propose hard x86. I don't have anything againts OpenSparc, XScale. I don't have anything against blond girls (actually, brunettes are cool too )

PicMaster