[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [f-cpu] Status quo



Le 2015-03-31 22:02, Tobias Bergmann a ÃcritÂ:
Hi,

I wanted to let some time pass before I give my opinion on the project
as well. Those that are able to contribute code should always have the
first say!

Sure. For now I'm evaluating the situation and looking around,
trying to understand who is there, where we could all progress etc.

I know that the only way to really help an open source project is by
contributing working code. Unfortunately I can't do that - at least I
got that impression when reading my work contract(*) - but I will keep
an eye on the project since my belief that the world needs a high
performance open and free CPU has only grown over the past decade.

People can contribute in other ways than just code.
I suppose you're german, and german translations are desirable.
It is my experience that translations are not easy but have a direct
benefit : they uncover the inconsistencies of the original texts.
I know I can write things "that I understand" without realising that
others may not have a clue what's in my tired mind...

If you know the CPU field, you can find semantic problems
and dark areas easily. Translation does not have the same
weight as actual working code but it's almost as important IMHO.

One way to avoid too many discussions is to close part of the project
and only work on implementation there to get it done.
All new, fancy, clever and cool ideas get pooled and evaluated for the
successor(s).
That way more than one generation can be in flight at the same time
without the clever ideas postponing the implementation indefinitely.
That makes sense though in practice, it's not as easy.

This mailing list is the historic place to discuss and I don't see how
to change that. I even wonder why I would change that and break the
community.

Instead, let's say that this list is "for discussion and organisation"
and actual implementations happen on other groups/lists.

Here is a list of potential lists :-D

 - YGWM : I started http://ygwm.org/ and this is a priority.
     there is a lot of code to clean up and refactor.
     This will become the basis for the future sites.
     If there are JavaScript lovers, this is your chance :-)

 - the YASEP : this is my sandbox http://yasep.org/ and
     I experiment ideas, tools, methods and approach,
     so this is also where the future development are influenced,
     not in their architecture but their implementation.
     The YASEP is also the microcontroller that will manage
     the F-GPU.

 - F-GPU : the first platform for the F-CPU needs hardware hackers !
     If you're a PCB/BGA/GPIO/FPGA specialist, please apply ;-)

I don't know yet where those lists or forums would be hosted.
I admit I'm not a great admin/webmaster. Again, help is welcome
in this domain :-)


Nonetheless there should be a way to discuss and examine *future*
ideas as well. They need time to mature. It even could be a separate
list if we want to avoid confusion.

I agree but I must admit that this list is the natural place where it occurs.
This is where people come to see and toss ideas around. It's a sort
of chaos but it's alive :-P
Subprojects-related lists will have less noise because only interested
people will go there, while F-CPU is still a pretty broad subject in itself.

I agree again that the world needs more openness and a thing like F-CPU
is long overdue. However I have seen that 1) the architecture is not everything,
2) the suitable design/development tools are absolutely necessary.
And we can't do like people on OpenCores who post their design files "and that's all".

So I'm sorry for the people want want to discuss CPU features.
For now I stick to the latest version of the architecture defined in the F-CPU manual, because it's important to work on actual results and build the necessary tools
that don't exist yet.

Fortunately today we have http://ghdl.free.fr/ and we don't have to deal
with a mess of diverse proprietary tools :-) Hopefully, with more work,
we'll have a coherent hardware and software development environment
and more people will be able to play with architectures :-)

Just my 2cents,
Tobias
yg

(*) I worked the last few years as a hardware designer on 5GHz+ mainframe CPUs
swoooon :-D

*************************************************************
To unsubscribe, send an e-mail to majordomo@xxxxxxxx with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/