Archive for the 'Largos' Category

Las version 0.10 “Flying Pigs” release available

Sunday, June 14th, 2009

I’ve released v0.10 of my Largos assembler so that all of you can have a go at alpha testing it. There is a README file in the archive that describes this release. I welcome error reports and feature requests, but I won’t be able to make any changes to Las until after UPEW 2009.

Las verion 0.10 “Flying Pigs” pre-release

Las “alpha test” release this week

Monday, June 8th, 2009

I’ve been finishing my large model assembler for the Propeller, and it is now generating correct code.

Las was designed from the ground up to make writing LMM code easy, and as such, it has to depart some from the Propeller pasm syntax for the pseudo operations. The name and syntax for all instructions remains the same.

I still have to finish some of the pseudo operations, however I expect to have an executable for people to try later this week. For now, the executable will be for Windows only, but later I plan Linux and OSX versions. I used some proprietary code in writing the assembler, so source will not be available.

“Las” stands for both “Large model Assembler” and “Largos Assembler”.

I should be posting the alpha version here by Friday, June 12th.

Some Largos information

Saturday, June 6th, 2009

I’ve posted a brief feature list of Largos on its own page – you can find it under “Pages” on the left navigation bar.

:-)

Largos processes over 2200 messages per second!

Saturday, June 6th, 2009

I’m working on Largos, my LMM operating system for the Propeller.

I am trying to get a decent “technolgy preview” ready for UPEW 2009 – and today I’ve been able to start writing code that accesses the kernel via my message passing object.

The message passing object is currently written in Spin, however the kernel is still able to process almost 2300 null messages (messages that don’t do anything except acknowledge that they executed) per second on an 80MHz Propeller!

The messaging overhead is quite acceptable for something written in Spin: 0.47ms

As time progresses, I will be replacing the Spin components with LMM code – I then expect the messaging overhead for LMM programs to be on the order of 5us-10us for null messages, or a speed increase of approximately 50x-100x.