During this holidays, we managed to create an architecture for our project HARP, but today, first day of school, we discovered that this architecture was inappropriate.
So we decided to build a new one,
In this one there will be still two parts:
one software part which will be on the computer and one system on the chip part. But we change what where on our pcb.
Data flow and data rate are two critics point in our project, we have to manage that each parts will be able to get a DATA applie some functions to it and send it as fast as it can .
For that we change three things for now:
- We will compressing the data between the computer and the pcb. in order to maximise the data flow
- It order to decompressing the data on the pcb we decided to add a gumsticks with wifi to our architecture, this gumstics will handle with the fpga and control all the part of the pcb.
- for now one we will have a massive parallel architecture : 4 Ram (one for each drivers) instead of one huge RAM, we decided to split it into 4 little RAMs with lower access time but still allowing writing/reading at the same time.
we started today a document with a summary of all our new architecture, we will published it normally tomorrow so will have much more technical information.
We also, Paulin and I, meet Mr Lefeuvre, a 3D specialist, in order to ask him some questions about how we could handle the data used for normal 3D software simulator in order to transform it into a cloud of point used in our project.
Paulin will soon publish a summary of this interview , and also what are the ways we have found to handel the 3D data
but now it’s time to go to bed, good night everyone.