thanks for your replay and for your work.
Let me check, please to fix a starting point:
- I'm interested and I'm working with the eLink Cluster of (4) parallella boards z7010, in the model of "interrupt_connected" example, described in the doc "Parallella Epiphany Cluster" by Aviv Burshtein, based on << pubuntu- 14.04-esdk.2015.1-20150130>>
- The same eLink cluster and example of "interrupts_connected" doesn't work at all with the last official release <<parabuntu-2016.11.1-hdmi-z7010.img>>
- The old << pubuntu- 14.04-esdk.2015.1-20150130>> doesn't have HDMI support, instead the last official <<parabuntu-2016.11.1-hdmi-z7010.img>> has HDMI support
- at present, the community is working on new release of parabuntu (2018.11.20 unofficial with eSDK2016.11) and eSDK (??)
- the img of the Unofficial Release <<Parabuntu 2018.11.20>>, parallella_e16_hdmi_gpiose_7010.img.gz, does work on my board
- at present Parabuntu 19 is under working as well as eSDK / Supercores (??)
Questions:
- Which test plan will be used to release the Parabuntu 19 and eSDK.2019.xx ? I mean all the original Adapteva test will be applied ?
- May we assume and add the issue about the cluster: the new eSDK.2019.xx should work according the "interrupts_connected" example as well as it worked with pubuntu- 14.04-esdk.2015.1-20150130 ?
Let me know if I can do anything, I could try to test the Parabuntu 19 img as it has been compiled by you or do something with eSDK.2019.xx.
Please take into consideration that I've ready my eLink Cluster of 4 z7010 boards to make some tests.
My project:
I'm working on a C application by which all the 3 slave boards of the cluster run at boot the "slave ready to work" task even if they are Headless.
The Master should be with HDMI and should be able to run anytime a task loading up to 64 different tasks for the 64 cores and receiving or visualize the results (graph)
State of art:
At present, according to a modified and extended version of the "interrupts_connected" example, I've to run anytime all the slave and the master, no way to put in stand by the slave task waiting for the Master upload and go. And it's well working as expected sharing input and outputs values among epiphany RAM and ARM's CPU RAM and synchronizing steps and results. My cores could work synchronized one step per time or anybody by its own.
Thanks a lot, keeping in touch and let me know what I can do.
My best, claudio4parallellaStatistics: Posted by claudio4parallella — Sun Aug 18, 2019 3:45 am
]]>