I am reading about checkpointing. Base on what I have read up now, there are 2 main checkpointing:
System-level checkpointing (SLC) – core-dump style snapshots of computations
Application-level checkpointing (ALC)– programs are self-checkpointing and self-restarting
I am interested to implement in C a checkpoint recovery algorithm in a function level. I am wondering if I can consider this inside "Application-Level" category.
Second of all, is there an available open-source library for that.
I put here a simple adding function as an example:
adding(int a0, int a1, int b0, int b1, int* res0, int* res1)
Here is the algorithm strategy:
Store on buffer (a0, a1, b0, b1, res0, res1)
adding(int a0, int a1, int b0, int b1, int* res0, int* res1)
while (error_flag <> 0) && (trial < 10)
Error_flag = 0;
Trial ++;
Get (a0, a1, b0, b1, res0, res1) from buffer;
adding(int a0, int a1, int b0, int b1, int* res0, int* res1);
If (error_flag == 0)
Unstore (a0, a1, b0, b1, res0, res1) on buf
Else error_message and stop program
Is there a way to write "Storing" section (first Line) in a general format. How about if the function gets different arguments types.