Abassi RTOS for megaAVR

Highlights


    1. Fully scalable footprint (as low as 1250 bytes on ATmega128; full: ~4425 bytes)
    2. Fast Execution (113 clock cycle context switch, 7.06μs on ATmega128 @16MHz; 23.3μs task switch on semaphore posting)
    3. Unlimited Tasks, Queues, Events, Semaphores, Mutexes and Mailboxes
    4. Flexible licensing options
    5. Available source code
    6. Comprehensive documentation
    7. Responsive Technical Support

Unmatched Feature Set

In addition to the standard feature set supported by most real-time kernels, the Abassi RTOS can be configured with many features unmatched in the industry. These features add robustness and code savings, and include:


    1. Intelligent starvation protectionGuarantees fair access to CPU, via enhanced priority aging, even on heavily loaded systems.
    2. Priority inheritance, including dynamic trackingAutomatically adjusts for priority aging, timeouts, etc. and propagationBetween entangled mutex owners, to prevent deadlock.
    3. Adaptive priority ceiling
    4. Deadlock detectionReports recursive mutex dependencies anywhere in the execution chain.
    5. AsymmetricTimeslice adjustable at a per task level. Round Robin scheduling
    6. Hybrid interrupt stack

Development Tool Choices

Abassi supports major megaAVR development tools, including AVR Studio and IAR.

Memory Requirements

Don't settle for half a RTOS! Even with all features enabled, Abassi still consumes less memory than the minimum configuration of most kernels. And you get advanced priority inheritance, asymmetric round-robin scheduling, starvation protection, and much more. Or, if memory footprint is paramount, Abassi scales to a minimum size that can't be matched.

"C" Code Memory Usage
Description Size
Minimal Build <950 bytes
+ Runtime service creation / static memory <1200 bytes
+ Multiple tasks at same priority <1325 bytes
+ Runtime priority change
+ Mutex priority inheritance
+ FCFS
+ Task suspension		 <2000 bytes
+ Timer & timeout
+ Timer call back
+ Round robin		 <2650 bytes
+ Events
+ Mailbox		 <3400 bytes
Full Feature Build (no name / no runtime creation) <3775 bytes
Full Feature Build (no names) <4125 bytes
Assembly Code Memory Usage
Description Size
Assembly code size 294 bytes
Vector table (per interrupt) +4 bytes
Interrupt prologue (per interrupt) +8 bytes
Hybrid Stack Enabled +42 bytes
Nested Interrupts Enabled +14 bytes
Data Memory Usage
Description Size
Base Kernel 106 bytes
+ Task 16 bytes
+ Semaphore 6 bytes
+ Request 6 bytes
+ Mailbox 14 bytes
Measured using current software release, configured for ATmega128-16 at 16 MHz and IAR Embedded Workbench for Atmel AVR, Version 6.11.1

For your exact data memory requirements, please consult the Porting Document that applies to your design environment.

Kernel Snapshot

  1. Fully preemptive
  2. Scalable
  3. Very small code and data footprint
  4. Code can reside in ROM
  5. Zero interrupt latency kernel
  6. Mutexes
    1. Configurable priority inheritance
    2. Adaptive priority ceiling
    3. Priority based and first-come, first-served
  7. Versatile semaphores
    1. Counting and binary
    2. Priority based and first-come, first-served
  8. Mailboxes
    1. Priority based and first-come, first-served
  9. Events
  10. Intelligent starvation protection
  11. Hybrid interrupt stack
  12. Deadlock detection
  13. Single or multiple tasks per priority level
  14. Configurable Round Robin
  15. Dynamic priority changes
  16. Compile time task and component creation
  17. Run-time safe service creation
  18. No fixed number of tasks or components
  19. Programmable timeout on components
  20. Selectable scheduling search algorithm
  21. Optional cooperative mode
  22. Debug logging