Search the Community
Showing results for tags 'thread'.
Found 4 results
-
Hi all, I have a systemC-tlm code that uses SC_THREAD( ) to implement parallel modules. I am working on latency analysis and hence included a few wait statements in each of the SC_THREAD functions. It was working fine for a long time, until yesterday. The SC_THREAD functions with wait(time, SC_NS) hang exactly at the wait statement. I tried debugging into the wait function, and found that the code throws an SC_REPORT_ERROR in wait( ) function in sc_wait.cpp. I am not sure if that error is caught, because I do not see any error in my console output. Does anyone have a clue on why I am seeing this behavior? -
Hi, I am new to SystemC. I have two questions. 1) Do sc_thread generates separate Linux thread for each call and what is the difference between the POSIX thread and sc_thread? 2) Why systemC has no parallelization scheme? The SystemC scheduler uses cooperative multitasking and cannot exploit fully the potential of SMP workstation.
-
I had modeled a spi module in which i have a thread for the module, and a testbench thread to test the module. I am using MOSI and MISO lines for data transfer. I have a CLK line, which will be zero if slave is reading or writing data. And 1 when master is writing or reading data. Master have the control over CLK. So slave has to wait for master to make CLK line zero and suspend itself. After reading and writing on the line, Slave now suspend itself to give master control again. But the problem is master is not suspending. It will just execute wait(); and go for the next line. I am including the lines of code where the error is occurring. Sensitivity has anything to do with wait() statement? Module Code: if (rd_buf == 0) { DRIVE.write(false); // Here, clock is made zero, and waiting for slave to write on MISO pin... wait(); // Not suspending and jumping to testbench... for (char i = 7; i >= 0; i--) { DRIVE.write(true); rd_buf|= MISO.read()<<i; cout << "MISO in spi " << MISO.read() << endl; DRIVE.write(false); wait(); } } Test Bench: for (i = 7; i >= 0; i--) { if (clock.read ()== 0) { cout << "write_buff contents " << ((write_buff >> i) & 1) << endl; miso.write(write_buff >> i & 1); wait(); } } Thank you
-
Hi All: I got a given SystemC code running without end_of_simulation called. And I know "end_of_simulation" won't be called if sc_stop is not called. I debug into the code and see following situation. Code looks like void thread_a() //it is a sc thread { while(xxx) { wait(sc_time(x,SC_NS)); break 1 cout<<"xxxxx"<<endl; break 2 ---logics---- } sc_stop; } The last time I was stopped at breakpoint 1, then I did step next. Then instead of stopping at breakpoint 2 , I stopped at sc_main after sc_start. So it looks like the entire thread is stopped by others? What else than "sc_stop" can terminate the thread? Since my code in "end_of_simulation" is not triggered, I am sure that there is no other sc_stop in my programme that calling sc_stop. So what can be the backend hand doing that? Thanks a lot for all your help
