Jump to content

AmeyaVS

Members
  • Content Count

    152
  • Joined

  • Last visited

  • Days Won

    25

Everything posted by AmeyaVS

  1. Hello everyone, Has anyone faced issues with Linux Build for SystemC 2.3.2 configured with PThreads as the threading library backend? I have configured the SystemC library using the following command: cmake -DCMAKE_BUILD_TYPE=Debug -DCMAKE_INSTALL_PREFIX=$HOME/apps/systemc-2.3.2 -DCMAKE_CXX_STANDARD=11 -DCMAKE_CXX_EXTENSIONS=OFF -DENABLE_PTHREADS=ON -DENABLE_PHASE_CALLBACKS=ON .. # Then followed by make # make check # and make install I having issues with the SystemC kernel getting stuck in deadlock. I am currently working on Ubuntu 18.04 using g++: g++ (Ubuntu 7.3.0-16ubuntu3) 7.3.0 Copyright (C) 2017 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Though I do not experience such issues with the QuickThreads back-end, but for now I need SystemC kernel with PThreads support. I have attached the back-trace of a minimal example/regression test from the test-suite which triggers this behavior:(systemc-regressions-2.3.2) TEST : systemc/1666-2011-compliance/living_dead_bug/living_dead_bug.cpp (11/871) Note: This behavior is also observed with SystemC 2.3.1a release. Best Regards, Ameya Vikram Singh SystemC.RaceCondition.GDB.BackTrace.txt
  2. Hello @katang, Did you set the CXX_STANDARD in CMakeLists.txt? Here is an example: https://github.com/AmeyaVS/SystemC_ramblings/blob/master/src/01_SystemCTest/CMakeLists.txt Also while building SystemC 2.3.2 did you configure it with -DCMAKE_CXX_STANDARD=11? Hope it helps. Regards, Ameya Vikram Singh
  3. Hello @katang, It seems you have mixed C++ standard in your build system. Please have a look here: Hope it helps. Ameya Vikram Singh
  4. Hello @ssingh.codesupport, There are multiple issues here. It would really good if you can go through the following discussions and develop an understanding about the SystemC simulation workflow. In-case you still have questions you can post you queries here. Hope it helps. Regards, Ameya Vikram Singh
  5. Hello @katang, The compiler message is quite right as mentioned by @Eyck. The Compiler sees the code something like this: void Dendrit_Set(int index, int value) { *(mNeuron->dendrit[index].write(value)); } Basically the operator precedence is at play here(refer here for more details). Compiler sees that you are trying to dereference a void type, since the write method of the sc_signal returns void. Hope it helps. Best Regards, Ameya Vikram Singh
  6. Hello @katang, Is there a specific reason for using SystemC 2.3.1 release? This issue has already been fixed in the SystemC 2.3.2 release as mentioned by @maehne. Best Regards, Ameya Vikram Singh
  7. Hello @kjhyland, While adding signals using sc_trace did you provide the hierarchical names of the signals? Best Regards, Ameya Vikram Singh
  8. Hello @LookBad, Sorry, I don't have a macOS environment readily available with me. Maybe someone else might be able to help you with that, or better would be if you could use a Linux VM for exploration. Regards, Ameya Vikram Singh
  9. Hello @SANDESH M K, You are probably looking for sc_buffer<T>. sc_buffer will generate the event notification for every write performed. Hope it helps. Regards, Ameya Vikram Singh
  10. Hello @LookBad, Can you provide some environment details? e.g.: OS, Compiler version etc. Regards, Ameya Vikram Singh
  11. Hello @YashJ, It is expected behavior. Here is a brief explanation: //... // Find the inline comments: while(true){ wait(); //< Wait for default event sensitivty(static sensitive event!). In you case it is in.value_changed() cout<<"\nIN = "<<in.read()<<" TIME" <<sc_time_stamp()<<endl; inter.write(in.read()); //< Write to the inter signal -> which means schedule and event for updating the sc_signal. cout<<"\nInter before wait= "<<inter.read()<<endl; wait(10,SC_NS); //< Relinquish control to the SystemC scheduler -> perform evaluate and update phase for all event scheduled. (Change sensitivity to dynamic scheduled after 10 ns) out.write(inter.read()); //< inter has the new value assigned at 3 lines before this statement. cout<<"\nInter after wait= "<<inter.read()<<endl; cout<<"\nOUT = "<<out.read()<<" TIME" <<sc_time_stamp()<<endl; cout<<"\nIN when outed = "<<in.read()<<endl; } ///.... Hope it helps. Regards, Ameya Vikram Singh
  12. Hello @ajinkya_bankar, Can you let us know what all have you tried? and also, what is it you are looking at SystemC for RTL design or high-level modelling? Most of the examples though dated, require minimal changes to be fully functional, and you can quickly update them to Did you look at examples directory within the SystemC open source implementation available here. If you looking at SystemC RTL design then probably this book: SystemC Primer would be good though also a bit dated. In-case you are looking for using SystemC for high-level modelling your a find these references to be quite useful: SystemC: From Ground Up (Second Edition) SystemC Language Reference Manual(LRM) Hope these resources help. Regards, Ameya Vikram Singh
  13. Hello @sumit_tuwien, You can turn off the G++ extensions using the CMake command-line option: # Enable C++11 compilation flag # -DCMAKE_CXX_STANDARD=11 # Disable C++ Compiler Extensions # -DCMAKE_CXX_EXTENSIONS=OFF cmake -DCMAKE_CXX_STANDARD=11 -DCMAKE_CXX_EXTENSIONS=OFF -DCMAKE_INSTALL_PREFIX=<Path to install location> .. You can find a reference here. Can you also list the version and platform information of the tools you are using (clang, CMake, etc.)? Hope it helps. Regards, Ameya Vikram Singh
  14. Hello @veeresh k, There are multiple issues in the sample code provided by you. Here are the heavily modified version: // main.cpp #include<systemc> //< Notice not systemc.h #include"head1.h" #include"head2.h" int sc_main(int argc, char* argv[]) { sc_core::sc_fifo<int> fifo(10); //< sc_core namespace specifier writer w("writer"); reader r("reader"); w.out(fifo); r.in(fifo); sc_start(1000, sc_core::SC_NS); //< Run simulation for a limited time since no stopping condition are provided upfront. Also time is also specified in sc_core namespace. return 0; } //head1.h // Notice the HEADER GUARD #ifndef HEAD1_H_ #define HEAD1_H_ // Changed the header file to systemc instead of systemc.h #include<systemc> SC_MODULE(reader) { // sc_fifo_in_if is a abstract class from which you cannot create an object. // sc_fifo_in is derived from the above class which actually implements the read method. sc_core::sc_fifo_in<int> in; void roperation() { int val; while (true) { wait(10, sc_core::SC_NS); // Time unit is defined in sc_core namespace. for (int i = 0; i <= 15; i++) { in.read(val); std::cout << val << std::endl; //< when using the systemc header you need to specify the std namespace. // Or you can use the statement after including all the header files: // using namespace std } } std::cout << "Availaible : " << in.num_available() << std::endl; //< Instead of num available it is num_available //< when using the systemc header you need to specify the std namespace. // Or you can use the statement after including all the header files: // using namespace std } SC_CTOR(reader) //< Your code mentions here writer the constructor name should be same as the class name. { SC_THREAD(roperation); //< Your code mentions woperation } }; #endif // HEAD1_H_ //head2.h // Refer the head1.h for similar details #ifndef HEAD2_H_ #define HEAD2_H_ #include<systemc> SC_MODULE(writer) { // Refer the head1.h for similar details sc_core::sc_fifo_out<int> out; void woperation() { int val = 0; while (true) { wait(10, sc_core::SC_NS); for (int i = 0; i <= 20; i++) { out.write(val++); } } } SC_CTOR(writer) { SC_THREAD(woperation); } }; #endif // HEAD2_H_ Here is a minimal makefile for building the example: CC=gcc CXX=g++ CXX_FLAGS= -g3 -Wall -c all: main.exe main.exe: main.o ${CXX} $< -o $@ -L${SYSTEMC_HOME}/lib-linux64 -L${SYSTEMC_HOME}/lib -lsystemc main.o: main.cpp head1.h head2.h ${CXX} ${CXX_FLAGS} -I${SYSTEMC_HOME}/include -o $@ $< clean: rm -rf main.o main.exe Note: For reference of the sc_fifo_in/out_if classes see here. For more details you can also refer to the SystemC LRM. Hope it helps. Regards, Ameya Vikram Singh
  15. Hello @sumit_tuwien, You can find the necessary details from the SystemC sources. You can look into these files for more details: RELEASENOTES README NOTICE You can also compare the RELEASENOTES from previous releases. Regards, Ameya Vikram Singh
  16. Hello @vasu_c, Can you let us know what platform you are using for the AArch64 build? and also, if possible the ABI specifications of the GCC compiler as well? Thanks and Regards, Ameya Vikram Singh
  17. Hello @veeresh k, This example does not have a dedicated clock generator. // These statements are what driving the clock signal to with High time of 1 ns and low time of 1 ns. for (i=0;i<10;i++) { clock = 0; sc_start(1, SC_NS);// Run simulation for 1 ns. clock = 1; sc_start(1, SC_NS);// Run simulation for 1 ns. } What message you receive from the SystemC kernel refers to the timescale of 1 ps e.g.: Info: (I702) default timescale unit used for tracing: 1 ps You can get the SystemC API documentation for the sc_clock from here: SystemC sc_clock api reference. If you need a sample you can find a use-case reference here(though a little bit-dated): https://github.com/AmeyaVS/SystemC_ramblings/blob/ff1a111063842bfcd6f5de6bb3db74917dc6331c/src/03_fir/firsytemmain.cpp#L26 Hope it helps. Regards, Ameya Vikram Singh
  18. Hello @Ganesan R, It seems you are trying to achieve the vector of vector instance, something like this: sc_vector < sc_vector < your module > > You can find necessary discussion on this here: http://forums.accellera.org/topic/5616-initialization-of-nested-sc_vector-sc_vector/ Without sample code it is very difficult to discern the scenario you are trying to replicate. Can you provide some minimal sample code? Regards, Ameya Vikram Singh
  19. Hi, It seems I have hit a bug in VCD tracing implementation in SystemC release 2.3.2. Here is the bracktrace captured in GDB: (minimal example available here: https://github.com/AmeyaVS/SystemC_ramblings/tree/dev) as one can observe the exception: SIGFPE in systemc-2.3.2/src/sysc/tracing/sc_trace_file_base.cpp:268 due to divide by zero error. GNU gdb (Ubuntu 8.1-0ubuntu3) 8.1.0.20180409-git Copyright (C) 2018 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-linux-gnu". Type "show configuration" for configuration details. For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>. Find the GDB manual and other documentation resources online at: <http://www.gnu.org/software/gdb/documentation/>. For help, type "help". Type "apropos word" to search for commands related to "word"... Reading symbols from /home/ameya/Documents/demo/SystemC_ramblings/src/02_adder/build/tests/AdderTest/AdderTest.run...done. (gdb) r Starting program: /home/ameya/Documents/demo/SystemC_ramblings/src/02_adder/build/tests/AdderTest/AdderTest.run [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". SystemC 2.3.2-Accellera --- May 1 2018 16:30:17 Copyright (c) 1996-2017 by all Contributors, ALL RIGHTS RESERVED At time 0 s :: (a, b, cin): 100 (sum, carry_out): 00 At time 0 s :: (a, b, cin): 010 (sum, carry_out): 00 At time 0 s :: (a, b, cin): 110 (sum, carry_out): 10 At time 0 s :: (a, b, cin): 001 (sum, carry_out): 10 At time 0 s :: (a, b, cin): 101 (sum, carry_out): 11 At time 0 s :: (a, b, cin): 011 (sum, carry_out): 10 At time 0 s :: (a, b, cin): 111 (sum, carry_out): 00 At time 0 s :: (a, b, cin): 111 (sum, carry_out): 01 Info: /OSCI/SystemC: Simulation stopped by user. Program received signal SIGFPE, Arithmetic exception. 0x00007ffff7b57799 in sc_core::sc_trace_file_base::timestamp_in_trace_units (this=this@entry=0x5555557a3730, high=@0x7fffffffc658: 93824994651128, low=@0x7fffffffc660: 140737488346560) at /home/ameya/apps/src/systemc/systemc-2.3.2/src/sysc/tracing/sc_trace_file_base.cpp:268 268 unit_type unit_divisor = trace_unit_fs / kernel_unit_fs; (gdb) bt #0 0x00007ffff7b57799 in sc_core::sc_trace_file_base::timestamp_in_trace_units (this=this@entry=0x5555557a3730, high=@0x7fffffffc658: 93824994651128, low=@0x7fffffffc660: 140737488346560) at /home/ameya/apps/src/systemc/systemc-2.3.2/src/sysc/tracing/sc_trace_file_base.cpp:268 #1 0x00007ffff7b5d8a2 in sc_core::vcd_trace_file::get_time_stamp (this=this@entry=0x5555557a3730, now_units_high=@0x7fffffffc658: 93824994651128, now_units_low=@0x7fffffffc660: 140737488346560) at /home/ameya/apps/src/systemc/systemc-2.3.2/src/sysc/tracing/sc_vcd_trace.cpp:2086 #2 0x00007ffff7b5e242 in sc_core::vcd_trace_file::~vcd_trace_file (this=0x5555557a3730, __in_chrg=<optimized out>) at /home/ameya/apps/src/systemc/systemc-2.3.2/src/sysc/tracing/sc_vcd_trace.cpp:2145 #3 0x00007ffff7b5e319 in sc_core::vcd_trace_file::~vcd_trace_file (this=0x5555557a3730, __in_chrg=<optimized out>) at /home/ameya/apps/src/systemc/systemc-2.3.2/src/sysc/tracing/sc_vcd_trace.cpp:2153 #4 0x000055555555cfc1 in sc_main (argc=1, argv=0x5555557a2bf0) at /home/ameya/Documents/demo/SystemC_ramblings/src/02_adder/tests/AdderTest/src/full_adder_main.cpp:37 #5 0x00007ffff7b32e24 in sc_core::sc_elab_and_sim (argc=1, argv=<optimized out>) at /home/ameya/apps/src/systemc/systemc-2.3.2/src/sysc/kernel/sc_main_main.cpp:87 #6 0x00007ffff7097b97 in __libc_start_main (main=0x7ffff7aa1fa0 <main(int, char**)>, argc=1, argv=0x7fffffffddb8, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffffffdda8) at ../csu/libc-start.c:310 #7 0x000055555555c08a in _start () (gdb) I seems to happen only when the simulation time is not progressing. One can apply the below patch to branch (dev) available under here for the following sub-directory: https://github.com/AmeyaVS/SystemC_ramblings/tree/dev/src/02_adder diff --git a/src/02_adder/tests/AdderTest/src/driver.cpp b/src/02_adder/tests/AdderTest/src/driver.cpp index e70aea1..26169b7 100644 --- a/src/02_adder/tests/AdderTest/src/driver.cpp +++ b/src/02_adder/tests/AdderTest/src/driver.cpp @@ -7,8 +7,14 @@ void driver::prc_driver() { d_a = pattern[0]; d_b = pattern[1]; d_cin = pattern[2]; - wait(5, sc_core::SC_NS); + //wait(5, sc_core::SC_NS); + wait(sc_core::SC_ZERO_TIME); pattern++; + if (pattern == 0) { + break; + } } + wait(sc_core::SC_ZERO_TIME); + sc_core::sc_stop(); } This behavior is not reproduced in SystemC release 2.3.1a. I will try to debug further and hopefully with probable fix. Regards, Ameya Vikram Singh
  20. Hello @Hook, The \ character is an escape sequence in C/C++. Here is a reference https://en.m.wikipedia.org/wiki/Escape_sequences_in_C I would recommend that you brush up your skills on C++ before continuing further. I would recommend the book "The C++ Programming Language Book by Bjarne Stroustrup". Hope it helps. Regards, Ameya Vikram Singh
  21. Hello @SystemCInDepth, The wait() call can only be used in SC_THREAD/SC_CTHREAD registered processes in SystemC library. You are trying to call wait even before the elaboration happens and you have started the simulation. It would be better if you could go through the book: SystemC from Ground Up by David Black to have a better understanding. Hope it helps. Regards, Ameya Vikram Singh
  22. Hello @Aaron0127, The entry point for creating the simulation models and starting simulation is supposed to be implemented within sc_main function. This was done to support the EDA tools vendors to support extensive debugging capability with their tools. In the available open source implementation the main function is provided by the SystemC library itself. Only the sc_main function implementation has to be provided by you. Hope it helps. Regards, Ameya Vikram Singh
  23. Hello @wasim6691, Did you build the SystemC library in debug mode and installed it? On Windows using MSVC you cannot mix and match the Debug and Release Build libraries together. You would need to configure SystemC library installation in Debug build to Debug your models. Regards, Ameya Vikram Singh
  24. Hello @fdoucet, Sorry for the late response. But the "WAIT_UNTIIL()" macro has been refactored to be "SC_WAIT_UNTIL()". You can find the reference to the changes in the "RELEASENOTES" of the SystemC 2.3.2 release: RELEASENOTES:388: - The non-standard macros WAIT, WAITN and WAIT_UNTIL have been RELEASENOTES:389: renamed to SC_WAIT, SC_WAITN, SC_WAIT_UNTIL. Hope it helps. Regards, Ameya Vikram Singh
  25. Hello @ANKUR SAINI, If you are on Linux/UNIX with the Accellera release of SystemC library with the GNU toolchain, you can use the following tutorials for reference: https://sourceware.org/gdb/onlinedocs/gdb/Threads.html https://stackoverflow.com/questions/1695268/multi-threaded-debugging-tutorial-for-gdb-and-c Due to the co-operative multi-tasking implementation of the SystemC library available from Accellera as @Roman Popov has mentioned, much of the debugging tasks are easier since all threads synchronize with the main library thread. Let me know if it helps. Regards, Ameya Vikram Singh
×
×
  • Create New...