C++ Program either gets caught in infinite loop or get aborted!

[anubane]

I have a multi-file (3 .cpp files and 2 .h files) C++ code that uses new, new[], delete, and delete[] operators to allocate large chunks of memory and performs some floating point operations (requires division by sqrt()). The program compiles and runs well with regular g++ and produces the required output. I now want to convert it to RISC-V instructions.

1. I first used the GNU RISC-V compiler toolchain to compile the program with following command: riscv32-unknown-elf-g++ -w -march=rv32imafc -mabi=ilp32f -DPREALLOCATE=1 -mcmodel=medany -fno-common -static -Iinclude/ f1.cpp f2.cpp f3.cpp -o executable here executable is the compiled binary file.

2. Since no errors are thrown at compile time, I proceeded to run the whisper simulator as follows: whisper --isa imafc --target executable --logfile mylog --profileinst prfl where mylog and prfl are output file paths.

3. The log file generated exceeds 6 GB in size and it doesn't terminate, so I assume that it is stuck in some infinite loop.

4. I tried adding the _start() function and tohost variable which didn't work either (same problem as 3 above).

5. I generated an object dump for executable and captured the (hex) addresses of starting address of main(), its ending address and address of tohost and passed them as parameters to --startpc, --endpc and --tohost respectively, in which case, the program was aborted after 64 illegal operations and ~300 MB of log file.

I think I am making some mistake. Kindly help me fix this issue or share the correct steps to generate the RISC-V assembly instructions for my C++ program.

[jrahmeh]

Hi Anurag,

The commands you used look good. Maybe we have a bug in whisper. Would it be possible to share the source files so that we can try to reproduce the problem and fix it.

Thanks,

Joe

[anubane]

cppcode.zip

Please find the code attached in the .zip file. (Please modify the makefile in case you have to use it.)

The code, when compiled and run with regular g++, will currently not produce any output as I have commented out all the print statements. However it runs fine.

One thing to note is that I have used inline functions - could that have caused the issue?

[jrahmeh]

Hi Anurag,

The program has a bug in the normalize function: float* hs = new float[rows]; for (int i=0; i<cols; ++i) { hs[i] = 0; //++memwrite; //++l_memwrite; } The cols in the loop should be replaced by rows. The bug causes memory corruption. Once bug is fixed, code runs fine in whisper.

Joe

[jrahmeh]

Hi Anurag,

The program has a bug that corrupts memory. The bug is in the normalize function at line 161 of symm_normalize.h. The code should be: for (int i=0; i<rows; ++i) {

After fixing the bug, the code ran fine on whisper.

I found the bug by compiling the code for x86 and running the valgrind tool on the executable: valgrind bin/executable

Joe

On Tue, Dec 1, 2020 at 12:24 AM Anurag Banerjee notifications@github.com wrote:

cppcode.zip https://github.com/chipsalliance/SweRV-ISS/files/5620613/cppcode.zip

Please find the code attached in the .zip file. (Please modify the makefil in case you have to use it.)

The code, when compiled and run with regular g++, will currently not produce any output as I have commented out all the print statements.

One thing to note is that I have used inline functions - could that have caused the issue?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/chipsalliance/SweRV-ISS/issues/3#issuecomment-736251729, or unsubscribe https://github.com/notifications/unsubscribe-auth/AASHQX2PZT27SHS5IWZ2PQDSSSD23ANCNFSM4UHRKGFA .

[anubane]

Hello Dr. Rahmeh,

Thank you for helping me out with this - I feel silly for having made such an obvious out-of-bound-access error. Further, I also made an error at symm_normalize.h:112 - it should have been D[i] instead of D[j] - but that is more of a logical error.

However, for anyone reading this in future - please also make sure that you have the latest whisper simulator (this repo) and not the older one (westerndigital). In case you are using -march=rv32imafc during compile time and --isa imafc during running whisper, make sure your data is in float datatype - I used sqrt() (returns double) which caused some error for me, but, when I used sqrtf() all was fine. My final log file was about 725 MB in size.

Also, for me, using --newlib was essential while running whisper.

[jrahmeh]

HI Anurag,

No worries. Thanks for using whisper. I will remove the old repository. I will also add notes about double/float issues.

Thanks,

Joe

On Wed, Dec 2, 2020 at 1:41 PM Anurag Banerjee notifications@github.com wrote:

Hello Dr. Rahmeh,

Thank you for helping me out with this - I feel silly for having made such an obvious out-of-bound-access error. Further, I also made an error at symm_normalize.h:112 - it should have been D[i] instead of D[j] - but that is more of a logical error.

However, for anyone reading this in future - please also make sure that you have the latest whisper simulator (this repo) and not the older one (westerndigital). In case you are using -march=rv32imafc during compile time and --isa imafc during running whisper, make sure your data is in float datatype - I used sqrt() (returns double) which caused some error for me, but, when I used sqrtf() all was fine. My final log file was about 725 MB in size.

Also, for me, using --newlib was essential while running whisper.

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/chipsalliance/SweRV-ISS/issues/3#issuecomment-737452564, or unsubscribe https://github.com/notifications/unsubscribe-auth/AASHQXZIWYL2K6V3LK25IH3SS2J7XANCNFSM4UHRKGFA .