search

harvard-acc / gem5-aladdin

End-to-end SoC simulation: integrating the gem5 system simulator with the Aladdin accelerator simulator.
BSD 3-Clause "New" or "Revised" License
210 stars 59 forks source link

SHOC/fft dynamic trace issue #13

Open gaurav-1993singh opened 4 years ago

gaurav-1993singh commented 4 years ago

Hi, I was trying to run the fft test in SHOC. In the normal mode (without dma) it makes the dynamic trace but it does not give out the power and area summary (it does give out the stats). When I try to make the trace using the dma-trace-binary, the fft-instrumented has a segmentation fault. Please let me know if there is a workaround. My main goal is to get power and area numbers.

Thanks

xyzsam commented 4 years ago

Thanks, I can reproduce. Investigating.

gaurav-1993singh commented 4 years ago

I will tell you my findings to hopefully save some of your time. The dmaLoad and dmaStore functions in the *.c files (testcases) are still on DMA_INTERFACE_V1, whereas the Makefile.tracer expect it to be on V3. Also, these files don't have the maparraytoaccelerator and invokeaccelerator commands. I think the makefiles should also automatically include the aladdin_sys_connection.c and aladdin_sys_constants.c files. I was able to get the power numbers for FFT by making all these hacks. In all, I think the .c files for test cases need update.

xyzsam commented 4 years ago

Yes, the discrepancy in expected DMA interface versions is the reason. We deprecated v1 but we missed updating most of these benchmarks (only the ones in our integration tests were fixed). Glad you were able to get the numbers you wanted. We'll get this issue fixed.