-
Notifications
You must be signed in to change notification settings - Fork 104
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Unable to run Benchmarks #141
Comments
Hi, I havent used embench in a while and this lief import is more recent change but I would expect you just need to run something like For benchmarking speed, you want to take a look at the scripts in https://github.com/embench/embench-iot/tree/master/pylib. For example, to run the benchmarks on your native machine, I believe the required option is |
Thanks, installing lief with pip seemed to fix the size testing. For speed however, when I set it to use the module "run-native", I get the following results: Benchmark Speed Looking into the log, I can see that for every single benchmark, the timing seems to be failing, like so: **Warning: Failed to find timing ./aha-mont64: ./aha-mont64: cannot execute binary file Warning: Failed to find timing ./crc32: ./crc32: cannot execute binary file This goes on in the log for the remaining 17 benchmarks. Not sure exactly what the issue with the timing is. |
the problem there is that to run on your native machine you also need to compile the benchmarks for that architecture (probably x86 and not rv32). Running natively is the easier bit and should work right away, running on some other target may be more complicated. Possibly https://github.com/embench/embench-iot/blob/master/pylib/run_gdbserver_sim.py could help you here, but I havent used that yet. |
I gave run_gdbserver_sim a shot and the times are still coming up empty unfortunately. Taking a peek at the logs shows this: Warning: Failed to find return code No symbol table is loaded. Use the "file" command. This series of messages just repeats 18 more times for the rest of the benchmarks in the log. It says that the monitor command is not supported by the target, which I take to mean that riscv32 won't work for this module either (please correct me if I'm wrong)? If so, is there a workaround for RISC-V processor testing for embench (i.e. how are others going about it)? |
Hi, I'm having the same issue while trying to run the speed benchmark. |
Hi rawan305, Unfortunately I have not been able to run it as of yet. |
Hi, Thanks in Advance |
I have compiled all of the Benchmark source files, but when I try to run the size and speed tests, I come across some issues. When running ./benchmark_size.py, I get the following:
Traceback (most recent call last):
File "./benchmark_size.py", line 26, in
import lief
ModuleNotFoundError: No module named 'lief'
When running ./benchmark_speed.py, I get:
benchmark_speed.py: error: the following arguments are required: --target-module
For benchmark_speed, what module should I be targeting exactly? For the record I built the benchmarks with the following supplied arguments:
./build_all.py --arch risc32 --board ri5cyverilator --cc riscv64-unknown-elf-gcc --clfags=-c --user-libs="-lm"
The text was updated successfully, but these errors were encountered: