search

darklife / darkriscv

opensouce RISC-V cpu core implemented in Verilog from scratch in one night!
BSD 3-Clause "New" or "Revised" License
2.1k stars 282 forks source link

Running CoreMark benchmark in DarkRISCV, And Change 'src' directory structure #51

Closed vgegok closed 2 years ago

vgegok commented 2 years ago

DarkRISCV Application (./src)

samsoniuk commented 2 years ago

congratulations! now we have a coremark benchmark! :)

vgegok commented 2 years ago

I have verified my changes and ensured that the original make command works. But, I don't see any updates to the master branch.

samsoniuk commented 2 years ago

that is strange:

samsoniuk approved these changes 35 minutes ago View changes

samsoniuk commented 2 years ago

ohh, it was approved, but not merged... the merge button was hidden when I closed the issue

vgegok commented 2 years ago

OK, great! Could you make me to be one of the contributors?

samsoniuk commented 2 years ago

I am not really much literate about git... I am feeling like I am missing something important!

samsoniuk commented 2 years ago

It is strange that you does not appears in the contributors list... maybe I triggered a bug in the github when I closed the pull request before the merge?!

vgegok commented 2 years ago

I don't know much about GitHub either. Last time, after my submission was merged, I did not appear in the contributor list.

samsoniuk commented 2 years ago

hmmm maybe there is something missing in your account:

https://docs.github.com/en/account-and-profile/setting-up-and-managing-your-github-profile/managing-contribution-graphs-on-your-profile/why-are-my-contributions-not-showing-up-on-my-profile

vgegok commented 2 years ago

I used my abandoned account to submit a pull request in the same way. It worked: https://github.com/vgego/github-test.git

I'm using another computer now. I'll let you know when I check it tomorrow.

Thank you!

vgegok commented 2 years ago

I found that my previous useremail configuration was wrong. I submitted a new pull request to update the list of contributors.

samsoniuk commented 2 years ago

I found that the smaller coremark/MHz 0.7 value was because the -O1 flag in the GCC, but with -O2 you get a coremark/MHz 0.9 which match with my Kintex-7 at 240MHz... I am not an expert in the coremark benchmark, but I guess it means that the code is peaking 0.91 clocks/instruction, which means that the DarkRISCV is almost as powerful as a 486SX!

vgegok commented 2 years ago

Yes, I have explained the coremark results of different optimization levels in src/coremark/README.md.