Suriya’s Design Notebook

Week: Sep.10 - Sep.16 /2023

work:

comments:

  • already had the WSL on my laptop so installing the required softwares wasn’t a hassle.
  • had the github for long but never worked on team, so looking forward to it.
  • cmake was new for me, did take some time to install and get to know the stuff.
  • the answers for the exercise are in Readme.md files of the respective labs
  • had to look up LFSR, did some problems on it previously. Everything else is straight Forward brushing up basics.
  • completed the verification and encounterd a problem in exercise 2, rectified it.
  • starting on verification.

Week: Sep.17 - Sep.23/2023

Work:

  • worked on lab3, not attaching the link yet

comments:

  • completed exercise 1,2 working on 3 and 4
  • in next week will complete the lab3 and maybe 4 too
  • researching on teams

Week: Sep.24 - Oct.1/2023

Work:

  • Completed lab3 lab3

Comments:

  • brushed upon C++; working on lab4
  • researching on which to join after the onboarding labs.
  • encountered a build issue, cleared it

Week: Oct.2 - Oct.8/2023

Work:

comments:

  • got to use some macros from catch2 dependecny
  • got hang of how all the packages work in the onboarding labs that we did
  • doing somemore work on them to get good hang of it.
  • planning on joining core team
  • don’t know the process but will text in the group.

Week: Oct.9 - Oct.15/2023

comments:

  • completed the onboardingLabs
  • revised the RV32 from ComputerOrganization&Design textbook.
  • worked on some modules and tried to implement a singlecycledatapath version of rv32
  • completed the design of the moudles, excpet topModuleWrapper and testbenches
  • will contact the core czar and start work on the assigned work

Week: Oct.16 - Oct.23/2023

Work:

  • got assigned to cache design

comments:

  • started working on it, after L1, will continue with others,
  • had the midterms couldn’t work on it good,
  • will complete this week

Weeks: Oct.23 - Nov.5/2023

Work:

  • completed the L1_data_cache documentation.
  • implemented the L1_data_cache module upto write and read hits
  • misses should be implemeted

comments:

  • no specifications for tha cache were there, so chose typical
  • write back, with write allocate and write through cache
  • LRU replacement policy, only implemented in L1, brainstorming on how to do the L2
  • opened a PR, and micheal gave some suggestions,
  • going to work on them and inorder to test the module i have complete datapath
  • haven’t worked with heirarchy caches so looking for resources

Weeks: Nov.6 - Nov.18/2023

Work:

  • Coded the typical version of a L1 cache,
  • tested the l1 module with a L2 simulation of different parameters

Comments:

  • should look for better way to test this module
  • and also this was implemented using registers directly
  • for simulating representing this in regsiters is good enough but SRAMs should be preferred
  • once testing is done will implement the SRAM version

Weeks: Nov.19 - Dec.2/2023

Work:

  • testing is successful for L1, tested for 2 types of various sets and LRU counter
  • need to implement ECC and improve the block size to reduce the miss rate

Comments:

  • tested with the same L2 simulation module, need to improve that. looking for various options
  • implemented the SRAM version, testing will be completed in 3 days, after that i will PR it
  • and look for suggestions in the similar way i am coding L2 with 4 way
  • havent tested for miss rate and penatly, need to check them and use optimization techniques as per the req.

Weeks: Dec.3 - Dec.17/2023

Work:

  • completed the testing for L1 Module, and started on L2
  • updated the documentation for overview and structural overview
  • made some changes to the module and tb for PR

comments:

  • as i was not comfortable in c++, i used sv for testing and after the testing is completed i started to convert it into c++, but i confirmed with michael and he’s fine with any testing as long its good.
  • Created Pull requests for all the work today and messaged micheal about this.
  • have 3 exams back to back so completely packed that why creating PR today.