From 945f55ac01db9973194355d08f0172edd9c4d02b Mon Sep 17 00:00:00 2001 From: Ibrahim Date: Sun, 8 Mar 2026 19:35:21 -0400 Subject: [PATCH 1/2] docs(dn): Ibrahim Hashim 03/08/26 --- src/design_notebooks/2026spring/irh8156.md | 47 +++++++++++++++++++++- 1 file changed, 46 insertions(+), 1 deletion(-) diff --git a/src/design_notebooks/2026spring/irh8156.md b/src/design_notebooks/2026spring/irh8156.md index f379b2b1..dc60ed9d 100644 --- a/src/design_notebooks/2026spring/irh8156.md +++ b/src/design_notebooks/2026spring/irh8156.md @@ -7,4 +7,49 @@ Project Work: * Went over onboarding labs 1-4. Comments: - * I created my first design notebook and submitted my first pull request of the semester. \ No newline at end of file + * I created my first design notebook and submitted my first pull request of the semester. + + +## Week of 1 Feb 2026 + +Project Work: + * Made sure I knew and rememebered everything from onboarding labs 1-4. + + +## Week of 8 Feb 2026 + +Project Work: +* Had no assigned work for this week. + + +## Week of 15 Feb 2026 + +Project Work: +* Had a team meeting this week, +* Went over the NYU Processor Design Team SoC CPU Core report + +Notes on the report: +* The overall design of the CPU is a 32 bit CPU using the RISC-V 32 instruction set +* It has a regular 5 stage pipeline like we took in comparch +* It uses a 2 bit branch predictor +* Its data cache is 2 way associative +* A question I have is if a different type of branch predictor would improve the performance of the CPU, specificially something more advanced, and what tradeoffs implementing something like that would have +* Why don't we use data forwarding when the hazard detection stalls the pipeline using NOPs? + + +## Week of 22 Feb 2026 + +Project Work: +* Went over the Design and Construction of RTL Toolchains at NYU report + +Notes on the report: +* Cmake is the main tool used for building and similation +* vcpkg handles dependencies and catch2 gives us the testing framework +* A question I have: Are there things that are easier/more accurate to test in HDL that become harder when everything's in C++? +* If we're depending to heavily on verilator, isn't it risky to put all our hope in this 1 tool since it could have bugs or limitations? + + +## Week of 1 Mar 2026 + +Project Work: +* Chose a subteam for the main project (core) From 40a887eb31cde164c1a68bd4b12bccdbfd5bbeb3 Mon Sep 17 00:00:00 2001 From: Ibrahim Date: Sun, 15 Mar 2026 07:21:57 -0400 Subject: [PATCH 2/2] docs(dn): Ibrahim Hashim 3/15/2026 --- src/design_notebooks/2026spring/irh8156.md | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/src/design_notebooks/2026spring/irh8156.md b/src/design_notebooks/2026spring/irh8156.md index dc60ed9d..3d035194 100644 --- a/src/design_notebooks/2026spring/irh8156.md +++ b/src/design_notebooks/2026spring/irh8156.md @@ -53,3 +53,15 @@ Notes on the report: Project Work: * Chose a subteam for the main project (core) + + +## Week of 8 Mar 2026 + +Project Work: +* No work assigned. Watched some videos about Von Neumann Architecture. + + +## Week of 15 Mar 2026 + +Project Work: +* Spring break! :) \ No newline at end of file