Toooba
ToolToooba is an open-source Bluespec/Verilog RISC-V core described as superscalar, out-of-order, and multi-core capable. It has been instrumented for TestRIG-style RVFI-DII verification, including changes to support direct instruction injection, superscalar fetch, compressed instruction fragments, and RVFI reporting.
First seen 5/30/2026
Last seen 7/19/2026
Evidence 17 chunks
Wiki v1
WIKI
Overview
Toooba is an open-source RISC-V core from Bluespec. The public GitHub description characterizes it as a superscalar, out-of-order, multi-core-capable RISC-V core based on MIT's RISCY-OOO design. The repository is listed as Verilog and, in the provided public metadata, has 187 stars, 51 forks, and was updated on 2026-05-20.
Microarchitecture role
NEIGHBORHOOD
4 nodes · 9 edgesgraph · Toooba · depth=1
RELATIONSHIPS
11 connectionsTestRIG is used to evaluate the Toooba processor.
Toooba is a CHERI-enabled RISC-V processor.
Toooba is a RISC-V out-of-order processor derived from RiscyOO.
Toooba is a fork of RiscyOO and implements superscalar out-of-order execution.
The thesis evaluates the CHERI-extended Toooba superscalar processor.
Toooba is implemented in Bluespec HDL as a fork of RiscyOO.
Toooba is used as a platform to research and implement temporal safety with CHERI.
Toooba is a fork of the RiscyOO processor from MIT, extended with CHERI modifications.
Toooba as an application-class processor uses virtual memory via an MMU.
Toooba is evaluated in the context of safe speculation for CHERI processors.
Toooba depends on Verilator as part of its build requirements.
CITATIONS
6 sources6 citations — click to expand
[1] Toooba is a Bluespec open-source RISC-V core described as superscalar, out-of-order, multi-core capable, and based on MIT's RISCY-OOO. bluespec/Toooba
[2] The Toooba repository public metadata identifies the language as Verilog, with 187 stars, 51 forks, and an updated_at timestamp of 2026-05-20T07:06:37Z. bluespec/Toooba
[3] TestRIG extends RVFI with Direct Instruction Injection, where DII is used for instruction input and RVFI for trace output, enabling interactive verification. Randomized Testing of RISC-V CPUs using Direct
[4] For superscalar Toooba, RVFI-DII work first substituted the vector of picked instructions before decode while keeping instruction-cache access, then moved to bypassing the instruction cache and providing 16-bit instruction fragments to exercise instruction picking and decode. Randomized Testing of RISC-V CPUs using Direct
[5] The RVFI-DII approach adapted to Toooba added superscalar fetch and assigned IDs to compressed instruction fragments to keep DII/RVFI synchronization through pipeline effects. Randomized Testing of RISC-V CPUs using Direct
[6] Extending the superscalar Toooba core for RVFI-DII required two extra records per instruction in the reorder buffer, present only when built for simulation with RVFI and therefore not a physical overhead for the design. Randomized Testing of RISC-V CPUs using Direct