| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| Ripes-v2.3.0-linux-x86_64.AppImage | 2026-10-19 | 34.3 MB | |
| Ripes-v2.3.0-mac-universal2.zip | 2026-10-19 | 68.6 MB | |
| Ripes-v2.3.0-win-x86_64.zip | 2026-10-19 | 25.4 MB | |
| README.md | 2026-10-19 | 3.5 kB | |
| Release build (v.2.3.0) source code.tar.gz | 2026-10-19 | 14.6 MB | |
| Release build (v.2.3.0) source code.zip | 2026-10-19 | 14.8 MB | |
| Totals: 6 Items | 157.7 MB | 0 | |
New features
- Qt 6: Ripes has been migrated to Qt 6, and the macOS release is now a universal (Apple Silicon + Intel) build.
- Ripes in the browser: a WebAssembly build of Ripes is available at ripes.dk.
- New processor models:
- A 5-stage multicycle processor (with one or two memories), based on the work of Francisco Valero Gil.
- An ISA simulator "processor" model (
RV32_ISS/RV64_ISS) for fast, non-pipelined execution. - Theming: Light/Dark/System color schemes with live theme switching.
- New memory-mapped I/O peripherals: a raster screen, a real-time clock, and an edge-capture/latching D-pad.
- Command line mode:
- ELF and C files can now be executed from the CLI.
- System input (stdin) handling for CLI execution.
- Cache simulation is now available in the CLI (
--cache-preset). - Assembler:
- GNU-style
offset(base)memory operands (e.g.lw t0, 4(sp),lw t0, (sp),jalr ra, 4(t0),c.lw a0, 4(a1),c.lwsp a0, 4(sp)). - Labels may now start with a dot.
- Caches: non-power-of-two associativities are now supported, and cache configuration now uses the term "sets" instead of "lines".
- UI:
- Added a memory viewer widget to the edit and processor tabs; register values can be copied from the register widget and followed into memory (right click).
- Panning in the cache graphic view, a modernized sidebar, and persisted window geometry.
- Added a setting for enabling VSRTL VCD tracing.
Bug fixes
- Processor models:
jalrdid not clear the least significant bit of the target address (#477).- Register shift instructions (
sll/srl/sra) used all bits ofrs2instead of only the lower 5 (RV32) or 6 (RV64) bits (#398). - Incorrect
mulh/mulhu/mulhsuresults on 64-bit processors. - Compressed
c.j/c.jalinstructions jumping to the wrong address for offsets below -1024 (#472). - Incorrect EX stage stall in the 5-stage pipeline (#444).
- Fixed RVC prime register encoding/decoding.
- Assembler:
c.j/c.jal/c.beqz/c.bnezencoded label targets as absolute rather than PC-relative addresses (#411).sw rs2, SYMBOL(rs1)with a symbolic offset was assembled as thesw rs2, label, rtpseudo-instruction (#467).- Superfluous instruction operands are now rejected rather than silently ignored.
- Disabled ISA extensions of a processor were still accepted by the assembler.
- Hex/binary representations of edge-case signed immediates (e.g.
0xFFFFFFFF) are now accepted. - Division by zero in constant expressions crashed the assembler (#347).
- Parentheses were matched before comments were removed (#421).
- Caches: fixed the random replacement policy, ensured a minimum of one way, and fixed numeric limits for way counts and presets.
- System calls: fixed the
readsyscall character count being halved, theO_CREATflag inopen, read-only file operations, and added a stack-heap collision check tobrk. - UI and misc:
- Auto-clocking was not stopped when a breakpoint was hit.
- A breakpoint at line 0 was also mirrored at line
num instructions. - I/O devices are now updated upon processor reset.
- Fixed a crash when hovering over the cache statistics chart.
- Fixed memory leaks in I/O simulation.
- Console input and clipboard usage in the WebAssembly build (#304).
- Source code highlighting with programs loaded at a non-zero start address.