Issues

• acc_latency = 4 drops accumulator writes ("Execute controller write to AccumulatorMem was skipped")

  ucb-bar/gemmini#424 · Filter by author youzi27 opened on Nov 1, 2026

• tiled_conv with a split kernel can pad a window more than its size (LoopConv asserts or hangs)

  ucb-bar/gemmini#423 · Filter by author youzi27 opened on Oct 31, 2026

• chipConfig fails to elaborate: None.get in AccumulatorMem

  ucb-bar/gemmini#421 · Filter by author youzi27 opened on Oct 31, 2026

• Wrong conv outputs with pixel repeats: the RS doesn't track the last rows of the mvin

  ucb-bar/gemmini#419 · Filter by author youzi27 opened on Oct 31, 2026

• [Security] Gemmini DMA may expose protected memory despite translation or access faults

  ucb-bar/gemmini#418 · Filter by author 90ICEDA opened on Oct 28, 2026

• LoopConv downsample with padding generates a zero-row mvin and stalls the load pipeline

  ucb-bar/gemmini#417 · Filter by author 90ICEDA opened on Oct 27, 2026

• Zero mvin into the accumulator with shrink set overruns the load DMACommandTracker byte count

  ucb-bar/gemmini#413 · Filter by author youzi27 opened on Oct 26, 2026

• Output-stationary with a_transpose = 1 and b_transpose = 0 adds a stale term when the execute queue runs empty during an accumulation

  ucb-bar/gemmini#412 · Filter by author youzi27 opened on Oct 26, 2026

• config_ex between a preload and its compute gives a wrong result for that matmul

  ucb-bar/gemmini#411 · Filter by author youzi27 opened on Oct 26, 2026

• Output-stationary matmul with C stride != 1 writes C rows to wrong addresses

  ucb-bar/gemmini#410 · Filter by author youzi27 opened on Oct 26, 2026

• mvins that write the same scratchpad/accumulator rows can complete out of order

  ucb-bar/gemmini#409 · Filter by author youzi27 opened on Oct 25, 2026

• Multi-row mvout reads incorrect scratchpad rows and corrupts output data

  ucb-bar/gemmini#408 · Filter by author 90ICEDA opened on Oct 18, 2026

• Request for a Private Security Vulnerability Reporting Channel

  ucb-bar/gemmini#407 · Filter by author 90ICEDA opened on Oct 18, 2026

• Output-stationary matmul with local C stride > 1 writes its result to the wrong scratchpad rows and overwrites other rows

  ucb-bar/gemmini#406 · Filter by author vera-rubin opened on Oct 11, 2026

• Timing violation in default Gemmini configuration at 3 ns clock period

  ucb-bar/gemmini#405 · Filter by author mtalhartl opened on Oct 7, 2026

• Install Chipyard issue

  ucb-bar/gemmini#402 · Filter by author hiennd49 opened on May 8, 2026
  2

• Possible FP32-to-BF16 truncation in BF16 mixed-precision PE OS dataflow

  ucb-bar/gemmini#401 · Filter by author Yanze-Xu opened on May 6, 2026

• quest: project maintainance

  ucb-bar/gemmini#400 · Filter by author tampler opened on Apr 19, 2026

• LoadMatmulLdB Loads in too many rows

  ucb-bar/gemmini#399 · Filter by author SeanStothers opened on Mar 4, 2026

• Gemmini Performance Scaling: 16x16 vs 64x64

  ucb-bar/gemmini#397 · Filter by author jhkim-voc opened on Oct 29, 2025
  1

• Gemmini + Rocket on Arty Z7-20: LUT over-utilization; trying RV32 + small core config

  ucb-bar/gemmini#396 · Filter by author Nauqchix opened on Oct 27, 2025

• [Chipyard/Gemmini] Simulation Fails with "ReservationStation.scala" Assertion for Softmax/Layernorm Workloads

  ucb-bar/gemmini#395 · Filter by author xsw632 opened on Sep 26, 2025
  1

• Question: Improving Fmax for Gemmini synthesis

  ucb-bar/gemmini#394 · Filter by author hojoojie opened on Sep 12, 2025

• LUTs consumption difference between OS and WS

  ucb-bar/gemmini#390 · Filter by author Ben-INL opened on May 16, 2025
  1

• Remove AXI4 from Gemmini

  ucb-bar/gemmini#389 · Filter by author Ben-INL opened on May 12, 2025
  1