NVIDIA / NVIDIA/cutlass

SM 120 (Blackwell GeForce RTX 50 Series) Block-Scaled MMA Runtime Assertion Failure

Open
#2,820 5 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

? - Needs Triage bug CUTLASS C++ inactive-30d inactive-90d
Dominant language
C++
Stars
10.5k
Forks
2.1k
Avg merge
3d 11h
Merged PRs (30d)
7

Description

Which component has the problem?

CUTLASS C++

Bug Report
Which component has the problem?

CUTLASS C++

Bug Report

Environment

  • GPU: NVIDIA GeForce RTX 5070 Ti (SM 120, Compute Capability 12.0)
  • Driver: 581.80 (Windows)
  • CUDA Toolkit: 13.0.88
  • OS: WSL Ubuntu 22.04 on Windows 11
  • CUTLASS: main branch (latest as of Nov 2025)

Issue

All Blackwell GeForce examples (79a-79d) fail at runtime with arch conditional MMA assertion:
Assertion 0 && "ERROR : Arch conditional MMA instruction used without targeting appropriate compute capability. Aborting.\n" failed.
../../include/cutlass/gemm/kernel/sm90_gemm_tma_warpspecialized_cooperative.hpp:366

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the failure with Blackwell examples 79a–79d on the stated RTX 5070 Ti environment. Read the assertion at include/cutlass/gemm/kernel/sm90_gemm_tma_warpspecialized_cooperative.hpp:366 and inspect how the examples target the compute capability. Done means the affected examples run without the arch-conditional MMA assertion.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
hpc
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.