CI run 4f4d66b¶

Run: link
Time: Denver 2025-08-26 07:14:18 MDT • Brussels 2025-08-26 15:14:18 CEST

GEMM Deployment Summary¶

Workflow: AIE Deployment Gemm
Commit: 4f4d66be1e8d7e5f268946d9bc36949d51904e93
Runner: venus
Run time: Denver 2025-08-26 07:13:58 MDT • Brussels 2025-08-26 15:13:58 CEST
Run: #54; Attempt 1

HW	M	K	N	Status	Note
single_col	128	128	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	128	128	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	128	128	64	✅ success
single_col	128	256	128	✅ success
single_col	128	256	256	❌ failed	Failed to open KMQ device due to invalid argument.
single_col	128	256	64	✅ success
single_col	128	64	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	128	64	256	❌ failed	Unexpected command state in qds_device::wait().
single_col	128	64	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	256	128	128	✅ success
single_col	256	128	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	256	128	64	✅ success
single_col	256	256	128	✅ success
single_col	256	256	256	❌ failed	Unexpected command state in qds_device::wait().
single_col	256	256	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	256	64	128	✅ success
single_col	256	64	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	256	64	64	✅ success
single_col	64	128	128	✅ success
single_col	64	128	256	✅ success
single_col	64	128	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	64	256	128	✅ success
single_col	64	256	256	✅ success
single_col	64	256	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_col	64	64	128	✅ success
single_col	64	64	256	✅ success
single_col	64	64	64	✅ success
single_core	128	128	128	✅ success
single_core	128	128	256	✅ success
single_core	128	128	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	128	256	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	128	256	256	✅ success
single_core	128	256	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	128	64	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	128	64	256	✅ success
single_core	128	64	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	256	128	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	256	128	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	256	128	64	✅ success
single_core	256	256	128	✅ success
single_core	256	256	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	256	256	64	✅ success
single_core	256	64	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	256	64	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	256	64	64	✅ success
single_core	64	128	128	✅ success
single_core	64	128	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	64	128	64	✅ success
single_core	64	256	128	✅ success
single_core	64	256	256	✅ success
single_core	64	256	64	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	64	64	128	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	64	64	256	❌ failed	Failed to open KMQ device: Invalid argument.
single_core	64	64	64	✅ success
whole_array	128	128	128	❌ failed	Undefined symbols during linking process.
whole_array	128	128	256	❌ failed	Undefined symbols during linking process.
whole_array	128	128	64	❌ failed	Undefined symbols during linking process.
whole_array	128	256	128	❌ failed	Undefined symbols during linking process.
whole_array	128	256	256	❌ failed	Undefined symbols during linking process.
whole_array	128	256	64	❌ failed	Undefined symbols during linking process.
whole_array	128	64	128	❌ failed	Undefined symbols during linking process.
whole_array	128	64	256	❌ failed	Undefined symbols during linking process.
whole_array	128	64	64	❌ failed	Undefined symbols during linking process.
whole_array	256	128	128	❌ failed	Undefined symbols during linking process.
whole_array	256	128	256	❌ failed	Undefined symbols during linking process.
whole_array	256	128	64	❌ failed	Undefined symbols during linking process.
whole_array	256	256	128	❌ failed	Undefined symbols during linking caused the failure.
whole_array	256	256	256	❌ failed	Undefined symbols during linking process.
whole_array	256	256	64	❌ failed	Undefined symbols during linking process.
whole_array	256	64	128	❌ failed	Undefined symbols during linking process.
whole_array	256	64	256	❌ failed	Undefined symbols during linking process.
whole_array	256	64	64	❌ failed	Undefined symbols during linking process.
whole_array	64	128	128	❌ failed	Undefined symbols during linking process.
whole_array	64	128	256	❌ failed	Undefined symbols during linking process.
whole_array	64	128	64	❌ failed	Undefined symbols during linking process.
whole_array	64	256	128	❌ failed	Undefined symbols during linking process.
whole_array	64	256	256	❌ failed	Undefined symbols during linking process.
whole_array	64	256	64	❌ failed	Undefined symbols during linking process.
whole_array	64	64	128	✅ success
whole_array	64	64	256	✅ success
whole_array	64	64	64	❌ failed	Failed to open KMQ device: Invalid argument.

Totals: ✅ 30 • ❌ 51 • All: 81

[single_col] M=128 K=128 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 16 | 24,175 | 1,070.00 | 30.62 | 47.85 | 21.69 | 33.89 | | tile2,1 | 16 | 24,176 | 1,070.00 | 30.62 | 47.85 | 21.69 | 33.88 |

[single_col] M=128 K=256 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 64 | 113,291 | 1,058.00 | 30.97 | 48.39 | 18.51 | 28.92 | | tile2,1 | 64 | 113,292 | 1,058.00 | 30.97 | 48.39 | 18.51 | 28.92 |

[single_col] M=128 K=256 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 32 | 52,410 | 1,058.00 | 30.97 | 48.39 | 20.01 | 31.26 | | tile3,1 | 32 | 52,411 | 1,058.00 | 30.97 | 48.39 | 20.01 | 31.26 |

[single_col] M=256 K=128 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 64 | 105,638 | 1,070.00 | 30.62 | 47.85 | 19.85 | 31.02 | | tile3,1 | 64 | 105,639 | 1,070.00 | 30.62 | 47.85 | 19.85 | 31.02 |

[single_col] M=256 K=128 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 32 | 50,629 | 1,070.00 | 30.62 | 47.85 | 20.71 | 32.36 | | tile2,1 | 32 | 50,629 | 1,070.00 | 30.62 | 47.85 | 20.71 | 32.36 |

[single_col] M=256 K=256 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 126 | 218,313 | 1,065.36 | 30.76 | 48.06 | 19.21 | 30.02 | | tile3,1 | 126 | 218,369 | 1,065.80 | 30.74 | 48.04 | 19.21 | 30.01 |

[single_col] M=256 K=64 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 32 | 60,032 | 1,150.00 | 28.49 | 44.52 | 17.47 | 27.29 | | tile2,1 | 32 | 60,033 | 1,150.00 | 28.49 | 44.52 | 17.47 | 27.29 |

[single_col] M=256 K=64 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 16 | 27,233 | 1,150.00 | 28.49 | 44.52 | 19.25 | 30.08 | | tile2,1 | 16 | 27,233 | 1,150.00 | 28.49 | 44.52 | 19.25 | 30.08 |

[single_col] M=64 K=128 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 16 | 23,188 | 1,046.00 | 31.33 | 48.95 | 22.61 | 35.33 | | tile3,1 | 16 | 23,701 | 1,046.00 | 31.33 | 48.95 | 22.12 | 34.56 |

[single_col] M=64 K=128 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 32 | 61,700 | 1,046.00 | 31.33 | 48.95 | 16.99 | 26.55 | | tile3,1 | 32 | 62,213 | 1,046.00 | 31.33 | 48.95 | 16.85 | 26.34 |

[single_col] M=64 K=256 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 32 | 59,621 | 1,058.00 | 30.97 | 48.39 | 17.59 | 27.48 | | tile2,1 | 32 | 60,137 | 1,058.00 | 30.97 | 48.39 | 17.44 | 27.24 |

[single_col] M=64 K=256 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 64 | 138,152 | 1,058.00 | 30.97 | 48.39 | 15.18 | 23.72 | | tile2,1 | 64 | 140,257 | 1,597.66 | 20.51 | 32.05 | 14.95 | 23.36 |

[single_col] M=64 K=64 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 8 | 13,317 | 1,134.00 | 28.90 | 45.15 | 19.68 | 30.76 | | tile3,1 | 8 | 13,834 | 1,134.00 | 28.90 | 45.15 | 18.95 | 29.61 |

[single_col] M=64 K=64 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 16 | 41,447 | 1,134.00 | 28.90 | 45.15 | 12.65 | 19.77 | | tile3,1 | 16 | 42,465 | 1,134.00 | 28.90 | 45.15 | 12.35 | 19.29 |

[single_col] M=64 K=64 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile3,1 | 4 | 5,781 | 1,134.00 | 28.90 | 45.15 | 22.67 | 35.43 | | tile2,1 | 4 | 5,781 | 1,134.00 | 28.90 | 45.15 | 22.67 | 35.43 |

[single_core] M=128 K=128 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 64 | 103,699 | 1,070.00 | 30.62 | 47.85 | 20.22 | 31.60 |

[single_core] M=128 K=128 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 128 | 209,746 | 1,070.00 | 30.62 | 47.85 | 20.00 | 31.25 |

[single_core] M=128 K=256 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 256 | 449,912 | 1,058.00 | 30.97 | 48.39 | 18.64 | 29.13 |

[single_core] M=128 K=64 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 64 | 107,388 | 1,078.00 | 30.40 | 47.50 | 19.53 | 30.51 |

[single_core] M=256 K=128 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 64 | 103,582 | 1,070.00 | 30.62 | 47.85 | 20.25 | 31.63 |

[single_core] M=256 K=256 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 256 | 437,335 | 1,058.00 | 30.97 | 48.39 | 19.18 | 29.97 |

[single_core] M=256 K=256 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 128 | 216,505 | 1,058.00 | 30.97 | 48.39 | 19.37 | 30.27 |

[single_core] M=256 K=64 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 32 | 55,759 | 1,150.00 | 28.49 | 44.52 | 18.81 | 29.38 |

[single_core] M=64 K=128 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 32 | 45,585 | 1,046.00 | 31.33 | 48.95 | 23.00 | 35.94 |

[single_core] M=64 K=128 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 16 | 21,654 | 1,046.00 | 31.33 | 48.95 | 24.21 | 37.83 |

[single_core] M=64 K=256 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 64 | 115,339 | 1,058.00 | 30.97 | 48.39 | 18.18 | 28.41 |

[single_core] M=64 K=256 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 128 | 235,010 | 1,058.00 | 30.97 | 48.39 | 17.85 | 27.89 |

[single_core] M=64 K=64 N=64

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 8 | 12,808 | 1,134.00 | 28.90 | 45.15 | 20.47 | 31.98 |

[whole_array] M=64 K=64 N=128

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,1 | 4 | 5,874 | 1,150.00 | 28.49 | 44.52 | 22.31 | 34.87 | | tile3,1 | 4 | 5,874 | 1,150.00 | 28.49 | 44.52 | 22.31 | 34.87 | | tile2,2 | 4 | 5,875 | 1,150.00 | 28.49 | 44.52 | 22.31 | 34.86 | | tile3,2 | 4 | 5,876 | 1,150.00 | 28.49 | 44.52 | 22.31 | 34.85 |

[whole_array] M=64 K=64 N=256

| Tile | Kernels | Total cycles | Avg cycles per kernel | MACs/cycle (kernel) | Peak eff. kernel % | MACs/cycle (system) | Peak eff. system % | |------|---------|--------------|-----------------------|---------------------|--------------------|---------------------|--------------------| | tile2,2 | 8 | 12,991 | 1,150.00 | 28.49 | 44.52 | 20.18 | 31.53 | | tile2,1 | 8 | 12,991 | 1,150.00 | 28.49 | 44.52 | 20.18 | 31.53 | | tile3,2 | 8 | 12,992 | 1,150.00 | 28.49 | 44.52 | 20.18 | 31.53 | | tile3,1 | 8 | 12,992 | 1,150.00 | 28.49 | 44.52 | 20.18 | 31.53 |