issues
search
opencompl
/
Quidditch
IREE compiler and runtime for Snitch
Apache License 2.0
6
stars
3
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Python 3.11 is a strict requirement for building
#134
compor
opened
1 month ago
0
Missing `setuptools`
#133
compor
opened
1 month ago
1
Add support for Reordering Instruction in Quidditch flow
#132
RRavikiran66
closed
2 months ago
0
[SnitchDMA] Split transfer legalization out of `DMAToLLVM`
#131
zero9178
closed
2 months ago
0
[DMA] Add `combine_token` op
#130
zero9178
closed
2 months ago
0
[DMA] Clarify concurrent behaviour of transfers and fix `start_tensor_copy`
#129
zero9178
closed
2 months ago
0
[SnitchDMA] Introduce `SnitchDMA` dialect
#128
zero9178
closed
2 months ago
0
[DMA] Split DMA operations into its own dialect
#127
zero9178
closed
2 months ago
0
[samples] Add `big_matvec` sample with pre-existing lowering config
#126
zero9178
closed
2 months ago
0
[TensorTile] Merge reduction tiling into L1 tiling
#125
zero9178
closed
2 months ago
0
expand README.md
#124
superlopuh
closed
2 months ago
0
[ConvertSnitchToLLVM] Fix bug in 1d size calculation
#123
zero9178
closed
3 months ago
0
[ConvertSnitchToLLVM] Improve DMA transfers contiguity check
#122
zero9178
closed
3 months ago
0
[quidditch_snitch] Remove redundant `start_tensor_copy` op in more cases
#121
zero9178
closed
3 months ago
0
[quidditch_snitch] Implement undef padding for `start_tensor_copy`
#120
zero9178
closed
3 months ago
0
[PromotePadsToL1] Implement pass lowering `tensor.pad`
#119
zero9178
closed
3 months ago
0
[Target] Implement pass inserting `tensor.pad` if required
#118
zero9178
closed
3 months ago
0
[PromoteToL1] Fix dominance bug
#117
zero9178
closed
3 months ago
0
[quidditch_snitch] Add padding capabilities to `start_tensor_copy`
#116
zero9178
closed
3 months ago
0
[quidditch_snitch] Simplify `specialize-dma-code` using interfaces
#115
zero9178
closed
3 months ago
0
Bump snitch_cluster past 0-size iDMA fix
#114
zero9178
closed
3 months ago
0
[quidditch_snitch] Add `start_zero_mem_transfer` operation
#113
zero9178
closed
3 months ago
0
[quidditch_snitch] Add `CompletedTokenAttr`
#112
zero9178
closed
3 months ago
0
[quidditch_snitch] Improve `wait_for_dma_transfers` without tokens syntax
#111
zero9178
closed
3 months ago
0
[ConvertToRISCV] Fallback to Linalg lowering if `xdsl-opt` fails
#110
zero9178
closed
3 months ago
0
[SpecializeDMACode] Properly lower `compute_core_index`
#109
zero9178
closed
3 months ago
0
[quidditch_snitch] Rename `cluster_index` to `compute_core_index`
#108
zero9178
closed
3 months ago
0
[samples] Properly check for correctness in `vec_multiply` tests
#107
zero9178
closed
3 months ago
0
[quidditch_snitch] Reintroduce `tensor.microkernel` op
#106
zero9178
closed
3 months ago
0
[samples] Speed up simulation by importing buffers
#105
zero9178
closed
3 months ago
0
[LowerL1Allocations] Support non-identity strides in L1 `memref.alloca`
#104
zero9178
closed
4 months ago
0
Bump IREE version
#103
zero9178
closed
4 months ago
0
[traces] Skip generators on stalls
#102
zero9178
closed
4 months ago
0
[quidditch_snitch] Implement `pipeline-copy-compute` pass
#101
zero9178
closed
4 months ago
0
[quidditch_snitch] Implement lowering of `pipeline` op
#100
zero9178
closed
4 months ago
0
[tracing] Fix invalid register for repition
#99
zero9178
closed
4 months ago
0
Bump xDSL version
#98
zero9178
closed
4 months ago
0
[quidditch_snitch] Introduce `pipeline` op
#97
zero9178
closed
4 months ago
0
[tracing] Add SSR state
#96
zero9178
closed
4 months ago
0
[tracing] Support single token DMA polls and barrier unlocks
#95
zero9178
closed
4 months ago
0
[SpecializeDMACode] Replace `start_dma_transfer` rather than erase
#94
zero9178
closed
4 months ago
0
[ConvertSnitchToLLVM] Integrate into new `ConvertToLLVM` pass
#93
zero9178
closed
4 months ago
0
[Target] Fork LLVMCPU's `ConvertToLLVM` pass
#92
zero9178
closed
4 months ago
0
[ConvertSnitchToLLVM] Implement single token wait
#91
zero9178
closed
4 months ago
0
[quidditch_snitch] Introduce asynchronous tensor copy ops
#90
zero9178
closed
4 months ago
0
[TensorTile] Add dedicated L1 tiling
#89
zero9178
closed
4 months ago
0
[samples] Use deterministic weights across runs
#88
zero9178
closed
4 months ago
0
Bump xDSL version
#87
zero9178
closed
4 months ago
0
[ConvertSnitchToLLVM] Make sure DMA strides are in bytes
#86
zero9178
closed
4 months ago
0
[quidditch_snitch] Fix missing `offset` application on destination po…
#85
zero9178
closed
4 months ago
0
Next