issues
search
rohany
/
taco
The Tensor Algebra Compiler (taco) computes sparse tensor expressions on CPUs and GPUs
http://tensor-compiler.org
Other
6
stars
3
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
legion/solomonikMM: modernize solomonikMM's code generation
#140
rohany
opened
2 years ago
0
DISTAL: stop linking libcudart into DISTAL runtime
#139
rohany
closed
2 years ago
0
stop requiring applications to use OpenMP flags
#138
rohany
opened
2 years ago
0
*: move format converter to JIT API in apps
#137
rohany
closed
2 years ago
0
scripts: make install script choose sensible defaults
#136
rohany
opened
2 years ago
0
DISTAL: log output to a Realm::Logger rather than random print statements
#135
rohany
opened
2 years ago
0
*: begin porting applications to using the JIT API
#134
rohany
closed
2 years ago
0
support distributed temporaries with precompute
#133
rohany
opened
2 years ago
0
*: support for JIT-ing GPU kernels
#132
rohany
closed
2 years ago
0
*: first pass at moving the DISTAL runtime into a library
#131
rohany
closed
2 years ago
0
DISTAL Library Checklist
#130
rohany
opened
2 years ago
1
extend TBLIS leaf kernel strategy to use CuTensor for GPUs
#129
rohany
opened
2 years ago
0
*: initial version of a DISTAL library with jit and aot modes
#128
rohany
closed
2 years ago
0
legion/legion: integrate newest control_replication to DISTAL
#127
rohany
closed
2 years ago
0
scripts: make install script more robust
#126
rohany
closed
2 years ago
0
*: implement Tensor Distribution Notation on dense tensors
#125
rohany
closed
2 years ago
0
Install script
#124
rohany
closed
2 years ago
1
cmake: use FindBLAS to find and link OpenBLAS / MKL
#123
rohany
closed
2 years ago
0
rewrite RectCompressedFinalizeYieldPositions CPU to not require explicit template specialization
#122
rohany
closed
2 years ago
0
OpenBLAS: bump version of OpenBLAS to pull in performance fixes for cascade lake
#121
rohany
closed
2 years ago
0
add installation script that manages and builds all dependencies upon installation
#120
rohany
closed
2 years ago
0
Add a leaf call transform for TBLIS
#119
TimothyGu
closed
2 years ago
4
[WIP]: Enable MKL support for DISTAL
#118
rohany
opened
2 years ago
1
spmv: add a call to a cuda kernel that correctly creates a new pos array for CUSPARSE calls
#117
rohany
opened
2 years ago
0
*: integrate with collective instances when ready
#116
rohany
opened
2 years ago
1
mappers: integrate backpressuring for index launches
#115
rohany
opened
2 years ago
0
lowerer: implement pos distribution with outer partitioning
#114
rohany
opened
2 years ago
0
distal: implement Tensor Distribution Notation syntax as in the SpDISTAL paper
#113
rohany
closed
2 years ago
0
*: several bug-fixes around implementing the batched SpMM schedule
#112
rohany
closed
2 years ago
0
*: progress on fixing all of the dense code with SpDISTAL
#111
rohany
closed
2 years ago
0
spdistal: implement a subset of SpDISTAL's tensor distribution notation extensions
#110
rohany
opened
2 years ago
0
*: support for controlling FieldIDs dynamically instead of compile time
#109
rohany
closed
2 years ago
0
Merge sparse formats branch into main, integrate with dense tensor ops
#108
rohany
closed
2 years ago
0
make field IDs for particular regions a GetProperty
#107
rohany
closed
2 years ago
0
scripts: experimenting with ways to display the GPU data
#106
rohany
closed
2 years ago
0
legion/spmv: several bugfixes for running SpMV weak scale
#105
rohany
opened
2 years ago
0
petsc: add an SpMV weak scaling benchmark
#104
rohany
closed
2 years ago
0
legion: progress towards getting a DISTAL spmv weak scaling benchmark
#103
rohany
closed
2 years ago
0
*: try a 2D decomposition of GPU SpMM
#102
rohany
opened
2 years ago
0
communicate: individual regions should allow for untrack, rather than the task as a whole
#101
rohany
opened
2 years ago
0
legion: temporary fix for dumping multi-dimensional data to files
#100
rohany
closed
2 years ago
1
*: efficient way to force nearest neighbor communication for benchmarks
#99
rohany
closed
2 years ago
0
spmm: bugfixes for SpMM CPU benchmarks
#98
rohany
closed
2 years ago
0
*: support a batched schedule for SpMM that uses less memory
#97
rohany
opened
2 years ago
1
legion/spmv: switch CuSparse over to old API
#96
rohany
closed
2 years ago
0
*: move to using a CuSparse kernel for SpMV instead of TACO
#95
rohany
closed
2 years ago
0
lower: use better heuristics for CPU load balancing of OpenMP loops
#94
rohany
closed
2 years ago
0
lower: change how partition generation of nonzero structure preservin…
#93
rohany
closed
2 years ago
0
*: switch generated code to use 64bit loop indices
#92
rohany
closed
2 years ago
0
*: use dynamically scheduled loops when non-statically partitioned loops
#91
rohany
closed
2 years ago
0
Next