* clean up opt
* don't let global kernels get too small
* 8192 -> 1024
* disable local shape for clang
* fix can_merge
* unroll the 5x5 depthwise convs in op
* load float4 check
* Less, LessOrEqual, Greater, GreaterOrEqual, Equal
* lint fix
* using built in functions
* overriding __eq__ breaks things
* backwards pass for less - foward only tests
* one other spot
* removing backwards for comparison ops to match pytorch
* raise runtime error
* more tests for comparison ops
* fixed the lineup
* added number upcast tests
* conv2d is an hlop
* shorter conv
* KOPT=-1
* alt imp
* MULACC
* smarter mulacc
* pop conv
* 7x7 -> 5x5
* didn't fix, that's not going to work
* this is faster and matches old behavior
* oh, non lazy just won't work with mulacc
* mulacc in torch
* bool types were creeping in
* optimizer is actually better with hlop conv
* fix pushing permutes issue
* refactor einsum_mulacc
* fix up readme
* update readme
* _image_conv2d
* fix bias addition location
* pushing permutes gets back to 200 kernels
* conv cleanup
* disable hlop conv
* don't hide that in helpers
* triton can add
* print stuff from triton
* write out file
* ops triton working
* reduce ops
* sort of works
* Triton bugfixes & implementation of remaining ops (#490)
* padding
* support pow, max, relu, gt0
* allocate return buffer
* Fix reduce
* Add tests for power op
* Fix triton illegal memory accesses and memory leak (#512)
* Fix mypy issue
* Add triton to setup.py
* Replace torch with pycuda
* Use one cuda stream for data transfer and kernels
* Remove triton submodule
* Fix memory leak by using weakrefs for caching
* Fix memory access by adding valid as mask for load
* Fix invalid kernel launches by flattening the grid (#515)
---------
Co-authored-by: Martin Loretz <20306567+martinloretzzz@users.noreply.github.com>
* Refactor getenv into helpers
* Remove unused os
* Fix default value
* Fix more defaults for CI
* Fix bracket
* Revert changes to openpilot/compile.py
* Use getenv from helpers when possible
* factor out opencl runtime
* don't use CL outside the runtime
* cuda runtime adds
* final_dimension
* tests pass with CUDA backend
* more cuda
* cuda simpler
* retain old functionality
* linter and typing
* move globalcounters out of runtimes
* oops, GlobalCounters in cuda
* MAX_OUTPUT_SHAPE=3 is fine for CUDA
* add image
* load + store + boring stuff:
* image tests pass
* thneed print GFLOPS
* op conv test
* more debugging
* hack for multiview image
* shapetracker creates less views
* disable image tests
* working better
* ugh, lkey not key
* print in DEBUG, and allow views
* works
* simple padding conv2d
* use index for image
* that was bad code
* debug print
* fix types
* less lines
* save lines
* working exec ast
* exec_ast is staticmethod
* GenericExecAST
* fold that sometimes
* ExplicitExecAST
* exec_ast for GPU
* gpu working
* get_lazyop_shape
* now gpubuffer is ExplicitExecAST
* dedup
* add a type
* RESHAPE in opencl code
* fix linter
* that too for linter
* cleanups
* remove dead code
* GenericShape is less lines
* add ALLOWED_KERNEL_COUNT to tests
* fix mypy
* that's gotta be recursive
* fix opencl shape processing
* remove unneeded lambda
* in progress
* big conv test works
* that's unneeded
* fix opencl with reduce
* rewrite contiguous_view_constant_fold
* clean up mids in loop code
* subidx
* print cl kernel before run
* no reduce, no loop
* Revert "no reduce, no loop"
This reverts commit 92777e40e9.
* option for matmul
* fixups
* fast like a nascar
* running
* thneed runner
* no buffer id makes no backing buffer
* move constant folding to the top
* runs on mac
* folded biases
* was v slow
* maybe just that
* elu touchup
* speed and float32
Co-authored-by: Comma Device <device@comma.ai>
