* update the backend to fix torch deprecation warning
* use param_hook to avoid full backward hook needlessly firing on inputs which do not require gradients
* fix indentation
---------
Co-authored-by: chenyu <chenyu@fastmail.com>
* better viz names
* delete unused
* don't use opacity, it's multiplicative
* keep styles
* scrollbar coloring
* pyrender doesn't work here
beautiful_mnist r_64_16_32_36@lower all index dtypes
* nak works
* TestOps::test_add works
* testop has no crashes
* fix bool casts
* fix typo
* add disassemble
* RANGE and locals/regs
* simplify NAKCompiler
* disass cleanup
* cleanup nir codegen
* almost all tests passing
* cleanup notes in extra/
* old notes
* only import nak if NIR=1
* fix new SPECIAL syntax
* fix local/shared memory
* more tests passing
* add DEFINE_VAR support
* llvmpipe kinda works
* diskcache
* some mypy stuff
* lvp passing test_ops.py
* fix imports
* actually fix imports
* remove 'stdout'
* fix llvm import
* fix mypy issues
* nicer errors
* simpler test_dtype skips
* test lvp in CI
* fix github action syntax
* fix more actions typos
* switch to mesa 25.1.0
* diskcache_put
* better generation for lvp nir_options
* b64encode shader blobs
* Revert diskcache changes
This reverts commits 930fa3de8a and 8428c694b3.
* general cleanup
* better error messages
* fix llvm import
* fix windows tests
* link with libm and libgcc_s
* fix some errors
* dont check for 'float4'
* NIR uses pointer arithmetic
* use tinymesa
* bump tinymesa
* bump tinymesa again
* update lvp nir_options
* print nir shader with DEBUG
* simplify LVPCompiler
* more tests
* "gated" STORE
* NAK is cacheable
* more tests
* all tests pass locally for NAK
* test autogen in CI
* autogen deps
* more deps
* fix uop_gc
* fix macos
* mypy
* save 2 lines
* save two more lines
* save 1 line
* save 4 lines
* save more lines
* Revert "save more lines"
This reverts commit dd3a720c5a.
* save more lines
* fix LVP on windows
* refactor
* reorganize some code
* refactor lib_gpu
* move LVP check
* out of order loads
* remove support.mesa
* bump tinymesa version
* simplify LVP jit
* macos
* macos ci
* shell: bash
* testing
* more testing
* compute brew prefix
* stupid typo
* actually fix
* lib
* stdout on macos
* inline gallivm_compile_module
* Revert "inline gallivm_compile_module"
This reverts commit b65983b151.
* elf macos
* semicolon
* inherit from CPULLVMCompiler
* ruff
* disas test
* fix libm linking
* default is fine actually
* arm works
* add elf loader link test
* fix NAK beam
* pylint is too smart by half
---------
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
Co-authored-by: nimlgen <138685161+nimlgen@users.noreply.github.com>
* tbgpu
* works
* cleaner
* this
* zero size
* h
* fix
* simpler
* prio over usb
* c
* not needed
* linter
* this way
* mappings
* mypy
* mypy
* mypy 2
* nn
* work on shape property
* reshape causing issues
* more mops
* all mops
* need to cache it
* _shape is like _device
* mostly works
* shape is good
* const uses _shape
* fix tests
* size doesn't use st
* close
* test is broken
* one less st
* hack for 3 op assign
* oops, i didn't mean to change that
* support emulate in the NullDevice
* reproed failure in emulation
* fix wmma
* feat: initial tinyfs device
* feat: don't allow compute on tinyfs device
* feat: tensor helpers to load and store
* feat: bufferview for tinyfs
* fix: keep copy sizes correct
* fix: recv large
* clean: unneeded
* feat: comment
* clean: unneeded
* clean: remove
* clean: remove
* feat: get request tag
* feat: rename to cloud
* feat: send request_id
* feat: start computing tree
* feat: compute store tree on this side
* feat: jank chunked load
* feat: more debugging
* feat: rename to just load and store
* feat: correct chunk count
* fix: fix load for < 1mb
* feat: comments
* feat: don't truncate on block devices
* feat: better way of testing block device
* feat: don't need to pad that much
* feat: connect to nodes directly on load
* feat: cache connections
* feat: don't hard code chunk size
* feat: close mmap when closing file handle
* feat: don't overwrite stuff on disk if storing from disk
* clean: debug print
* fix: close mmap
* feat: await workers
* feat: fast copy from tinyfs to disk
* feat: don't copy to device on last
* feat: use single socket per device
* feat: raid in tinyfs
* clean: remove import
* clean: type
* feat: maintain single event loop
* feat: lower worker count
* feat: use connection pool
* feat: fetch mapping in its own process
* fix: release lock
* feat: don't fetch if exists
* feat: req id only on stores
* feat: always fetch
* fix: rangeify
* feat: allow specifying raid root
* fix: dealloc buffer
* feat: start support non 0 offset
* clean: use cleaner
* feat: don't pass to threadpool
* clean: typing
* move where clauses to load
* shorten line
* drop clauses if they are duplicated
* add rule for swapped where branch
* where on ungated load
* dont move clause if load is in the clause
* parse_valid returns None
* no data dependent branches
* fix rule
* enable swapped rule
* remove those
* disable cpu_access in the sqtt buffer allocation
not sure if this is required, it results in a very slow call to
pcie_mem_write over USB GPU, removing it worked fine.
* fix itrace_se_mask on gfx12
on gfx11 it gave 6 se, on gfx11 this value is 2 so no instructions were
traced.
* Revert "fix itrace_se_mask on gfx12"
This reverts commit 0644adbcd1.
