qazal
52d8afde2b
new viz unittests, isolate the ctx bug ( #7069 )
...
* start new test_viz
* test_rewrite_twice
* test_rewrite_with_ctx
* add back some of the old tests
* lints
2024-10-15 18:53:56 +03:00
nimlgen
9f00eacde5
nv tagged memory + resnet failed kernel ( #7061 )
...
* nv tagged memory
* linter
* metal fix?
2024-10-15 18:19:58 +03:00
chenyu
2008bac6bf
use validhack logic to rewrite buffer idx ( #6740 )
...
* use validhack logic to rewrite buffer idx
saved a whopping one mod in the conv backward kernel...
* cleanup more
2024-10-14 16:47:31 -04:00
chenyu
a99e42cf2f
clean up test_uop_symbolic.py ( #7058 )
...
enable more tests and remove dead tests
2024-10-14 15:35:58 -04:00
chenyu
fbaab30fe3
add timing to fuzz_linearizer ( #7056 )
...
and applied smaller FUZZ_MAX_SIZE. this is getting quite slow in CI
2024-10-14 11:57:41 -04:00
qazal
0ef186d4be
scheduler internal api cleanups [pr] ( #7052 )
...
* delete external_benchmark_ast.py [pr]
* cleanup 2
* random
2024-10-14 15:56:10 +03:00
George Hotz
f85c9ba00a
rewrite max to use cmplt + where ( #7037 )
2024-10-14 20:00:51 +08:00
George Hotz
b20b22a738
hotfix: add test_tiny, because many times it's what you want
2024-10-14 16:32:33 +08:00
Louis Novy
2ac5aec66b
Fix exponential complexity in _is_padding_okay [pr] ( #7008 )
...
* preliminary test
* missed Optional
* don't check for cache during recursion
* match style from st_fixup... may be marginally faster?
* pathological test case: strongly connected DAG
* move to test_schedule as this isn't really a fusion
* oops this shouldn't be edited
* Revert "oops this shouldn't be edited"
This reverts commit 487cb027dc48d8c550ce
2024-10-14 02:34:47 +03:00
chenyu
bd8ecf7fd6
remove NumNode ( #7035 )
2024-10-13 16:42:19 -04:00
chenyu
c4c806a210
generate new kernel dataset ( #7034 )
...
* generate new kernel dataset
pre req to remove NumNode
```
extra/optimization/generate_dataset.sh
gzip -k /tmp/sops
mv /tmp/sops.gz extra/datasets/
```
* fix var range in fuzz_linearizer
2024-10-13 16:19:41 -04:00
chenyu
1a27417262
remove arbitrary multiplication case ( #7033 )
...
adds the wrongly simplified kernel in test_linearizer_failures
#7019
2024-10-13 15:06:05 -04:00
chenyu
13575f080a
remove bitcast backward in function.py ( #7031 )
...
bitcast cannot backward
2024-10-13 10:08:27 -04:00
qazal
13846930cd
hotfix: extract_dataset.py ( #7029 )
2024-10-13 11:18:23 +03:00
chenyu
04d9b46d51
derivative of softmax is indepedent of max ( #7009 )
...
* derivative of softmax is indepedent of max
* update test
2024-10-12 15:59:23 -04:00
chenyu
cae1c41755
test case of softmax backward kernel count ( #7022 )
2024-10-12 15:46:32 -04:00
George Hotz
85a45164fb
remove pyint [pr] ( #7016 )
...
* remove pyint
* bump time on tp [pr]
* dont truncate in const fold
* remove dead code
* Revert "dont truncate in const fold"
This reverts commit 29c81db0f7
2024-10-12 22:36:24 +08:00
George Hotz
38d45dfba5
hotfix: no rng in test/external/external_benchmark_schedule.py
2024-10-12 22:03:04 +08:00
George Hotz
a71bb09ec3
remove symbolic file [pr] ( #7012 )
2024-10-12 18:44:44 +08:00
George Hotz
5ae2de9845
UOp.variable ( #7010 )
...
* UOp.variable [pr]
* fix tests
* clean
* improve name rendering
* last bug
2024-10-12 18:20:44 +08:00
Markiian Novosad
8831c691e2
Add slice parameter type checking to disallow Tensor usage for slices ( #6967 )
...
* add support for single el tensors for slices
* rm trailing spaces
* cleanup long lines
* remove tensor in slice support, add comprehensive err msg
* cleanup getitem, add slice type check
* Edit err message
2024-10-11 16:20:21 -04:00
qazal
7451812bbf
delete AST_REWRITE ctx var ( #6995 )
2024-10-11 11:33:16 +03:00
George Hotz
e7a0ffe46a
break out linearization [pr] ( #6994 )
2024-10-11 15:27:33 +08:00
George Hotz
e441794c4b
remove custom op support, we waste time maintaining this ( #6991 )
...
* remove custom op support, we waste time maintaining this
* customop is over
2024-10-11 14:31:09 +08:00
George Hotz
c08521e823
minor cleanups from toonygrad ( #6990 )
2024-10-11 14:19:10 +08:00
George Hotz
f50d0e0ee0
cloud device [pr] ( #6964 )
...
* first try at cloud device [pr]
* real separation
* we're free
* clang works
* unhappy with timeout
* better timeouts and free
* unrelated
* use http verbs + add test
* lines + better test
* fix DELETE
* shorter cloud
* split key
* fix sending renderer
* PTXRenderer serialization
* add sessions
* http.client
* minor timeout bump
* fix keep-alive
* inc server timeout
* real fix timeout
* that one too
2024-10-11 12:24:06 +08:00
Bhavya Gada
23c09f4b4c
add support for padding='same' in nn.conv ( #6975 )
...
* add support for padding='same' in nn.conv
* express concisely
* simplify loop
* test same padding with dilation and conv1d
* fix bad indentation
* make loop one liner
2024-10-11 11:39:07 +08:00
qazal
4ef5310039
track viz context even if rewrite errors [pr] ( #6976 )
2024-10-10 18:33:15 +03:00
chenyu
592e5f1df2
skip test_viz test_no_dedup_different_opts ( #6979 )
2024-10-10 11:10:24 -04:00
chenyu
e3dc10f8f6
improve fold_unrolled_divs ( #6977 )
...
addressed #6935
the first few terms in fold_unrolled_divs might have been folded already, so the check should first try to add those terms back. there is a case that every but one term is folded which is not an add chain anymore, so just added as a failed test case for now
2024-10-10 10:52:05 -04:00
qazal
3481468702
bring viz to core ( #6970 )
...
* move viz to core
* pathfix
* move test_viz to core
* cleanup test_viz diff
* use contextvars
2024-10-10 16:56:26 +03:00
qazal
3724a66716
move test_viz to test/, prereq for tinygrad/viz [pr] ( #6972 )
2024-10-10 11:40:46 +03:00
qazal
20d3c2d113
unify UOps.SHAPETRACKER and UOps.SWIZZLE with UOps.VIEW ( #6955 )
...
* add UOps.VIEW
* update hardcoded asts
* update sops.gz
2024-10-09 02:00:17 +08:00
qazal
2800520dd5
even smaller process_replay.py [pr] ( #6941 )
...
* even smaller process_replay.py [pr]
* delete those tests
* dedup asts
2024-10-08 20:43:22 +08:00
czhu
08bfa8632b
embedding shape ( #6930 )
2024-10-08 14:42:20 +08:00
chenyu
e4c0743188
failed example for logcumsumexp ( #6936 )
...
need cummax for numerical stability
2024-10-07 10:55:45 -04:00
qazal
b82023c97e
process replay cleanup to generic _pmap [pr] ( #6929 )
...
* process replay cleanup to generic _pmap [pr]
* delete `COMPARE_SCHEDULE`
2024-10-07 13:57:05 +08:00
qazal
16312b4c59
rip out old scheduler process replay stuff, diff pure UOps [pr] ( #6927 )
2024-10-07 13:20:35 +08:00
wozeparrot
9eb6eef441
seed in tensor ( #6869 )
2024-10-06 14:46:58 -04:00
jeffzh4ng
19a7e41113
implement logcumsumexp ( #6921 )
...
* implement logcumsumexp
* change axis=None to axis=0
2024-10-06 10:45:36 -04:00
chenyu
75d9dcf000
support dtype in softmax and log_softmax ( #6914 )
...
matches torch. for mixed precision training, we would want to use float for softmax
2024-10-06 07:18:15 -04:00
chenyu
08414d7b7c
cleanup test_uop_symbolic.py ( #6894 )
...
no more test_symbolic for reference, so force expected output to be exact instead of a set
2024-10-04 20:53:10 -04:00
ignaciosica
555bcb5e54
static access for code_for_op ( #6889 )
2024-10-05 07:38:01 +08:00
vladov
5f6b6162b3
Suppress warnings in transcendental tests. ( #6891 )
2024-10-05 07:37:17 +08:00
George Hotz
4df5c7a4ef
move lazy to engine [pr] ( #6886 )
...
* move lazy to engine [pr]
* engine.lazy
2024-10-04 23:19:26 +08:00
George Hotz
6b063450df
move hcq device to runtime [pr] ( #6879 )
...
* things that are only used in one place don't belong in helpers [pr]
* start moving hcq device [pr]
* fix paths
2024-10-04 22:26:50 +08:00
George Hotz
8ca506ee37
remove the magic methods for moving between devices [pr] ( #6881 )
...
* remove the magic methods for moving between devices [pr]
* remove unneeded clang
2024-10-04 20:27:52 +08:00
George Hotz
a0cb16ac61
node cleanup + local metal test speed [pr] ( #6880 )
...
* node cleanup [pr]
* fix tests, including the double one on metal
* no time tqdm tests
2024-10-04 18:14:23 +08:00
George Hotz
cdff1d75b6
things that are only used in one place don't belong in helpers [pr] ( #6878 )
...
* things that are only used in one place don't belong in helpers [pr]
* pretty print moved
2024-10-04 17:27:38 +08:00
George Hotz
f4ec39fe58
switch symbolic from old to uops, final PR ( #6872 )
...
* switch symbolic from old to uops, final PR
* two wrong answers
* not needed resolves
* symbolic ops passes
* symbolic ops passes
* progress
* tests pass (almost)
* fix last test
* fix some tests
* global binding and unbinding
* Revert "global binding and unbinding"
This reverts commit 9456725630
2024-10-04 16:42:27 +08:00
