geohot
da35edbb55
reenable that upcast
2025-04-01 17:09:02 +08:00
geohot
661431ee75
correctness
2025-04-01 17:01:46 +08:00
geohot
8340d9c1c2
disable padding
2025-04-01 16:27:54 +08:00
geohot
910cddbbca
correct but slower
2025-04-01 16:11:47 +08:00
geohot
e6e0c0ec86
should work
2025-04-01 15:25:15 +08:00
geohot
d0eedb5a79
hack
2025-04-01 15:05:13 +08:00
geohot
f69deddbd4
opt
2025-04-01 14:43:36 +08:00
geohot
be11fbbf78
works
2025-04-01 14:38:38 +08:00
geohot
812c391617
fp mul
2025-04-01 13:43:16 +08:00
geohot
3306083f42
YOU DIDNT FOIL
2025-04-01 12:32:00 +08:00
geohot
18d7e9d3f1
oops
2025-04-01 11:56:57 +08:00
geohot
1c3f249ecf
fix multicore flop tracking
2025-04-01 10:16:01 +08:00
nimlgen and GitHub
bb7b89475c
dsp multicore 2 ( #9644 )
...
* dsp multicore 2
* hmm
* better
2025-03-31 23:56:54 +08:00
geohot
8005e6c974
write test pkl imagenet
2025-03-31 19:37:28 +08:00
geohot
a3d61a0372
save pkl from benchmark
2025-03-31 19:31:48 +08:00
geohot
c73e35aa24
non const fix
2025-03-31 19:10:06 +08:00
geohot
0b4b9f61b9
simpler
2025-03-31 19:03:06 +08:00
geohot
ee3ddfcdc1
many l2fetch
2025-03-31 18:58:52 +08:00
geohot
220d682489
prefetch l2 is so winning
2025-03-31 18:29:12 +08:00
geohot
9c388c3539
try to be smarter
2025-03-31 18:23:49 +08:00
geohot
4b3a4c8c46
fix prefetch l2
2025-03-31 18:09:48 +08:00
geohot
eb606d7230
MULTICORE=1 PYTHONPATH=. QUANTIZE=1 DEBUG=2 DEVECTORIZE=0 python3 extra/replay_pkl.py /tmp/im.pkl
2025-03-31 15:37:07 +08:00
geohot
49d52a2763
support acc in __builtin_HEXAGON_A2_vraddub
2025-03-31 15:12:00 +08:00
geohot
a59c3dd09a
err, that's a bug
2025-03-31 14:56:15 +08:00
geohot
a640292aed
delete extra
2025-03-31 14:35:32 +08:00
George Hotz and GitHub
2f48c12441
Merge branch 'master' into dsp_search
2025-03-31 14:27:27 +08:00
George Hotz and GitHub
e4c545b396
linearizer fix from dsp branch ( #9641 )
...
* linearizer fix from dsp branch
* revert that
2025-03-31 14:26:39 +08:00
geohot
be3b5efc64
fix precommit a bit
2025-03-31 14:26:19 +08:00
geohot
996d0ac1d2
multicore all the way
2025-03-31 14:17:19 +08:00
George Hotz and GitHub
ec405b919f
Revert "Revert "do not block gc in UOp.toposort ( #9623 )" ( #9624 )" ( #9639 )
...
This reverts commit 7ef02d0e1c
2025-03-31 14:03:38 +08:00
George Hotz and GitHub
77e897b3b1
Merge branch 'master' into dsp_search
2025-03-31 13:03:29 +08:00
George Hotz and GitHub
49b1c46d16
good changes from the dsp branch ( #9638 )
2025-03-31 13:02:53 +08:00
geohot
273dde69bd
remove range split support
2025-03-31 12:43:21 +08:00
geohot
a64030d8c8
ignore hacks
2025-03-31 12:36:39 +08:00
qazal and GitHub
9d67d3a2f3
simpler viz codeblocks ( #9636 )
...
* simpler viz codeblocks
* err
2025-03-31 11:48:35 +08:00
geohot
9b19129e87
mc
2025-03-31 11:34:22 +08:00
geohot
48221d9024
2 global dim
2025-03-31 11:25:12 +08:00
geohot
bcfcd60f55
opt weights
2025-03-31 11:02:03 +08:00
chenyu and GitHub
60eb0c4ed7
exclude slow tests on PYTHON ( #9634 )
2025-03-30 22:55:05 -04:00
geohot
abc90024ac
hand coded opts
2025-03-31 10:44:09 +08:00
chenyu and GitHub
5012ba3f04
cumalu touchup [pr] ( #9632 )
2025-03-30 22:43:11 -04:00
chenyu and GitHub
d8d7ac1bb1
fix bert free_intermediates ( #9633 )
...
fix when only run eval `TRAIN=0 BERT_SIZE=tiny examples/mlperf/training_submission_v5.0/tinycorp/benchmarks/bert/implementations/tinybox_green/dev_beam.sh`
2025-03-30 22:42:52 -04:00
qazal and GitHub
ff984c807d
hotfix: less lines for viz helpers ( #9631 )
2025-03-31 10:10:34 +08:00
George Hotz and GitHub
f0e6d8394c
Merge branch 'master' into dsp_search
2025-03-31 10:01:19 +08:00
qazal and GitHub
c206a7ae6d
refactor viz state updates ( #9630 )
...
* refactor viz state updates
* onclick
2025-03-31 09:54:54 +08:00
6652003839
Add cumprod to Tensor ( #9629 )
...
* probably how cumprod should look like
* update _cumalu to work with MUL
* shorter
* cumprod testing
* clean
* more cleanup
* add cumprod to torch backend.
* make it look like cumsum
* mypy fix
---------
Co-authored-by: chenyu <[email protected]
2025-03-30 21:49:18 -04:00
d52e91db7b
ONNX ops clean ups ( #9622 )
...
* combine work from remove numpy and onnx ops tests
* clippy
---------
Co-authored-by: chenyu <[email protected]
2025-03-30 21:39:22 -04:00
uuuvn and GitHub
962c0f65f8
Fix generate_am ( #9626 )
...
This should be a comment
2025-03-31 01:15:44 +08:00
uuuvn and GitHub
2a4247b8c2
RDNA 3.5 support ( #9627 )
2025-03-31 01:15:20 +08:00
a08b07b4da
Bump onnx==1.17.0 ( #9618 )
...
* bump
* remove resize tf_crop_and_resize
---------
Co-authored-by: chenyu <[email protected]
2025-03-30 03:21:51 -04:00
geohot
