github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-14 17:38:06 -05:00
Code Issues Packages Projects Releases Wiki Activity
11,106 Commits 203 Branches 14 Tags
python_speed
Commit Graph

11106 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
nimlgen
ebbcdd6577 cpu: use suppress_finalizing (#12071) 2025-09-08 18:28:09 +03:00
qazal
73ca0e870c viz: index visible rects (#12070) 2025-09-08 17:37:17 +03:00
chenyu
d40f5b766b default BEAM_PADTO to 0 (#12069) 
seems incorrect, disable by default now
 2025-09-08 10:17:03 -04:00
Sieds Lykles
75b58fe2d3 move simplify_valid pat to sym (#12065) 
* move simplify_valid pat to sym

* fix expectedfailure
 2025-09-08 07:01:26 +02:00
chenyu
56861852be enable IMAGE for test_mnist and test_mnist_backward (#12064) 
passes now
 2025-09-07 09:06:39 -04:00
nimlgen
ef71acc88a hcq: cleanup fileio iface (#12063) 
* hcq: cleanup fileio iface

* typo

* _
 2025-09-07 15:43:27 +03:00
nimlgen
35ddfc3d39 change default cpu_count (#12062) 2025-09-06 23:30:20 +03:00
nimlgen
97187bf8b6 cleanup win and arch checks (#12060) 
* cleanup win and arch checks

* stupid mypy
 2025-09-06 23:08:46 +03:00
Sieds Lykles
f326df8ae8 add type: ignore (#12059) 2025-09-06 21:17:35 +02:00
George Hotz
c66935f7b9 only run hcopts once (#12053) 
* only run hcopts once

* same?
 2025-09-06 11:14:52 -07:00
qazal
801be5f7b9 viz: memory graph cleanups (#12057) 
* delete the total nbytes tooltip

* split pixel rescaling from layout
 2025-09-06 19:44:53 +03:00
nimlgen
10ac427aaa cpu threading (#11951) 
* start cpu threading

* fix

* fix2

* fix

* hacks?

* threads

* minor

* no dsp

* dsp 2

* n

* more

* test

* xm

* cleaner

* readable

* f

* reorder

* when no threads

* rangeify

* typos

* not needed

* reapply

* remoev this

* linter

* fixed cpu count in ci

* fix

* fixes

* rm

* typo

* sort based on speed

* test if test works in ci

* Revert "test if test works in ci"

This reverts commit 1f05edb531.

* do not pad thread
 2025-09-06 16:13:43 +03:00
nimlgen
2b1844da27 cpu: support several threads in runtime (#12055) 2025-09-06 13:29:31 +03:00
nimlgen
f37b836618 factor out _globalizable_rngs (#12054) 2025-09-06 13:29:23 +03:00
nimlgen
1630c87d0e run optimize_local_size only when locals supported (#12056) 2025-09-06 13:29:09 +03:00
Jordan Chalupka
48ec5efad9 only run autogen tests on change (#12049) 
* only run autogen tests on change

* example change

* rm example change
 2025-09-05 23:53:01 -07:00
Sieds Lykles
581b2388c2 add dtypes.index (#12015) 
* add dtypes.index

* cast shape, stride and mask to dtypes.index in view.create

* move pm_lower_index_dtype to ops

* DEFINE_VAR is dtype.index by default

* merge var_val_using_str

* remove int from commutative

* fix test_rewrite_map

* change that to dtypes.index

* change some int to index

* shorten those

* remove old cast in renderer

* cleanup

* change that back

* add comment

* delete comment

* just delete those

* view doesnt have to cast anymore

* adjust comment
 2025-09-06 06:03:44 +02:00
Sieds Lykles
c6c16b2946 var_vals uses str for var (#12011) 
* var_vals is str,int

* remove imports

* remove print

* fix test

* change var_vals in hcq

* update test_hcq

* fix multitensor _device_num var

* fix syminfer test

* shorten line

* p.vars stays list[Variable]

* shorten line

* vars is back to tuple[Variable, ...]

* change var_vals in extra

* change var_vals from shapetracker

* var_vals is str:int

* fix signature
 2025-09-06 04:16:12 +02:00
George Hotz
8658a97197 hotfix: name the shift rewrite better + no ctx there 2025-09-05 19:01:59 -07:00
George Hotz
6ef3270fc8 fix opt gate (#12050) 2025-09-05 18:59:54 -07:00
George Hotz
66c5206b42 hotfix: minimal scheduler copy 2025-09-05 18:24:00 -07:00
George Hotz
478e758755 Revert "fix scheduler copy (#12048)" 
This reverts commit 51b7c40788.
 2025-09-05 18:21:55 -07:00
George Hotz
51b7c40788 fix scheduler copy (#12048) 
* fix scheduler copy

* hand coded opt only runs once
 2025-09-05 17:17:49 -07:00
George Hotz
0123c394e5 early simplfy_merge_adjacent (#12045) 
* do simplify_merge_adjacent before schedule

* do simplify_merge_adjacent before schedule

* disable that slow test
 2025-09-05 16:39:20 -07:00
George Hotz
8423c06144 delete unused bufs_from_lin (#12044) 2025-09-05 16:08:28 -07:00
George Hotz
38dcadf07b delete kernel.py (#12040) 
* delete kernel.py

* delete that file

* rip and tear

* don't test search

* imports

* fix torch frontend

* not a part of regen
 2025-09-05 15:52:07 -07:00
George Hotz
ee4f696086 delete more tests (#12043) 
* delete more tests

* delete and simplify

* flaky on windows

* a few more, those remained
 2025-09-05 15:31:30 -07:00
George Hotz
12c7b1bb01 cleanup lin tests without Kernel (#12041) 
* cleanup lin tests without Kernel

* no kernel.py there

* remove that test
 2025-09-05 15:13:14 -07:00
Sieds Lykles
8435d2d23b fix openpilot speed regeression (#12039) 
* set local_size=None if special.arg[0]=='i'

* add cast back
 2025-09-06 00:05:45 +02:00
George Hotz
e00858a2c3 only POSTOPT (#12038) 2025-09-05 14:46:33 -07:00
George Hotz
433581f8ed make POSTOPT=2 the default (#12034) 
* make POSTOPT=2 the default

* more matching tc

* fix winograd

* fix that test

* add matvec to Scheduler

* flip tc sort order

* similar speed

* fix beam on image

* disable slow tests

* slow
 2025-09-05 14:34:05 -07:00
chenyu
3b41a04b96 remove test_openpilot in test_onnx (#12037) 
openpilot is tested in compile3
 2025-09-05 16:20:03 -04:00
Sieds Lykles
290521f68e add check for z3>=4.12.4 (#12035) 2025-09-05 20:33:26 +02:00
George Hotz
870f63d9cc add WARP axistype, fix postopt bugs (#12033) 
* postopt is 83% match

* warp is bright CYAN

* beautiful mnist beam works

* fix shutdown bug
 2025-09-05 10:36:55 -07:00
chenyu
4c2d4f683a lower universal_test_unary cos domain (#12032) 
flaky
 2025-09-05 12:19:44 -04:00
chenyu
a340723bf1 SKIP_SLOW_TEST=1 for nv CI (#12031) 2025-09-05 11:52:02 -04:00
chenyu
ce7163e9b4 clean up skip slow tests in PYTHON (#12028) 
skip with SKIP_SLOW_TEST and decorators
 2025-09-05 11:35:26 -04:00
qazal
f08299d2ec viz: small profiler resizing improvements (#12026) 
* switch to ResizeObserver

* set a fixed size for device-list

* less

* height from devices

* int

* side rect, more const
 2025-09-05 18:29:03 +03:00
chenyu
5dcc4c7f1b skip test_linalg in windows unit test (#12030) 2025-09-05 11:28:40 -04:00
George Hotz
f8e2dd4dd1 investigate opts mismatches (#12020) 2025-09-05 07:40:29 -07:00
chenyu
e0da644171 lower sample count in test_multinomial (#12027) 2025-09-05 10:10:28 -04:00
chenyu
9b6f1b86cb add Tensor.maximum in test_dtype_alu (#12025) 
works except nan
 2025-09-05 09:48:39 -04:00
nimlgen
3e1c04bcdf jit: noopt for copy buffers (#12023) 2025-09-05 16:04:35 +03:00
qazal
ab413ce72f viz: give tooltips a max-width (#12022) 
* viz: give tooltips a max-width

* better
 2025-09-05 14:25:38 +03:00
qazal
f461ccf407 exclude op2 nan lt in test_dtype_alu (#12024) 
failure: https://github.com/tinygrad/tinygrad/actions/runs/17490320000/job/49679581331?pr=12022#step:6:125
 2025-09-05 14:14:22 +03:00
nimlgen
4fcea8493d viz: add label to tooltip (#12021) 2025-09-05 13:06:33 +03:00
George Hotz
2b5a73ac65 improve test_linearizer (#12016) 
* improve test_linearizer

* tweaks

* simpler

* get_prg

* that one doesn't have to return

* fix postopt bugs

* fix rng
 2025-09-04 20:44:05 -07:00
chenyu
7f3df6ea21 exclude nan in test_dtype_alu lt (#12019) 2025-09-04 23:38:37 -04:00
Sieds Lykles
f5404ca53c Divmod combine - associative variations (#12017) 
* add rule and test

* more rules and tests

* add all four variations

* fix test

* test fixed!

* adjust commment

* add new variations

* disable intel tensor core ops count test for bigger_matmul_half
 2025-09-05 03:44:02 +02:00
chenyu
677220ae7e test_tesnor_data to unit/ (#12013) 2025-09-04 19:58:27 -04:00