Commit Graph

9161 Commits

Author SHA1 Message Date
George Hotz
5dc1bc6070 switch get_kernel -> get_program [pr] (#10817)
* switch get_kernel -> get_program [pr]

* fix tests
2025-06-15 12:26:50 -07:00
George Hotz
a36b09a715 universal device import [pr] (#10818) 2025-06-15 12:01:02 -07:00
George Hotz
cc5e4e54b8 move type verify to codegen [pr] (#10816) 2025-06-15 12:00:52 -07:00
George Hotz
27cf836958 split ocelot out for autogen, fix CI (#10819)
* split ocelot out for autogen, fix CI

* mac ocelot
2025-06-15 11:37:23 -07:00
Ahmed Harmouche
c380efc220 Support aarch64 linux on webgpu (#10802) 2025-06-14 14:57:18 -04:00
Sieds Lykles
37d3ca152e Adapt >> for division by power of two to all ints (#10803)
* Change divison by power of two to always use shift

* Change test to test int instead of uint

* simplify condition

* add old rule back with comment

* remove import

* use sresolve instead of simplify

* use keyword in simplify instead of sresolve

* webgpu cast y to uint

* remove comment

* explicitly set dtype in wgsl

* without simplify

* undo simplify kwarg

* change test to test both int32 and uint32
2025-06-14 14:55:51 -04:00
chenyu
652db5702b move test_conv_shapetracker and some test_search util into unit test (#10812) 2025-06-14 13:29:32 -04:00
George Hotz
754667093f remove IGNORE stuff (#10796)
* remove IGNORE stuff, was this even tested? [pr]

* delete IGNORE op
2025-06-14 09:59:45 -07:00
leopf
118a09ddcf xor self folding (#10806)
* xor folding

* tests + z3 bitwise xor
2025-06-14 10:01:17 -04:00
qazal
8e6ac18436 viz: make sidebar list responsive to keyboard smashing (#10811)
* expanded is a static style

* only draw the list once

* identify with ids

* state isn't used here anymore

* only toggle states

* less
2025-06-14 13:52:27 +03:00
chenyu
8c28b5d833 move dtype spec tests into unit test (#10808)
* move dtype spec tests into unit test

can clean up more after the split

* skip CI test_backward_sum_acc_dtype
2025-06-13 22:21:22 -04:00
chenyu
7a6df0a161 remove .relu() call in several conv tests in test_ops (#10807)
* remove .relu() call in several conv tests in test_ops

testing negative parts double the effectiveness. keep the relu between two convs and the tests that explicitly test relu

* relax tol
2025-06-13 17:10:16 -04:00
nimlgen
b6e574fcdf am: smu 14.0.3 is smu 14.0.2 (#10714) 2025-06-13 23:07:56 +03:00
chenyu
7d5c769c6b fix compile4 (#10797) 2025-06-12 22:28:56 -04:00
wozeparrot
c01b20fd83 amd: more verbose out of memory error (#10798) 2025-06-12 19:06:58 -07:00
geohotstan
806b68c2b3 Add fallback dtype to ONNX (#10788)
* start

* still need the float16 workaround in

* tiny nit for correctness

* idk hacks, I need to understand this device stuff better

* no-op?

* remove that assert for true nooooooop

* add fallback_context
2025-06-12 20:39:21 -04:00
George Hotz
dcd1928f29 tensor cores for gfx1200 [pr] (#10795) 2025-06-12 16:33:29 -07:00
qazal
a113c5e3ae viz: update browser test to properly shutdown [pr] (#10793)
Using `await page.evaluate` can cause non deterministic `TargetCloseError`
exceptions if it cannot find the elements on the page, Puppeteer
doesn't cleanly stop when `browser.close()` is called.
[Failing CI](https://github.com/tinygrad/tinygrad/actions/runs/15596803685/job/43928961323?pr=10763#step:9:61)
2025-06-12 17:58:42 +03:00
Dan German
24e7aed74b ramp.py: correct UOp and Ops import path from tinygrad.uop to tinygrad.uop.ops (#10791) 2025-06-12 10:07:03 -04:00
qazal
c066baea65 viz: enter key only expands steps (#10792)
It shouldn't be changing any step or context state. Those are handled
explicitly by the arrow keys (or clicking).
2025-06-12 16:00:14 +03:00
qazal
822e2dcb20 viz: back button returns to the kernel graph (#10790)
* create space

* viz: back button returns to the kernel graph
2025-06-12 15:19:48 +03:00
chenyu
4242b9874e remove AMD_LLVM=0 in mlperf and search ci (#10785)
tinybox updated to llvm 20
2025-06-11 21:10:31 -04:00
wozeparrot
eb739bb96a hotfix: lower threshold (#10786) 2025-06-11 19:36:20 -04:00
wozeparrot
53edd49a33 feat: bump to llvm20 (#10784) 2025-06-11 16:04:18 -07:00
chenyu
7d8939908f AMD_LLVM=0 for resnet cron (#10780)
similar pf on llvm19 and fine on 20
2025-06-11 16:28:40 -04:00
qazal
a6af8db4d3 viz work from the profiler (#10781)
* inline ansistrip

* refactor to changeStep + explicitly set expandSteps
2025-06-11 23:20:41 +03:00
Sieds Lykles
10b61157b9 Support symbolic slice with no start [pr] (#10775)
* add symbolic slice with no start

* reshape the test

* step must be int

* just add a cast...

* more cast...
2025-06-11 16:00:38 -04:00
chenyu
d465ef4acb AMD_LLVM=0 for sdxl search (#10779)
hangs with llvm19 but seems fine with llvm20
2025-06-11 14:56:45 -04:00
uuuvn
0d45e1a3ec Explicitly use CUDA_KERNEL_NODE_PARAMS v1 (#10776) 2025-06-11 16:27:50 +03:00
George Hotz
a38947b4bb move symbolic and transcendental to uop [pr] (#10771) 2025-06-10 20:51:22 -07:00
chenyu
81e296d7b8 remove Tensor.test() in retinanet (#10770)
test was removed
2025-06-10 22:14:57 -04:00
chenyu
25304c3dd0 default AMD_LLVM=1 (#10253) 2025-06-10 18:19:21 -04:00
George Hotz
9d0383634d bump cache and include full python version [pr] (#10768)
* bump cache and include full python version [pr]

* stupid windows

* really stupid windows
2025-06-10 15:07:30 -07:00
chenyu
612cdf5146 move fuzz_shape_ops to run with other fuzzer (#10767)
* move fuzz_shape_ops to run with other fuzzer

* don't skip CPU
2025-06-10 17:43:04 -04:00
chenyu
5e7ad70aae don't run linearize().uop tests in get_action_space test (#10766)
* don't run linearize().uop tests in get_action_space test

this part takes 2 minutes in CI and has nothing to do with action space. also not sure if the "for some reason" comment is still relevant

* -n=auto test/models
2025-06-10 17:23:53 -04:00
b1tg
52c49dd4f3 fix onnx ci (#10762)
Co-authored-by: b1tg <b1tg@users.noreply.github.com>
2025-06-10 14:28:40 -04:00
qazal
9e1d1ebc52 print tag in UOp [pr] (#10755) 2025-06-10 21:16:07 +03:00
chenyu
14fa62c61d move high level tests to unit (#10760)
either no need a backend, or running on one to check suffice
2025-06-10 12:55:44 -04:00
George Hotz
0fbf3f5554 Revert "Revert "Update autogen ci runner to ubuntu 24.04 (#10736)" (#10757)" (#10758)
This reverts commit a6dba9b9d9.
2025-06-10 09:32:27 -07:00
George Hotz
a6dba9b9d9 Revert "Update autogen ci runner to ubuntu 24.04 (#10736)" (#10757)
This reverts commit 1d15374c7a.
2025-06-10 09:31:51 -07:00
uuuvn
1d15374c7a Update autogen ci runner to ubuntu 24.04 (#10736)
For `kfd.AMDKFD_IOC_EXPORT_DMABUF`
2025-06-10 08:33:02 -07:00
Adrian Wijaya
78b9c30640 move idiv to MathTraits [pr] (#10748) 2025-06-10 08:32:09 -07:00
Sieds Lykles
0daa4c6ed0 Add DType.min and DType.max properties (#10749)
* add properties

* cleaner test

* remove added newline
2025-06-10 08:31:34 -07:00
qazal
5d9c274924 keep UOp tags if sources are replaced (#10754)
* keep UOp tags in unified_rewrite

* add failing test, print tag if defined

* remove the repr change
2025-06-10 08:30:14 -07:00
qazal
3de4c9839f viz: display UOp tags (#10751)
* viz: display UOp tags

* g.tag
2025-06-10 16:02:23 +03:00
nimlgen
800d1796d5 am_smi: kill process group (#10750) 2025-06-10 15:23:39 +03:00
qazal
5bd4ad2e8b viz: remove unused arg (#10747) 2025-06-10 12:00:09 +03:00
George Hotz
413e223d6e Revert "remove cpu graph, it's different from the others (#10743)" (#10745)
This reverts commit 3d64a98432.
2025-06-09 22:40:48 -07:00
George Hotz
3d64a98432 remove cpu graph, it's different from the others (#10743)
* remove cpu graph, it's different from the others

* remote was blacklisting CPUGraph
2025-06-09 22:17:10 -07:00
George Hotz
245b1d3a46 move add/mul to MathTrait [pr] (#10741)
* move add to MathTrait [pr]

* both add and mul
2025-06-09 21:48:55 -07:00