github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-22 21:38:10 -05:00
Code Issues Packages Projects Releases Wiki Activity
8,764 Commits 204 Branches 14 Tags
08ab184dfd0ddfafeb8516101d0e85a3901f8893
Commit Graph

8764 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
nimlgen
08ab184dfd usbgpu: copyin over 100mb/s (#10259) 
* usbgpu: over 100mb/s

* align

* h
 2025-05-12 16:52:43 +03:00
Kirill R.
4c7c139102 Use cmod/cdiv in sym_infer (#10258) 
* Use cmod/cdiv in sym_infer

* test

---------

Co-authored-by: chenyu <chenyu@fastmail.com>
 2025-05-12 09:07:28 -04:00
chenyu
0015b3921f sleep more in CI Remove amdgpu (#10261) 
see if this is less flaky
 2025-05-12 08:13:44 -04:00
qazal
95c6a736a9 fix FUSE_ARANGE=1 for bert (#10255) 2025-05-12 14:44:05 +03:00
Sieds Lykles
7c4b381fbf Extra simplify valid test [pr] (#10256) 
* add test

* Change the range

* add todo test
 2025-05-12 07:32:03 -04:00
b1tg
7eeb35ba6f fix AMD LLVM compile error for bf16 cifar (#10254) 
* fix AMD LLVM compile error

* remove llvm_bf16_cast

---------

Co-authored-by: b1tg <b1tg@users.noreply.github.com>
 2025-05-12 01:57:07 -04:00
uuuvn
a0ed1ec1ae Faster remote server (#10235) 2025-05-11 19:15:05 -07:00
b1tg
41f5ece877 add nsw flag (#10249) 
Co-authored-by: b1tg <b1tg@users.noreply.github.com>
 2025-05-11 19:14:32 -07:00
George Hotz
8864ff894b hotfix: that repeat_kv belongs outside the if 2025-05-11 18:43:01 -07:00
George Hotz
98c84a711d min rectified flow example [pr] (#10252) 
* work on minrf example

* more

* jit sample

* t is tensor not const

* fixes

* more convs

* fix dropout

* don't print

* 504

* big patch

* onehot

* touch

* use embeddings

* dumb uses final layer

* act

* non fl

* match

* tp

* 3

* of

* ppsz

* normal

* add adln

* no t

* weird transformer

* weird transformer

* contig

* actual speed fix

* dumb

* cb

* 0

* t is 0

* mort-t

* args

* dumb days are over

* readable

* contig

* no more t mask

* mask_t

* init to zero

* clean

* steps

* work

* tt

* t

* solid
 2025-05-11 18:36:44 -07:00
chenyu
70c797b107 train bert tests (#10248) 
added a working bert tiny test, and a failed bert FUSE_ARANGE test
 2025-05-11 08:42:08 -04:00
George Hotz
b2df4cb696 add support for amdgpu-flat-work-group-size to AMD LLVM IR (#10246) 
* add support for amdgpu-flat-work-group-size to AMD LLVM IR

* don't spam llvm init

---------

Co-authored-by: qazal <77887910+Qazalin@users.noreply.github.com>
 2025-05-10 19:11:10 -07:00
qazal
9210280811 add v_fmac_f16 vop3 instruction to remu (#10247) 
* fmac vop3

* from the box
 2025-05-10 23:48:25 +03:00
George Hotz
697259a8a1 amd_comgr_action_info_set_options was deprecated [pr] (#10245) 
* amd_comgr_action_info_set_options was deprecated [pr]

* more standard
 2025-05-10 11:59:04 -07:00
Kevin Buhler
2e0990c4e9 even spacing in viz nodes (#10168) 
* even spacing in viz nodes

* precise dy value

* dominant-baseline text-after-edge

* add STROKE_WIDTH constant, delete dominant_baseline attr

---------

Co-authored-by: qazal <77887910+Qazalin@users.noreply.github.com>
 2025-05-10 10:35:10 +03:00
chenyu
d0e9b74f40 minor div_and_mod_folding cleanup [pr] (#10243) 
remove type ignore and one walrus
 2025-05-09 22:42:01 -04:00
Adam Van Ymeren
a28ca0680f update dead link (#10242) 2025-05-09 19:59:52 -04:00
nimlgen
2145bce3f9 usbgpu: copyin size is 16k (#10240) 
* usbgpu: copyin size is 16k

* ush
 2025-05-09 22:12:54 +03:00
Sieds Lykles
74e40aafa0 use cdiv in div and mod folding (#10216) 
* use cdiv

* use cdiv and cmod there as well

* Add tests

---------

Co-authored-by: chenyu <chenyu@fastmail.com>
 2025-05-09 12:37:24 -04:00
Sieds Lykles
8da9c070ca take gcd out of trunc div (#10238) 2025-05-09 12:08:10 -04:00
qazal
e2292f6663 TRACEMETA>=2 displays UOp metadata in VIZ (#10237) 2025-05-09 17:42:00 +03:00
qazal
d5686f33a9 delete KernelContext dataclass [pr] (#10236) 2025-05-09 17:36:21 +03:00
qazal
467daf8d4c remap UOp metadata in graph_rewrite_map [pr] (#10234) 
* remap metadata in graph_rewrite_map [pr]

* fix

* merge loops

* UOp.metadata returns Metadata|None

* shorter
 2025-05-09 17:20:53 +03:00
nimlgen
4c75b124b6 usb: copy into mv is faster (#10233) 
* usb: copy into mv is faster

* missing

* bytes
 2025-05-09 14:53:36 +03:00
nimlgen
d08ce62553 hcq: do not reread signal in wait (#10232) 2025-05-09 14:38:36 +03:00
nimlgen
0464a31000 usbgpu: no overrun check needed (#10231) 2025-05-09 14:20:24 +03:00
nimlgen
116390083f nvme speed write example (#10230) 2025-05-09 14:20:01 +03:00
chenyu
9846435c2e fix test_div_numerator_negative (#10229) 
the simplification was wrong with negative const_factor
 2025-05-09 06:19:59 -04:00
chenyu
cba508c8c3 update uop symbolic tests (#10228) 
clean up TODOs and update tests
 2025-05-09 01:55:53 -04:00
chenyu
56def6c319 better bound for mod negative number (#10227) 2025-05-09 01:19:47 -04:00
chenyu
99f6d89dfb tighter idiv bound for symbolic denominator (#10226) 2025-05-08 22:38:56 -04:00
uuuvn
82a6160ff7 Detect metal paravirtualization bug via device name instead of CI (#10225) 2025-05-08 19:31:47 -07:00
Xingyu
a21369d039 Enhance tensor random functions with dtype support (#10214) 
* Enhance tensor random functions with dtype support
- Updated `aten.uniform_` and `aten.normal_` to include dtype parameter in backend.py
- Added unit tests for uniform and normal tensor generation with specific dtypes in test.py

* Refactor test name for clarity
- Renamed `test_normal_dtype` to `test_normal` in `extra/torch_backend/test.py`
- Aims to improve readability and better reflect the test's purpose
 2025-05-08 20:48:07 -04:00
qazal
b6904bbf83 Revert "split grouper into insert and finalize stages [pr] (#10222)" (#10224) 
This reverts commit 2594e4db15.
 2025-05-09 03:02:38 +03:00
qazal
2594e4db15 split grouper into insert and finalize stages [pr] (#10222) 2025-05-09 02:36:22 +03:00
George Hotz
0b7e3e86d0 single device copy [pr] (#10221) 
* single device copy [pr]

* simpler
 2025-05-08 15:23:22 -07:00
qazal
1d0f239df7 use Tensor.train() in schedule test + typo [pr] (#10220) 2025-05-08 23:46:42 +03:00
qazal
ff2aa6d0b2 buffer in create_kernel is optional [pr] (#10218) 
* buffer in create_kernel is optional [pr]

* pylint
 2025-05-08 22:35:55 +03:00
qazal
40560e77c2 minor grouper + viz fixup [pr] (#10217) 
* minor grouper + viz fixup [pr]

* gitignore mypy_cache

* reorder create_kernels

* replace with realized

* use tensor_map + viz before spec

* lint

* add that back
 2025-05-08 21:39:44 +03:00
George Hotz
0411b09763 small changes from new multi [pr] (#10213) 2025-05-08 07:04:27 -07:00
Sieds Lykles
a0580e8d3c Cleanup in div_and_mod_folding [pr] (#10178) 
* Refactor binary var simplification

* Simplify the congruence logic

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
 2025-05-08 06:25:32 -07:00
nimlgen
267ba9b592 usbgpu: better names in copy speed benchmark (#10212) 2025-05-08 16:12:37 +03:00
hooved
7b4f05fd00 Add test for correctness of Infinity in WebGPU (#10201) 
* use function for infinity instead of uniform

* test infinity math locally

* test infinity math in CI

* make pytest available to MacOS (WebGPU)

* revert to master except failing webgpu test
 2025-05-08 05:20:05 -07:00
nimlgen
e24fe1c746 usbgpu: pci cache (#10207) 2025-05-08 14:31:01 +03:00
nimlgen
7d6ed1b1e9 hotfix: mac ci (#10210) 
* fixed?

* cmnt
 2025-05-08 14:13:23 +03:00
nimlgen
ba52fce4b2 usbgpu: benchmark in ci (#10208) 
* usbgpu: benchmark

* usbgpu: benchmark
 2025-05-08 12:02:04 +03:00
qazal
d0e3449992 remove view_supported_devices, check allocator instead [pr] (#10209) 2025-05-08 11:45:02 +03:00
nimlgen
5a7f6b4d8e am: fix launch on rdna4 (#10206) 2025-05-08 09:46:12 +03:00
George Hotz
8d4c563c01 all COPY can be clone (#10205) 
* match old behavior

* simple

* it means the naive thing before the multi

* fix
 2025-05-07 20:31:39 -07:00
hooved
8e76c40aea Refactor test: Enable generality in testing UOp alu expressions (#10200) 
* use function for infinity instead of uniform

* test infinity math locally

* test infinity math in CI

* make pytest available to MacOS (WebGPU)

* revert to master except failing webgpu test

* isolate test refactor
 2025-05-07 19:39:44 -07:00