mirrors/tinygrad
2
0
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-06-24 02:14:17 +00:00
Code Issues Projects Releases Packages Wiki Activity
11,919 commits 280 branches 15 tags 325 MiB
Commit graph

11,919 commits

This branch
This branch
All branches
Author SHA1 Message Date
George Hotz
8c79751937 Q5_K 2026-01-29 17:04:03 +08:00
George Hotz
aeacd3b2fb ggml_type_13 2026-01-28 21:19:52 +08:00
George Hotz
202b74b369
 		assembly/amd: continue refactors (#14386) 
* simpler

* merge

* flat

* no ctx

* use the correct apis

* dup code

* write clean code

* remove bad helpers

* bits junk remove

* junk remove

* smem test

* fix tests

* correct fix + tests

* Fmt matters it seems

* wmma refactor

* a lil more

* kimi cleanups

* line
 2026-01-28 17:33:03 +08:00
qazal
5bffa17f82
 		llama train: better NULL=1 EMULATE=AMD_CDNA4 dev experience (#14395) 
* beam opens devices

* switch to hip renderer

* amd: true?

* llvm true is for test_autogen
 2026-01-28 17:31:22 +09:00
qazal
0294014108
 		fix bufferize cost function for multi, improve VIZ=-1 cli (#14394) 
* improve cli

* remove_bufferize change
 2026-01-28 15:53:18 +09:00
qazal
c158acea29
 		failing multi ram usage test from llama gemm (#14392) 2026-01-28 14:32:32 +09:00
Christopher Milan
067e27857e
 		nested composite actions don't work (#14393) 2026-01-28 00:13:30 -05:00
Christopher Milan
9dddf3d478
 		don't save caches for PRs, try 2 (#14391) 2026-01-27 23:30:17 -05:00
Christopher Milan
68fe5d8b36
 		Revert "don't save caches for PRs (#14389)" (#14390) 2026-01-27 23:22:26 -05:00
Christopher Milan
4ab228b498
 		don't save caches for PRs (#14389) 2026-01-27 23:21:31 -05:00
Christopher Milan
5e36482314
 		decompose long to ints where unsupported, try 2 (#14383) 2026-01-27 23:20:43 -05:00
wozeparrot
e496547720
 		llama3 gradacc (#14291) 2026-01-27 19:48:10 -08:00
George Hotz
88bc5ee212
 		assembly/amd: rename to better names (#14384) 
* assembly/amd: rename to better names

* might help fuzzing segfault

* emu2 -> emu
 2026-01-28 10:00:54 +08:00
George Hotz
065b95cfb0
 		Revert "add retry to fetch (#14370)" (#14385) 
This reverts commit dc4d7f2d55.
 2026-01-28 09:35:37 +08:00
Eitan Turok
dc4d7f2d55
 		add retry to fetch (#14370) 2026-01-27 14:04:25 -08:00
chenyu
8d1f3c8885
 		fix copysign for inf input (#14381) 
* fix copysign for inf input

* llvm olt
 2026-01-27 16:45:48 -05:00
Christopher Milan
289a3e415e
 		also skip test_nonoverlapping_shrink_assignment (#14382) 2026-01-27 16:26:26 -05:00
Christopher Milan
f34efc1ad1
 		DISABLE_FAST_IDIV actually works as a ContextVar (#14378) 2026-01-27 16:12:42 -05:00
chenyu
8c899e4aaf
 		fix copysign for -0 (#14380) 
test both x and 1/x < 0 work too. and found another big with the * 0 hack
 2026-01-27 15:44:58 -05:00
chenyu
62884585a7
 		failed test case for copysign -0.0 (#14379) 
* failed test case for copysign -0.0

* skip those
 2026-01-27 14:37:17 -05:00
nimlgen
ec1b28bc2c
 		am: exit early in case of failures (#14376) 
* am: exit early in case of failures

* sorry, pre-linter

* reset when error state
 2026-01-27 22:10:02 +03:00
chenyu
cd22ee9ed0
 		add InvalidType to ConstType [pr] (#14373) 
* add InvalidType to ConstType [pr]

TYPED=1 python test/test_tiny.py passes.
added PyConst = float|int|bool for some Tensor level input types

* hcq
 2026-01-27 14:09:34 -05:00
Christopher Milan
5b42a1357b
 		SCACHE=0 works with DEBUG (#14377) 2026-01-27 13:12:43 -05:00
chenyu
db010a31be
 		IGNORE_OOB -> CHECK_OOB [pr] (#14374) 
flip the meaning
 2026-01-27 12:20:59 -05:00
chenyu
c22667b0c4
 		also skip test_overlapping_shrink_assignment_reverse (#14375) 
crashing
 2026-01-27 12:20:39 -05:00
nimlgen
e52d58b041
 		autogen: update amd (#14372) 2026-01-27 19:53:14 +03:00
nimlgen
cbf94a0a95
 		nv: exit early in case of failures (#14363) 
* nv: exit early in case of failures

* f

* cleaner
 2026-01-27 19:16:22 +03:00
nimlgen
ec691cb299
 		am: print sq intrs (#14366) 
* am: print sq intrs

* cleaner
 2026-01-27 18:28:13 +03:00
qazal
a5f3d46423
 		hcq: do not assume kernel names are unique (#14371) 
* hcq: do not assume kernel names are unique

* colored kernel name
 2026-01-27 23:03:15 +09:00
George Hotz
e5df7e640b
 		fix branches in amd_asm_matmul (#14369) 2026-01-27 20:48:42 +08:00
George Hotz
0ced258726 HOTFIX: skip crashing assign test 2026-01-27 20:35:17 +08:00
George Hotz
131ae604de
 		force_transcendental on sqrt (#14368) 2026-01-27 20:24:41 +08:00
imaolo
14574c68fa
 		Add ContextVar to disable the scheduler cache (#14257) 
* add scheduler cache ContextVar

* test scheduler cache context var

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
 2026-01-27 19:55:29 +08:00
George Hotz
bfc88bcfb8
 		assembly/amd: emu refactors + enable PYTHON_REMU by default (#14361) 
* assembly/amd: start refactors

* cleanups

* those are global

* methods on ctx

* const cleanup

* range helper

* types and imports

* cleanups

* cleanups

* remove stale name

* fix emu2 types

* more typing

* more mypy

* cleanups

* fxns

* scc cleanup

* cleanups

* cleanups

* simpler parse_pcode

* laneid

* no defaults for pcode

* pcode is not optional

* cleanups

* functions cleanup

* splat

* expr_parser functions

* single tok

* invert global loops

* try_eat

* minor

* run parser on all

* no silent 0

* tests
 2026-01-27 17:42:24 +08:00
Christopher Milan
2e72625652
 		Revert "decompose dtypes.long to ints where unsupported (#14261)" (#14362) 2026-01-27 02:04:59 -05:00
qazal
f866b2a513
 		mfma loop in asm dsl (#14349) 
* mfma loop in asm dsl

* work
 2026-01-27 11:11:37 +09:00
Christopher Milan
0793319929
 		decompose dtypes.long to ints where unsupported (#14261) 
* add works

* use carry not overflow

* bitwise ops

* use tag instead of vec

* cleaner

* mul somewhat works

* mul actually works

* SUB and NEG work

* SHL/SHR

* ulong support

* this should work?

* oops

* fix indexing

* all ALU mostly works

* refactor

* test_dtype passing

* signed division works

* format

* clean

* some tests

* ruff
 2026-01-26 18:34:13 -05:00
wozeparrot
a987a4abc3
 		feat: llama8b dev_beam.sh (#14358) 2026-01-26 14:51:23 -08:00
Christopher Milan
c9c533fc78
 		libclang path is homebrew on macos (#14357) 
* libclang path is homebrew macos

* typo

* ugh

* typo

* regen

* no LIBCLANG_PATH
 2026-01-26 17:32:09 -05:00
chenyu
d641e63189
 		improve min/max for AND (#14356) 2026-01-26 15:44:18 -05:00
chenyu
f16372487a
 		fix assign hazard on shrink (#14355) 
* fix assign hazard on shrink

possible to have race if both assign src and dest are shrink

* test_nonoverlapping_shrink_assignment
 2026-01-26 14:46:30 -05:00
chenyu
145df879c1
 		find_permutes -> fix_assign_hazard [pr] (#14354) 
some noop tweaks and comment updates
 2026-01-26 14:05:19 -05:00
nimlgen
e152f1b0f5
 		llama: use ALL2ALL (#14353) 2026-01-26 22:01:53 +03:00
nimlgen
3f25eb3026
 		am: ih (#14346) 
* am: ih

* um

* fix

* line

* no trap and fix ring

* keep

* fix
 2026-01-26 20:11:04 +03:00
chenyu
823bc17fb5
 		failed test case for shrink overlap assigns (#14350) 
* failed test case for shrink overlap assigns

current logic can create a race resulted in wrong output

* skip for now
 2026-01-26 11:58:45 -05:00
George Hotz
204f51e739
 		assembly/amd: bug fixes for PYTHON_REMU (#14347) 
* default PYTHON_REMU to 1

* mockgpu

* less size

* normal compile path

* uniqie

* more

* fix clamp

* Change PYTHON_REMU default to 0 in _try_dlopen_remu
 2026-01-27 00:48:22 +08:00
chenyu
231305603d
 		remove REAL_DEV [pr] (#14337) 
it's just Device.DEFAULT now
 2026-01-26 10:08:16 -05:00
Martin Szewieczek
9cbe99348a
 		func meshgrid: change param index to type str (#14331) 2026-01-26 10:07:56 -05:00
George Hotz
3b43d26f10
 		assembly/amd: emu speed (#14344) 
* assembly/amd: emu speed

* fix spec

* go

* don't do this

* simpler

* no stupid consts

* hack

* simpler

* no index

* no where

* faster linearizer

* fix spec

* no index dtype
 2026-01-26 22:21:34 +08:00
George Hotz
774a454bb5
 		assembly/amd: fix scratch SVE (#14340) 
* assembly/amd: default python REMU

* mem_used

* no lane

* sve

* remove that

* needs s_code_end in tests
 2026-01-26 21:03:51 +08:00