Commit Graph

Select branches

Hide Pull Requests

abstractions_4

add_local_buffer

after_qol

algebraic_upat

all_bv_no_idx

all_sink

amd_isel

amd_refac

amd_smu_reset_patch

amd_sqtt

amd_uop

amd_uop_cleanups

assign_copy

assign_try_3

before2

better_spec

better_syn

bitcast_spec

broadcast_shape_2

bt_ei

buf_heur

bump_nv_2404

cache_reshape

call_inline

call_is_kernels

callify

cast_vec_as_cast

cat_mop

changes_from_rangify

check_clsetarg

check_dims

cifar_30

cifar_30_clean

clean_load

clean_spec

clone_tg

codegen_try_2

contig_on_after

continue_mat

deepsek

default_late_allreduce

delete_kernel_py

delete_ones

delete_slow_rangeify

delta_attention

delta_attention_scan

demote_op

disable_rules

dsp_search

dsp_search_merged

dtype_shape

earlier_gater

emu_cdna

emulate_ctx

end_all_ranges

endrange

er_prioity

execution_unit

failing_assign

fancy_usb

fast_llm

fix_fa_fwd_clang_22

fix_fuse

fix_llvm_cast

fix_range_merging_sd

fix_rangeify_tests

fix_rc

fix_sched_copy

fix_slice_store

fix_src_spec

fixup_fix_kernel

full_spec

fun_w_egraphs

gate-on-load-store

gate_rewrite

gemma4_gpt

gen_pdf_fast

gh-pages

gl_dims

gldims

glm_flash

grad_outerworld

grok_x86_novec

group_div_rules

hcq2_vram_copies

heur_4096

heur_group

hk_gemm

http_threading

hwtest_fixes

id_store

if_range_cleanups

ignore_oob

image_idx_xy

image_in_mc

image_no_vec

image_refactors_stack

index_to_slice

invalid_try_3

jinja_template

kernel4

kernel_is_call

kernelless

kimi_slop

kitten_matmul

late_Locals

late_add_load

late_allreduce

late_codegen

late_decomp_is_late

late_lower_idx

late_view

late_weakint

lds_double_buffering

less_wmma_arg

lil_dg_change

lil_rangeify_td

llama_trainer

llm_262k_minimal

llm_flash_attn

llm_qwen

llm_serve_fixes

load_view_late

local_cache

locals_have_device

loop_end_op

loop_is_rng

lowerer_hacks

mac_pytest

manual_kernel_3

master

mega_lowerer

min_new_lin

min_outer_range

minor_views

mixins7

mlperf_training_v4.1

mlperf_training_v5.0

mlperf_training_v6.0

modern_scan

more_apps

more_early_comps

more_folding_rules

more_kernelcount

more_pcode

more_rangeify_pm_tests

more_ren_clean

more_uop_gc_test

more_uop_mm

more_vmap

move_devec

move_dtype_decomps

move_expander

move_gates_to_load_store

move_mop_cleanup

move_schedule

move_simp_views

moveleftright

mt_typing

multi_range

multioutput

new_ck

new_devec

new_invalid_and_image

new_linearizer

new_shape

new_test

no_assign_2

no_buf_0

no_decimals

no_late_simplify_marg

no_merge_ends

no_merge_views

no_mstack_in_allreduce

no_one_ranges

no_sip_usbgpu

no_tsink_base

no_tuplize

no_user_ck_contig

no_vec_x86_4

no_weakint_variable

noop_index

novecimg

num_batches_tracked_shape

olmo3_support

one_compiler

one_hcopt

only_assign_buffers

only_loop_backends

only_reg_emu

only_reg_emu2

outer_range

outerworld_work

pad_invalid_work

pad_mixin

param_call_is_function

parse_sqtt_more

pci_iommu

port_nir

postrange_hcopts

precompile_backward

prefill_qwen_fast

print_special_ops

ps_clean

pyremu_try

python_prg

python_speed

q4k_support

qcom-openpilot-13ms

qcom_830

qcom_mmapeak

qcom_speed_codex

qol_string

qwen36_27b_amd_900

qwen36_27b_amd_minimal

qwen36_amd_bugfixes

qwen_mergable

r_cleanups

r_new_opt

ranged_call

rangeify_2

rangeify_const_folding

rdna3_vibes

rdna4_asm

rdna4_gemm

real_pyrender

recurse_sub

reg_coalese

relu_pattern

remove_assign

remove_cast_hack

remove_ck_contig

remove_define_var

remove_forced_re

remove_on_stack

remove_programspec

remove_pylint

remove_slow_tests

remove_vec_2

remove_vec_3

renderers_use_shapes_2

replace_if_with_range

reprocess_node_experiments

reshape_trait

resnet-ci

revert-11547-fix_mismatch_red

revert-11715-revert-11711-amd_llvm_tc_cast

revert-11809-no_ki_in_gpudims

revert-11811-better_div_nesting

revert-11876-memplan_vbig

revert-12104-simplify_valid_in_symbolic

revert-12308-bufferize_early

revert-12731-jitlink_init_check

revert-12813-after_cleanups

revert-13046-better_cleanup_arange_buffers

revert-13311-prof_link_sqtt

revert-13583-rocr_desc_2

revert-14734-tload

revert-17141-full_rewrite_to_sink

revert-17408-_broadcasted-weak

revert-9438-simple_matmul_np_rand

revert_wrong_ck_fix

rewrite_rangeify2

rtoposort

scan_assign

scan_op_8

sched_cache

sched_lin

sd_load_simple

self_type

set_testpath

shrink_in_render

simp_spec

simple_gidxs

simple_pool

simple_priority

simple_rangeify

simpler-sd-compile

simpler_end

simpler_fusion

simpler_image

simpler_postrange

simpler_reduce

slice_like_shrink

spec_3

spec_cleanup

spec_ren

speed_one_linear

split_buf_idx

split_decomp_pass

split_sub

stack_0

store_is_void

string_ranges

support_opts_in_contig

swizzle_in_kernel

sym_llm

sym_work

symbolic-chunk-split

szdiff_bot_test

tc_is_heur

tensor_graph_single_sink

test_amd_crash

test_fa

test_from_postopt

test_hgemm

test_unit

tiny_scan

tmr

tool_call_support

topoprop

ttomsa-new_linearizer

tuplegtuple

two_stage_remove

typed_checks

typing_work

unused_permute_arg

uop_prg

upd_test4

update_benchmark

update_benchmark_search

update_mlperf

user_failure

vecless_x86_glm

verify_sink_dims_clean

vf_ish

view_after_const

view_in_codegen

viz_cbroadcast

warp_fun

wgpu-f16-packed

working_postopt

x86_moves

x86_numel

v0.10.0

v0.10.1

v0.10.2

v0.10.3

v0.11.0

v0.12.0

v0.13.0

v0.4.0

v0.5.0

v0.6.0

v0.7.0

v0.8.0

v0.9.0

v0.9.1

v0.9.2

Mono Color

• fe8c05b96f allow disabling method cache geohot 2023-03-11 08:57:49 -08:00
• 5e1380df6a profiling llama + cache is_contiguous geohot 2023-03-11 08:23:21 -08:00
• 01f39b19dc move to shapetracker.py geohot 2023-03-11 07:49:20 -08:00
• f3ac52aee8 Mypyc (#680) George HotzandGitHub 2023-03-11 07:33:30 -08:00
• 22905dd657 speedups from llama branch geohot 2023-03-10 22:01:32 -08:00
• 0b03216cc3 losing lines (#678) George HotzandGitHub 2023-03-10 21:57:05 -08:00
• d7cb8e3e56 multithreaded fake_torch_load_zipped geohot 2023-03-10 19:16:27 -08:00
• b1206bcb18 third try at torch loading (#677) George HotzandGitHub 2023-03-10 19:11:29 -08:00
• 8b7a16cf85 Add conv binops_no_rerun test assertions (#665) Connor HendersonandGitHub 2023-03-10 22:09:48 -05:00
• 8bf75a7fdd fix stable diffusion and CI geohot 2023-03-10 17:48:12 -08:00
• c7d17c25d9 ugh, that's getting ugle geohot 2023-03-10 17:41:19 -08:00
• 4780f9a6df llama runs (slowly) in master geohot 2023-03-10 17:36:51 -08:00
• 1826ff6b89 dtypes nice and clean (#673) George HotzandGitHub 2023-03-10 16:56:07 -08:00
• d26345595d more llama stuff geohot 2023-03-10 10:48:10 -08:00
• 442e1bcd5a typo + EARLY_STOPPING geohot 2023-03-10 10:43:07 -08:00
• 6142e63a3e touchups, print GB/s geohot 2023-03-10 10:37:37 -08:00
• 036737a12a mem_estimate tracks bytes, not items geohot 2023-03-10 09:44:12 -08:00
• 1a039306d2 good changes from llama branch (#671) George HotzandGitHub 2023-03-09 20:51:22 -08:00
• de1b6d3e08 check shrink is actually smaller geohot 2023-03-09 12:59:45 -08:00
• dbbaa0bdd7 int32, and refactor pad/shrink geohot 2023-03-09 12:57:17 -08:00
• fb5ee9260f add pad tests to shapetracker geohot 2023-03-09 12:51:18 -08:00
• da7fb4b227 Fixed DDPG example (#667) jspielerandGitHub 2023-03-09 20:49:52 +01:00
• 022c5835fc fix GPU import error and old python Tuple geohot 2023-03-08 12:22:11 -08:00
• c22afc52db move the custom function example to a test geohot 2023-03-08 10:05:04 -08:00
• 7d3b9d0e95 oops, things relied on that API. the global cache needs access to the ASTRunner class geohot 2023-03-08 08:39:31 -08:00
• 4f957423c3 jitting custom ops + OPTLOCAL assignment bugfix geohot 2023-03-08 08:30:37 -08:00
• 7285de41a1 tinygrad supports CUSTOM functions geohot 2023-03-08 07:50:33 -08:00
• 00641aa45d add challenge tests George Hotz 2023-03-07 19:39:04 -08:00
• e0244baf60 3 letters for graph op George Hotz 2023-03-07 19:20:48 -08:00
• 46df02115d bring back SHUFFLE_PAD_OPS as OPT>=4 George Hotz 2023-03-07 17:42:34 -08:00
• 4eb880550f enable contract test George Hotz 2023-03-07 17:32:28 -08:00
• d885d2d0f5 Allow 1s for contraction detection (#663) Alex WangandGitHub 2023-03-08 09:31:28 +08:00
• b561256a0e allow all reduces (#661) George HotzandGitHub 2023-03-07 15:36:01 -08:00
• b14d31d6db ConvNeXt + extras (#657) George HotzandGitHub 2023-03-06 22:10:56 -08:00
• d8dda2af3a openpilot fixups v0.5.0 George Hotz 2023-03-06 14:14:44 -08:00
• 4b9bc1615b While fusion (#654) George HotzandGitHub 2023-03-06 09:13:23 -08:00
• 066a65dad5 remove tflop number, i'll never update that when it's fast geohot 2023-03-06 08:30:31 -08:00
• 6e763dc446 matmul example in readme geohot 2023-03-06 08:25:13 -08:00
• 5dc227dba6 fix bug in ENABLE_METHOD_CACHE and enable for llvm geohot 2023-03-06 07:43:40 -08:00
• 8c5dea8d72 fix CUDA float4 issues geohot 2023-03-06 07:16:38 -08:00
• 7dbcc26582 fix up external tests geohot 2023-03-06 06:52:28 -08:00
• 50012f679b move get_contraction to shapetracker geohot 2023-03-06 06:42:57 -08:00
• 64ecbd91b5 Refactor contraction and add integration test cases for push permute (#650) Alex WangandGitHub 2023-03-06 22:36:55 +08:00
• cb5be9697c One less line in consume_flops (#651) Peter McDevittandGitHub 2023-03-06 00:34:45 -07:00
• f028accae3 print upcasted colors geohot 2023-03-05 22:12:42 -08:00
• b2d4b2c06e refactor imports and EOL whitespace geohot 2023-03-05 20:52:03 -08:00
• 382f346523 clean up opt (#649) George HotzandGitHub 2023-03-05 20:49:36 -08:00
• 7930c6ab5c CLImage backing bug + test_vec_mul geohot 2023-03-05 16:32:05 -08:00
• 8de24e3b05 accumulator can be a float4 (#647) George HotzandGitHub 2023-03-05 15:44:41 -08:00
• c10131ddf5 reduce number of lines (#645) Cyril RoumégousandGitHub 2023-03-06 00:42:32 +01:00
• 7989f79820 using image from mad branch saves 1ms on op model geohot 2023-03-05 14:38:42 -08:00
• 7940ad258e fix dropout test geohot 2023-03-05 12:24:04 -08:00
• 3072e098c0 local workgroup optimizer geohot 2023-03-05 12:08:12 -08:00
• b1ba78ac38 move applegpu disassembler geohot 2023-03-05 11:21:12 -08:00
• e8de3f5736 Revert "less lines (#643)" (#644) George HotzandGitHub 2023-03-05 08:41:11 -08:00
• 30f2238994 less lines (#643) Peter McDevittandGitHub 2023-03-05 09:37:14 -07:00
• 3da56ab41d adreno disassembler Comma Device 2023-03-05 10:32:03 -06:00
• 16b03f3c3b wow, can't believe that was broken (#642) George HotzandGitHub 2023-03-04 22:28:28 -08:00
• 4a607f7d65 more ext gpu tests geohot 2023-03-04 21:00:08 -08:00
• 69198a73d2 test_1x1_24_6 geohot 2023-03-04 20:37:46 -08:00
• f281f707bd better function names geohot 2023-03-04 18:27:37 -08:00
• a77d792aff Codegen gpu cleanups (#640) George HotzandGitHub 2023-03-04 15:31:51 -08:00
• 10d40d3cf2 Expand the inline loop to prevent stack overflow from _deepwalk (#638) Patrick GenevaandGitHub 2023-03-04 18:14:17 -05:00
• b02a392d69 Improve local (#635) George HotzandGitHub 2023-03-04 09:30:49 -08:00
• 117111825c Fix windows file permission error (#634) Patrick GenevaandGitHub 2023-03-04 12:23:55 -05:00
• 528cb3b3b9 fix ast test geohot 2023-03-04 07:49:25 -08:00
• 8bc9277587 G.nodes isn't always valid geohot 2023-03-04 07:24:43 -08:00
• 85f69b5489 metal needs the Cocoa geohot 2023-03-03 23:22:15 -08:00
• 28a6ada4ce line reduction in metal geohot 2023-03-03 23:14:40 -08:00
• 893f136fe0 lines from helpers geohot 2023-03-03 23:07:46 -08:00
• 81cda2b672 zero out s == 1 strides geohot 2023-03-03 22:57:02 -08:00
• aef336c079 merge_views is very powerful geohot 2023-03-03 22:53:59 -08:00
• b5b4edf59b comments geohot 2023-03-03 22:39:31 -08:00
• cfb050e2d1 simple modrange, thanks Jacky geohot 2023-03-03 22:37:04 -08:00
• 3dab721f9f lazy cleanup geohot 2023-03-03 22:01:03 -08:00
• c53efb3635 optimize for CL (#633) George HotzandGitHub 2023-03-03 22:00:09 -08:00
• 9d97d97b26 Opens image in default viewer after saving. (#612) Pankaj DohareyandGitHub 2023-03-04 06:58:49 +05:30
• 1a84976d4d fix thneed gflops geohot 2023-03-03 16:52:59 -08:00
• 7a1d96fd76 No negative (#632) George HotzandGitHub 2023-03-03 16:48:14 -08:00
• 8c475ea86a relax atol, merge_view geohot 2023-03-03 07:48:44 -08:00
• b9ce20c374 openpilot test wasn't running, factor out image idx geohot 2023-03-03 07:41:53 -08:00
• 9bd2cdee08 skip broken bn training test for speed geohot 2023-03-03 06:52:11 -08:00
• 999b44c274 fix external test + speed geohot 2023-03-03 06:46:16 -08:00
• 8919ca8163 test cleanups geohot 2023-03-03 06:36:06 -08:00
• 459488bba2 fix linter (#630) George HotzandGitHub 2023-03-02 20:06:20 -08:00
• 3915c89fb6 symbolic improvements (#629) George HotzandGitHub 2023-03-02 19:50:38 -08:00
• b842cdf11f support wait in cuda George Hotz 2023-03-02 10:39:26 -08:00
• dc88ad3342 fix ops print bug George Hotz 2023-03-02 10:33:03 -08:00
• 0335cb86b9 refactor comparison. there's a bug in the method cache George Hotz 2023-03-02 10:10:16 -08:00
• 11f257ddf9 clean up cmp tests George Hotz 2023-03-02 09:35:16 -08:00
• 8902764167 fit nits in compare geohot 2023-03-02 08:15:26 -08:00
• 52204a7b88 adding comparison operators (#616) DiogoandGitHub 2023-03-02 11:10:44 -05:00
• 2e26286294 speed like you wouldn't believe (#626) George HotzandGitHub 2023-03-02 07:49:19 -08:00
• 51fb6aeb45 Fix cuda runtime (#625) Martin LoretzandGitHub 2023-03-02 15:52:34 +01:00
• fca055bd66 NOOP means contiguous geohot 2023-03-01 21:54:51 -08:00
• d062cc82b8 put restrict back geohot 2023-03-01 21:34:45 -08:00
• 201d9a2d58 remove extra copy on output geohot 2023-03-01 21:19:44 -08:00
• b442e75c7a test speed v torch geohot 2023-03-01 19:50:12 -08:00
• bfcec234a2 Refactor ASTs (#622) George HotzandGitHub 2023-03-01 18:57:29 -08:00
• 5e41d5857c Add tests for randomness (#621) Jacky LeeandGitHub 2023-03-01 15:39:20 -08:00