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

• 4c2faa4140 functools.partial keeps mypy compiler working geohot 2023-02-08 18:04:32 -06:00
• cfd13c083b refactor GenericShape for a big line reduction geohot 2023-02-08 18:01:08 -06:00
• c656513591 GPURunner class will replace CL cache eventually geohot 2023-02-08 17:31:36 -06:00
• a5a55ac19e GlobalCounters cache + assign in optim geohot 2023-02-08 17:10:55 -06:00
• d9555bc478 that turned out to be dumb geohot 2023-02-08 16:52:29 -06:00
• 3d63934995 refactor to keep cl in the runtime (#545) George HotzandGitHub 2023-02-08 16:46:09 -06:00
• 8c8a5a77dd refactor llvm into runtime and ops geohot 2023-02-08 16:28:32 -06:00
• 45ce4de6f3 improve typing geohot 2023-02-08 12:48:21 -06:00
• 2e1bdc889a write out all the functions, no auto binding (#543) George HotzandGitHub 2023-02-08 12:41:39 -06:00
• d854337f0d nn/optim.py compiles now geohot 2023-02-08 11:25:18 -06:00
• 1029deccb1 refactor ops_cpu and ops_torch to not share code geohot 2023-02-08 11:11:42 -06:00
• ee18420c13 dyn add of math ops geohot 2023-02-08 10:04:30 -06:00
• 2844482a60 Mypy fun (#541) George HotzandGitHub 2023-02-08 09:56:51 -06:00
• 996e0a10b7 update cpu and torch to hold buffers (#542) George HotzandGitHub 2023-02-08 09:40:45 -06:00
• ae4f0aeb5f NumPy-like semantics for Tensor.__getitem__ (#506) Mitchell GoffandGitHub 2023-02-08 06:59:46 -08:00
• 0ac3286af0 factor out Device geohot 2023-02-07 16:08:20 -06:00
• 2aeebd70a6 mypy will compile the shapetracker, no speed up geohot 2023-02-07 15:43:44 -06:00
• 185d2e3678 fix map_buffer and add some __slots__ geohot 2023-02-07 15:32:48 -06:00
• aebe75d9a2 remove val expansion (#539) George HotzandGitHub 2023-02-07 15:14:05 -06:00
• 001cc96e25 Lazy refactor (#538) George HotzandGitHub 2023-02-07 11:53:21 -06:00
• 02d8cb0959 lazy cleanup geohot 2023-02-07 07:39:53 -06:00
• d93563f39f fix KOPT geohot 2023-02-07 06:56:12 -06:00
• 7604b17fbf TestZeroViewShapeTracker fix test (#481) Jared ZandGitHub 2023-02-07 06:17:55 -06:00
• c073271f20 more symbolic correctness geohot 2023-02-07 00:03:14 -06:00
• e961fd3a04 more symbolic test, ModNode is wrong geohot 2023-02-06 23:43:21 -06:00
• 8cfeb118d6 symbolic new test geohot 2023-02-06 23:27:26 -06:00
• 7c5a5ecdac even simpler symbolic geohot 2023-02-06 22:47:00 -06:00
• 8b05de1841 symbolic cleanups geohot 2023-02-06 22:12:11 -06:00
• 2a924e2b77 fix sz.sh for llvm geohot 2023-02-06 15:36:05 -06:00
• 0d405fd5bc Parallelize CI tests (#535) James RobertsandGitHub 2023-02-06 23:27:44 +02:00
• 4977d6f225 using tuples in isinstance (#534) AndreyandGitHub 2023-02-06 17:40:26 -03:00
• d56c57b112 adding more robust install method (#532) timmermansjoyandGitHub 2023-02-06 20:12:05 +01:00
• fd3807c479 delete cherry and old cuda accel, promote llvm geohot 2023-02-06 10:02:41 -06:00
• 90529d3750 tests are 20% faster (#529) George HotzandGitHub 2023-02-06 07:56:14 -08:00
• 039de1b332 oops, pytest is for testing geohot 2023-02-06 09:30:12 -06:00
• 6eb0e6a650 shuffle deps: always tqdm, make linting category geohot 2023-02-06 09:27:01 -06:00
• 1d80639646 make linter test install testing deps geohot 2023-02-06 09:21:48 -06:00
• 60bb64811c merge mypy into linters, no useless package update geohot 2023-02-06 09:14:00 -06:00
• c3d81bba2a test_train: Adam -> SGD geohot 2023-02-06 08:55:41 -06:00
• 36c26a57b1 make slow LLVM opt optional geohot 2023-02-05 20:20:38 -06:00
• f7291f6ca3 fixes big KOPT, breaks opencl (#505) George HotzandGitHub 2023-02-05 10:46:17 -08:00
• 97f0a82be7 Cache pip packages in github actions (#522) Martin LoretzandGitHub 2023-02-04 05:04:20 +01:00
• 4ad67b4bbc Refactor triton buffer to use CLBuffer of cuda runtime (#524) Martin LoretzandGitHub 2023-02-04 05:02:41 +01:00
• ad4f6aa2cf Add test for quick_gelu (#526) Jacky LeeandGitHub 2023-02-03 20:01:39 -08:00
• db0a9b0a2d Refactor CL.time_sum into GlobalCounters (#519) James RobertsandGitHub 2023-02-02 06:13:56 +02:00
• 45e847d284 Update triton to work in master (#517) Martin LoretzandGitHub 2023-02-01 21:58:14 +01:00
• 5e37f084db stable diffusion: clean up constant folding geohot 2023-02-01 12:53:16 -08:00
• 175c38d1b3 triton: it already was GT0 geohot 2023-02-01 12:00:33 -08:00
• 486f023e81 Rename Normalize and move to nn (#513) Jacky LeeandGitHub 2023-02-01 11:55:03 -08:00
• cd97b036cc A Triton backend for tinygrad (#470) 3 people 2023-02-01 11:53:57 -08:00
• 4e24002bbe no generic exceptions geohot 2023-02-01 11:14:37 -08:00
• 54c68defc7 Replace SIGN with GT0 (#511) 3 people 2023-02-01 11:01:39 -08:00
• 799b3f185a Refactor getenv into helpers (#508) Jacky LeeandGitHub 2023-01-31 15:09:09 -08:00
• d91b6711ea oops, broke BN geohot 2023-01-31 08:18:48 -08:00
• 21f2af08d5 getenv + graphing geohot 2023-01-30 19:15:03 -08:00
• 491e78d203 Add symbolic tests for correctness (#494) Jacky LeeandGitHub 2023-01-30 18:40:16 -08:00
• 60ccddb58b reenable SWAP geohot 2023-01-30 17:31:52 -08:00
• c1a769b68b fix bug in gpu copy out geohot 2023-01-30 16:51:28 -08:00
• e87410c531 fix multiple accumulators geohot 2023-01-30 16:22:26 -08:00
• aea55eb196 found failing upcast geohot 2023-01-30 16:12:56 -08:00
• b67f997864 tests pass w/o float4 geohot 2023-01-30 15:40:49 -08:00
• c6f570a2e6 improve progress bar geohot 2023-01-30 14:50:28 -08:00
• 4685c9c095 Big changes (#498) Kevin GilpinandGitHub 2023-01-30 14:42:22 -08:00
• 7118602c97 goat progress bar geohot 2023-01-30 14:37:26 -08:00
• 7ee0d99c70 CLCACHE geohot 2023-01-30 14:02:06 -08:00
• 7457f0d755 KOPT=2 geohot 2023-01-30 13:28:06 -08:00
• cccfea4b25 factor out KOPT code geohot 2023-01-30 13:13:55 -08:00
• de2c419fd4 make_pair and first attempt at hlb_cifar10 geohot 2023-01-30 11:07:23 -08:00
• 7b6b1f32b1 [Fix] fix typo: test_mnist -> datasets (#492) AllentDanandGitHub 2023-01-30 13:30:47 +08:00
• 2db272c7f7 Kernel Optimizer (#489) George HotzandGitHub 2023-01-29 17:15:00 -08:00
• 43abbd3d00 Use force_create to allocate return buffer (#491) Martin LoretzandGitHub 2023-01-30 02:13:10 +01:00
• bb0cdc2442 111.51x speedup for reduce geohot 2023-01-29 03:06:00 -08:00
• 45c0aa6e2d search with SHIFT, REDUCE geohot 2023-01-29 02:42:20 -08:00
• 87879cf4b6 improve search more geohot 2023-01-29 02:08:57 -08:00
• f6bbd43cb8 improve search geohot 2023-01-29 01:33:47 -08:00
• ebdec2b72f fix optimizer geohot 2023-01-29 00:23:06 -08:00
• a9cabce791 oops, broke mem estimates geohot 2023-01-28 20:21:31 -08:00
• a500e79bd1 don't OPTWG on OS X, it's way slower geohot 2023-01-28 20:02:33 -08:00
• b0df4d99a0 os x profiling: this ratio is exact i believe geohot 2023-01-28 19:02:51 -08:00
• c0963b723e should fix tests geohot 2023-01-28 15:13:03 -08:00
• b134a4f3d1 don't upcast already upcasted geohot 2023-01-28 14:58:28 -08:00
• 2f194aadad loop unrolling upcast geohot 2023-01-28 14:51:24 -08:00
• 381f3e92da fix prints, add third conv geohot 2023-01-28 14:10:27 -08:00
• 92001a06e1 openpilot/go.sh geohot 2023-01-28 13:57:43 -08:00
• aea29f8a6e fix CUDA reduce geohot 2023-01-28 13:38:58 -08:00
• 0f34c24aeb move expr_idxs to shapetracker geohot 2023-01-28 12:25:05 -08:00
• f2e81f7208 line reduction and cleanups geohot 2023-01-28 12:17:40 -08:00
• 03dd1201dc local buffer implied geohot 2023-01-28 12:06:28 -08:00
• b3e4e678e8 Use ShapeTracker for tracking shapes in kernels (#485) George HotzandGitHub 2023-01-28 11:56:32 -08:00
• 259c48f235 discord image is invite link geohot 2023-01-28 11:42:11 -08:00
• d748000ada tinygrad discord geohot 2023-01-28 11:36:15 -08:00
• ae810eb558 minor cleanups geohot 2023-01-28 08:59:15 -08:00
• 713318745d padding size in get_conv_args geohot 2023-01-28 08:47:18 -08:00
• 299d1cdc9c lil cleanup of load ldr geohot 2023-01-28 00:31:57 -08:00
• 2b5bc5d4a1 factor out image_idx geohot 2023-01-28 00:22:54 -08:00
• bd8a5c2ced Simple CUDA Runtime (#480) George HotzandGitHub 2023-01-27 16:26:24 -08:00
• 6d5e1a8029 GEMM kernel search geohot 2023-01-27 10:08:57 -08:00
• 123993156d refactor group_for_reduce a little geohot 2023-01-27 08:51:23 -08:00
• 82e58108e3 add flake8 to precommit geohot 2023-01-26 22:31:45 -08:00
• f4b571039b fix shape types geohot 2023-01-26 22:29:20 -08:00