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

autogen-comgr-2

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

ci-deps

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

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

explicit-dev

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

gitea-test-ci

gl_dims

gldims

glm_flash

grad_outerworld

grok_x86_novec

group_div_rules

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-chat-template

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

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

onnx-cpu-count

outer_range

outerworld_work

pad_invalid

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

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-lp

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

view_after_const

view_in_codegen

viz_cbroadcast

warp_fun

wgpu-f16-packed

working_postopt

x86_moves

x86_numel

#1

#2

#3

#4

#5

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

• db5a632e8c multicat + test onnx is generic onnx geohot 2022-06-11 11:50:47 -07:00
• a710b3a210 it's a real test now geohot 2022-06-11 11:33:33 -07:00
• 8440dbfa5d support inputs geohot 2022-06-11 11:21:45 -07:00
• 08de1aa636 add flatten to tinygrad geohot 2022-06-11 11:15:16 -07:00
• aee251cc41 op model test geohot 2022-06-11 11:06:03 -07:00
• d061ce8d5e add ELU support geohot 2022-06-11 10:47:23 -07:00
• 50ba554a14 Ops layer of indirection geohot 2022-06-11 10:04:45 -07:00
• 1d29780b75 use conv2d for transpose when able geohot 2022-06-11 09:32:13 -07:00
• 861323c121 cleanup convdw torch geohot 2022-06-11 08:32:57 -07:00
• 041ad55ed1 clean up torch processing op geohot 2022-06-11 08:27:09 -07:00
• fc7eabb86f processing op geohot 2022-06-11 08:12:02 -07:00
• 72186ebd5a movement ops, reshape is a copy now geohot 2022-06-10 20:01:47 -07:00
• e5d694490f reshape makes a copy geohot 2022-06-10 19:49:04 -07:00
• c8bacd0d8e rename transpose to permute geohot 2022-06-10 19:41:50 -07:00
• 462f1ce0da Remove Matmul (#323) George HotzandGitHub 2022-06-10 19:26:23 -07:00
• 66cb86c9ec fix strided conv George Hotz 2022-06-10 15:17:53 -07:00
• 8864b37333 fix torch convdw George Hotz 2022-06-10 15:04:39 -07:00
• aac1a9b419 this breaks tests George Hotz 2022-06-10 12:20:42 -07:00
• c4238187be oops, we should have a linter in a precommit hook geohot 2022-06-10 10:35:50 -07:00
• 160f651c25 conv_args not in llops interface geohot 2022-06-10 10:35:10 -07:00
• 8c084b8c12 dispatcher: now this is nice geohot 2022-06-09 10:28:46 -07:00
• 70bb3a7976 remove more weird logic in the dispatcher geohot 2022-06-09 10:06:01 -07:00
• e58d344759 simpler dispatch logic since the mlops are universal now geohot 2022-06-09 10:01:10 -07:00
• 273af8d732 remove unused CPUBuffer functions geohot 2022-06-09 09:45:39 -07:00
• 9d85067ad8 no splatting conv_args geohot 2022-06-09 09:40:58 -07:00
• 259b536e3a clean up reduce a little geohot 2022-06-09 09:34:19 -07:00
• d3e7238bdd remove dead code geohot 2022-06-09 09:25:40 -07:00
• 40f8eb7383 namedtuple for conv_args geohot 2022-06-09 09:19:52 -07:00
• d0c3204996 get_conv_args geohot 2022-06-09 09:10:40 -07:00
• 6dcaee2361 ops enum cleanup geohot 2022-06-09 08:42:42 -07:00
• e01ed64d7c restore that naming geohot 2022-06-09 08:38:34 -07:00
• e7ebbecf97 proper llops geohot 2022-06-09 08:33:25 -07:00
• d841fc4392 stupid, but the tests should pass geohot 2022-06-08 23:43:10 -07:00
• 214fb8c974 ctx.op geohot 2022-06-08 23:39:54 -07:00
• 5a533fc073 llops torch test passes geohot 2022-06-08 23:30:56 -07:00
• 1e3db466cc torch llops geohot 2022-06-08 23:15:03 -07:00
• 30ab2249eb match order geohot 2022-06-08 11:46:51 -07:00
• 4a9882d495 hlops geohot 2022-06-08 11:46:09 -07:00
• e046a2fd9f readme fix typos geohot 2022-06-08 11:43:05 -07:00
• 4b09ca90a1 readme: still WIP geohot 2022-06-08 11:41:19 -07:00
• 91d97c8313 mlops for cpu geohot 2022-06-08 11:20:10 -07:00
• 1764a36bd6 llops cpu passes geohot 2022-06-08 10:57:45 -07:00
• 81d16d105e llcpu convs work geohot 2022-06-08 10:51:09 -07:00
• 6bdcf5ef59 llops cpu geohot 2022-06-08 10:43:02 -07:00
• 582ddef59c 14 llops too geohot 2022-06-08 10:01:10 -07:00
• 60a48455ad still over line count, maybe test pass geohot 2022-06-08 09:51:28 -07:00
• 70561f3d90 way over the line limit geohot 2022-06-08 09:36:31 -07:00
• 4f7ee235c5 not a real test now geohot 2022-06-08 09:00:59 -07:00
• ae33060dae early float4 stuff for binary geohot 2022-06-08 08:59:54 -07:00
• 82f29b5dbf better GPU block geohot 2022-06-08 08:01:04 -07:00
• 42ae78241e only run test on GPU geohot 2022-06-08 07:54:40 -07:00
• cdf4b5f142 opencl perf test geohot 2022-06-08 07:49:08 -07:00
• 892ac661e1 enet readability geohot 2022-06-07 10:23:05 -07:00
• 9b27ba650b load new torch files geohot 2022-06-07 10:06:48 -07:00
• 33c9c6c025 float4 for unary_op geohot 2022-06-06 10:40:30 -07:00
• 69a401d596 use cl.enqueue_copy and add padding to GPUBuffers geohot 2022-06-06 10:27:08 -07:00
• d8ee8a39ac sgd threestep graph is so pretty geohot 2022-06-06 09:45:37 -07:00
• aa3fb10a92 hmm, not allowed geohot 2022-06-06 09:40:15 -07:00
• c143c92828 adam threestep geohot 2022-06-06 09:38:28 -07:00
• c26ed64cef slightly cleaner graph geohot 2022-06-06 09:31:48 -07:00
• d302049e53 don't use div geohot 2022-06-06 09:25:31 -07:00
• 7229d65235 shortcut nops geohot 2022-06-06 09:15:41 -07:00
• b482e1db85 minor cleanups, one less buffer geohot 2022-06-06 08:54:24 -07:00
• 1ec786960f fix ci with f geohot 2022-06-06 08:34:05 -07:00
• f177cd550f make binary ops respect needs_input_grad geohot 2022-06-06 08:29:08 -07:00
• a1dff4061b minor cleanups geohot 2022-06-06 08:14:52 -07:00
• 613f0ca6e5 remove new_buffer geohot 2022-06-06 07:57:39 -07:00
• 30f55eaaba rename llops to opencl geohot 2022-06-06 07:49:36 -07:00
• 233c71a7ba support requires_grad geohot 2022-06-06 07:47:31 -07:00
• 9f9cf076c0 comments geohot 2022-06-05 21:51:31 -07:00
• a7ea00dae7 remove useless geohot 2022-06-05 21:37:12 -07:00
• 5526f36f44 all buffer alloc in top level geohot 2022-06-05 21:34:34 -07:00
• 85e130a0c7 line count geohot 2022-06-05 21:12:02 -07:00
• d9e57954d8 factor out unary and binary op allocations geohot 2022-06-05 20:48:36 -07:00
• bb7e7b48f9 output edges aren't intermediates either geohot 2022-06-05 17:27:32 -07:00
• 927a238dc8 better printing for intermediate tensors geohot 2022-06-05 17:19:47 -07:00
• 3dac8fa728 this fix the gc geohot 2022-06-05 17:16:40 -07:00
• 0ee21ba115 add ViT test and car geohot 2022-06-05 17:12:43 -07:00
• ebd7290f56 simpler batchnorm geohot 2022-06-05 16:51:51 -07:00
• 30025f2bc7 bn: don't track_running_stats if you aren't training geohot 2022-06-05 16:41:20 -07:00
• 1de75b67d5 fix bug in graph with use of id geohot 2022-06-05 16:31:20 -07:00
• c193088c97 fix bug in purple edges geohot 2022-06-05 16:04:51 -07:00
• 81c9438ea1 keepdim avoids reshapes geohot 2022-06-05 15:56:42 -07:00
• 60b8689ea2 code2 was unneeded for reduce op geohot 2022-06-05 15:15:04 -07:00
• 46aaac5aaf clean up loops geohot 2022-06-05 15:00:21 -07:00
• 44168c4c40 remove all ctxes geohot 2022-06-05 14:55:39 -07:00
• f01bad36c2 factor out convs geohot 2022-06-05 14:48:42 -07:00
• b49bfb6e02 fix conv args. fix spacing geohot 2022-06-05 14:35:31 -07:00
• 365e62a609 refactor out matmul geohot 2022-06-05 14:33:06 -07:00
• 11d0cfec77 more readable and faster geohot 2022-06-05 14:13:08 -07:00
• d1b6f9822c UnaryOp class should make linecount geohot 2022-06-05 14:07:02 -07:00
• 24ed95e667 gpu refactors geohot 2022-06-05 13:55:49 -07:00
• 7a3fe34db1 GPU llops geohot 2022-06-05 13:49:39 -07:00
• f0fe37bd34 simpler graph demo geohot 2022-06-05 12:40:12 -07:00
• 89acf6742d more graph docs geohot 2022-06-05 12:16:50 -07:00
• 88de42fb6e document graph mode geohot 2022-06-05 12:13:05 -07:00
• 2097d814f6 Sum doesn't need to save the tensor geohot 2022-06-05 12:04:51 -07:00
• c8b569a8c7 cleaner comments geohot 2022-05-14 21:28:39 -07:00
• 8a7bbcc240 Accel/cuda (#319) Adrian BakkeandGitHub 2022-05-15 06:25:30 +02:00
• 7ad60eb8b2 remove redundant list comprehensions (#317) Drew HintzandGitHub 2022-04-05 12:30:50 -05:00