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

• b3755370ae substitute assign source for KERNEL + comments [pr] (#9255) qazalandGitHub 2025-02-26 12:44:29 +02:00
• 941559098b do not lockup VIZ when rendering big graphs [pr] (#8795) qazalandGitHub 2025-02-26 10:15:26 +02:00
• e162aa862d is_realized only if buffer is allocated (#9253) qazalandGitHub 2025-02-26 09:58:08 +02:00
• b603af373e run some tests from torch [pr] (#9252) George HotzandGitHub 2025-02-26 15:42:22 +08:00
• 3f4eb9006a test for device mismatch [pr] (#9250) George HotzandGitHub 2025-02-26 13:06:33 +08:00
• 9c4d9d9f10 Acc first (#9232) 4 people 2025-02-26 04:10:15 +01:00
• 979e84f30e RESET_STEP in bert setup and beam (#9248) chenyuandGitHub 2025-02-25 19:15:10 -05:00
• 2676c9d46e dsp: raise exec errors as RuntimeError for beam (#9246) nimlgenandGitHub 2025-02-25 19:22:35 +03:00
• 70db8c3003 hcq: dyn alloc signals (#9238) nimlgenandGitHub 2025-02-25 17:22:24 +03:00
• 6610ad58ab hotfix bert no shard with only one device (#9243) chenyuandGitHub 2025-02-25 09:05:11 -05:00
• bba9c22f53 implement the new subbuffer spec for DISK [pr] (#9241) qazalandGitHub 2025-02-25 14:36:23 +02:00
• 48dfed064a remove const/var from the kernel graph [pr] (#9240) qazalandGitHub 2025-02-25 13:21:55 +02:00
• b4c3780df0 hotfix: interop example (#9237) nimlgenandGitHub 2025-02-25 10:32:00 +03:00
• 8c7be428e5 update bert BS to 78 (#9236) chenyuandGitHub 2025-02-24 22:47:35 -05:00
• 990c240b82 Stable pow gradient (#9226) 3 people 2025-02-25 02:54:26 +01:00
• 731d14e718 hotfix bump testmetal2 timeout-minutes to 20 (#9235) chenyuandGitHub 2025-02-24 20:23:56 -05:00
• cbfe95d306 bring cast before view back (#9230) 3 people 2025-02-25 01:50:39 +02:00
• 90c3ed17c5 move cast to before softmax in attention (#9213) chenyuandGitHub 2025-02-24 17:24:59 -05:00
• f0b24d230c add test_onnx_ops.py (#8569) 3 people 2025-02-25 05:15:22 +08:00
• 56288243e6 metal PyTorch interop (#9229) 3 people 2025-02-24 22:36:08 +03:00
• 687d157906 delete cast early folding from ops [pr] (#9228) qazalandGitHub 2025-02-24 20:00:51 +02:00
• c9493e41a6 reorder expand (#9051) 3 people 2025-02-24 20:55:47 +08:00
• 14aa2395d0 allow VIEW(BUFFER) in Tensor UOps [pr] (#9210) qazalandGitHub 2025-02-24 14:06:15 +02:00
• 1d06d61b16 from_blob for cuda (#9223) 3 people 2025-02-24 14:02:06 +03:00
• fc32ff80d6 torch and numpy dtype interop [pr] (#9224) George HotzandGitHub 2025-02-24 18:26:49 +08:00
• 24615db5f5 hotfix: torch cuda interop example geohot 2025-02-24 09:02:48 +00:00
• fd731e740a hotfix: add note on backend2.py geohot 2025-02-24 11:23:03 +08:00
• f2dd9c1562 simplify c++ code (#9221) albanDandGitHub 2025-02-23 22:04:41 -05:00
• d12efc95d4 support custom name function in viz [pr] (#9219) qazalandGitHub 2025-02-24 03:03:25 +02:00
• b3ae664d5d fix gradient of pow(t, int) (#9217) chenyuandGitHub 2025-02-23 17:42:09 -05:00
• 12b5b83821 set TRACK_MATCH_STATS=0 for real_strides [pr] (#9216) qazalandGitHub 2025-02-23 23:26:31 +02:00
• 9db0ec46a7 simpler buf_uop [pr] (#9215) qazalandGitHub 2025-02-23 20:23:14 +02:00
• 898aafe6fd move split_reduceop to scheduler + enable it for multi (#9214) qazalandGitHub 2025-02-23 18:30:04 +02:00
• 05e3202fba remove unused memsize_to_str and minor cleanups [pr] (#9211) ShikChenandGitHub 2025-02-23 22:58:37 +08:00
• 81a71ae0f6 hotfix: skip test_exclude_const_metadata (#9208) qazalandGitHub 2025-02-22 23:26:04 +02:00
• e0adb1fc76 really run test_ops with TINY_BACKEND in ci (#9206) chenyuandGitHub 2025-02-22 15:51:24 -05:00
• e6d20c47e3 simpler becomes_map update [pr] (#9201) qazalandGitHub 2025-02-22 21:50:58 +02:00
• 4578c3e8fd simpler tensor metadata mapping + tests [pr] (#9203) qazalandGitHub 2025-02-22 21:18:46 +02:00
• b711c6343a no early return + allow childless const/bind/var in kernel graph [pr] (#9202) qazalandGitHub 2025-02-22 20:28:22 +02:00
• 97bc723538 torch backend works for ResNet-18 (#9200) George HotzandGitHub 2025-02-22 22:16:23 +08:00
• f92820d30d torch backend tests (#9198) George HotzandGitHub 2025-02-22 16:01:49 +08:00
• 4e6665bda5 different way to write torch backend (#9197) George HotzandGitHub 2025-02-22 14:42:26 +08:00
• 041b6d5678 am: load fw in batches (#9185) nimlgenandGitHub 2025-02-21 23:21:31 +03:00
• 1db4341e9f move viz graph to lib/graph [pr] (#9196) qazalandGitHub 2025-02-21 22:04:07 +02:00
• 6587c7879b simple fixes to onnx (#9195) geohotstanandGitHub 2025-02-22 02:10:06 +08:00
• 2318d7ac51 Add missing tinygrad.runtime.autogen.am to packages (#9194) Simon RandGitHub 2025-02-21 14:38:24 +01:00
• 8bb80b6e5e reorder AST matchers + comments [pr] (#9193) qazalandGitHub 2025-02-21 15:31:15 +02:00
• 2eab8021fb remove inputs+outputs attributes from ScheduleItem [pr] (#9192) qazalandGitHub 2025-02-21 14:48:11 +02:00
• e87be0131e torch backend start (#9191) George HotzandGitHub 2025-02-21 16:57:28 +08:00
• d3a21cced2 hotfix: bump version to 0.10.2 v0.10.2 geohot 2025-02-21 10:43:49 +08:00
• 2e7c2780a9 CLANG -> CPU (#9189) chenyuandGitHub 2025-02-20 18:03:09 -05:00
• f986e12f91 metal: choose compile spec based on macos (#9188) nimlgenandGitHub 2025-02-21 00:43:39 +03:00
• 3e22747799 run unit test on windows ci (#9187) chenyuandGitHub 2025-02-20 14:40:41 -05:00
• 287de4ecc6 use torch in test_gradient (#9186) chenyuandGitHub 2025-02-20 12:26:11 -05:00
• 574a905291 Fix running VIZ=1 after package installation + test (#9183) qazalandGitHub 2025-02-20 16:02:00 +02:00
• 1692087db5 _one_hot_along_dim input needs to be int (#9179) chenyuandGitHub 2025-02-20 09:00:43 -05:00
• bf36967883 cuda hooking (#9180) George HotzandGitHub 2025-02-20 19:20:01 +08:00
• 3b37cc898b add bert tiny config (#9177) chenyuandGitHub 2025-02-19 14:57:03 -05:00
• 5662c898f1 correctly step through bottom_up_rewrites in viz [pr] (#9176) qazalandGitHub 2025-02-19 20:20:57 +02:00
• b1ddb2a1a6 fix win32 CPUProgram missing cache flush (#9171) peppingdoreandGitHub 2025-02-19 16:38:51 +03:00
• 1bb9d78c7a hotfix: add output buffer back to kernel parents + comment [pr] (#9174) qazalandGitHub 2025-02-19 15:22:01 +02:00
• 975c318dbc bert use int32 for input ids (#9173) chenyuandGitHub 2025-02-19 08:17:27 -05:00
• e4a8bf28ea scheduler cleanups + better cycle assert [pr] (#9172) qazalandGitHub 2025-02-19 14:30:58 +02:00
• cf315d544b rename can_pad arg to cache [pr] (#9170) qazalandGitHub 2025-02-19 13:24:59 +02:00
• 2fc8bf115d remove support for VIEW with two sources in ops [pr] (#9168) qazalandGitHub 2025-02-19 12:10:18 +02:00
• a2afa523a0 Only add enable f16 directive if ShaderF16 is supported (#9163) Ahmed HarmoucheandGitHub 2025-02-19 10:20:03 +01:00
• 0f94b98646 Force WebGPU backend type [pr] (#9164) Ahmed HarmoucheandGitHub 2025-02-19 10:19:39 +01:00
• 4bc708a9b0 do not create buffers we never realize in scheduler (#9165) qazalandGitHub 2025-02-19 11:11:47 +02:00
• 1c4e9bc363 image fixup tensor map [pr] (#8611) 3 people 2025-02-19 16:11:06 +08:00
• 2a5fe3e700 whitespace changes from the map_tensors branch [pr] (#9167) qazalandGitHub 2025-02-19 09:52:59 +02:00
• a773ff73e3 match image cast folding on the cast itself [pr] (#9166) qazalandGitHub 2025-02-19 09:31:34 +02:00
• 9a20063837 create subbuffer immediately before constructing ScheduleItem [pr] (#9162) qazalandGitHub 2025-02-18 22:07:52 +02:00
• 1c92534bff hotfix: viz should show if there's a rewrite [pr] (#9161) qazalandGitHub 2025-02-18 20:11:03 +02:00
• a330f3338c save applied opts in ProgramSpec [pr] (#9150) George HotzandGitHub 2025-02-19 00:40:03 +08:00
• ff05bff221 put bert data shard inside jit (#9160) chenyuandGitHub 2025-02-18 10:36:54 -05:00
• 679291e26a assert only base maps to buffer [pr] (#9159) qazalandGitHub 2025-02-18 16:46:47 +02:00
• 4f592eeea6 hotfix: remove extra matcher for copy/buffer_view [pr] (#9157) qazalandGitHub 2025-02-18 14:21:24 +02:00
• ff9b985d9f hotfix: View Base AST geohot 2025-02-18 18:48:34 +08:00
• 30f470eaa3 UNIQUE UOp for buffer instead of arg (#9156) George HotzandGitHub 2025-02-18 16:59:59 +08:00
• 38f5ea2132 increment writable buffers refcount from the kernel graph [pr] (#9153) qazalandGitHub 2025-02-18 10:20:02 +02:00
• ddddcc165b colors back in DEBUG=2 [pr] (#9155) George HotzandGitHub 2025-02-18 16:17:57 +08:00
• 6d62966bf7 add support for named rewrites [pr] (#9152) George HotzandGitHub 2025-02-18 16:07:04 +08:00
• caee42e8a6 Revert "name from uops [pr] (#9151)" (#9154) George HotzandGitHub 2025-02-18 16:06:44 +08:00
• 28897be9a2 name from uops [pr] (#9151) George HotzandGitHub 2025-02-18 15:52:03 +08:00
• a4dab3ec3f add name uop (#9149) George HotzandGitHub 2025-02-18 15:26:58 +08:00
• 2db8b4046a minor linearizer refactor to finalize in rewrite [pr] (#9148) George HotzandGitHub 2025-02-18 12:42:22 +08:00
• df3b320f46 rewriter -> devectorizer [pr] (#9147) George HotzandGitHub 2025-02-18 12:42:08 +08:00
• 5dc1257ce0 clean up bert fake data iterator [pr] (#9145) chenyuandGitHub 2025-02-17 20:03:38 -05:00
• 751c517b6c cancel viz request after the kernel clicked away [pr] (#9144) qazalandGitHub 2025-02-17 21:19:09 +02:00
• 465421b525 fix Tensor.isclose (#9143) chenyuandGitHub 2025-02-17 12:03:12 -05:00
• 36741cbbc1 enable real_size assert for test_conv_2x2_backward_one_view [pr] (#9142) qazalandGitHub 2025-02-17 18:53:44 +02:00
• e9ff4ef4f7 s/ScheduleContext/GrouperContext [pr] (#9141) qazalandGitHub 2025-02-17 18:14:17 +02:00
• 96cc9f59e0 refactor to kernel context [pr] (#9140) qazalandGitHub 2025-02-17 17:57:14 +02:00
• df6781332e remove var_vals from the scheduler context [pr] (#9139) qazalandGitHub 2025-02-17 17:43:50 +02:00
• 35e9c4657b Use proper units when printing beam time (#9103) Ali LadjevardiandGitHub 2025-02-17 19:11:38 +03:30
• a7f91224eb add Tensor.isclose() (#8844) 3 people 2025-02-17 23:11:40 +08:00
• 2b787c3b17 hotfix: lower ul.disabled opacity for viz [pr] (#9138) qazalandGitHub 2025-02-17 16:16:48 +02:00
• 660c034da6 KERNEL op try 3 (#9061) qazalandGitHub 2025-02-17 15:47:54 +02:00
• ec80df5115 add PROGRAM renderer to viz [pr] (#9137) qazalandGitHub 2025-02-17 15:46:08 +02:00
• 7b09a72682 don't display void dtype in viz nodes [pr] (#9136) qazalandGitHub 2025-02-17 14:49:36 +02:00