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

• 61e255d197 use max for gpt2 and llama (#2949) chenyuandGitHub 2023-12-28 23:26:00 -05:00
• c7b106bf9c hotfix float4 only supports float and half (#2948) chenyuandGitHub 2023-12-28 20:23:52 -05:00
• 2f67f1e580 remove obsolete TODO in beautiful_mnist (#2946) chenyuandGitHub 2023-12-28 17:09:23 -05:00
• 50f2e31d26 cleanup float4 grouping in global_load and global_store (#2942) chenyuandGitHub 2023-12-27 14:10:04 -05:00
• 54629b56d2 minor cleanup in kernel and linearizer (#2937) chenyuandGitHub 2023-12-26 12:05:32 -05:00
• 820f2e054e fix PADTO optimization (#2935) chenyuandGitHub 2023-12-25 22:52:49 -05:00
• dca5e4fe74 tensor == tensor should be bool (#2916) qazalandGitHub 2023-12-25 19:38:47 +02:00
• 8a8aed23d2 test dtypes of return values of cumsum, argmax/min, multinomial (#2933) chenyuandGitHub 2023-12-25 11:33:17 -05:00
• 12996d3a7d green linearizer asserts for ops (#2800) 3 people 2023-12-25 17:41:54 +02:00
• 1fb815e77e hotfix fix coder. RMSNorm cannot have float16 input (#2932) chenyuandGitHub 2023-12-25 02:28:11 -05:00
• b469fe3723 add CMPEQ (#2931) chenyuandGitHub 2023-12-25 00:15:55 -05:00
• 016aebcd84 Fixed Tensor.randint() not accepting tuple shapes (#2923) WillandGitHub 2023-12-25 11:32:26 +10:00
• 2dc99af169 clean up manual cast dtypes in tensor.py (#2930) chenyuandGitHub 2023-12-24 13:03:41 -05:00
• 8de1fc2539 Einsum space fix (#2927) Isalia20andGitHub 2023-12-24 10:23:27 +04:00
• b55b55d56e use at least int32 and uint32 for sum output (#2926) chenyuandGitHub 2023-12-24 01:14:54 -05:00
• d424babe2c tensor.py cleanup around Tensor.slice (#2921) chenyuandGitHub 2023-12-22 19:46:39 -05:00
• 089703a390 cleanup test_dtype_alu (#2919) chenyuandGitHub 2023-12-22 17:29:31 -05:00
• 3ba591c3fd less outdated abstraction.py (#2917) chenyuandGitHub 2023-12-22 15:31:02 -05:00
• 50927defad s/lazydata.realized/lazydata.base.realized/g (#2914) chenyuandGitHub 2023-12-22 14:45:13 -05:00
• 2783e1b50d bugfix Tensor.item when it's unbased (#2913) chenyuandGitHub 2023-12-22 13:50:06 -05:00
• c3133adb8c Disk shm refactor (#2912) Oleg RybalkoandGitHub 2023-12-22 20:23:37 +03:00
• 3855432265 don't use numpy to create Tensor(None) (#2909) chenyuandGitHub 2023-12-22 01:07:44 -05:00
• 50cfb1fb3a update onnx model links (#2908) chenyuandGitHub 2023-12-22 00:19:41 -05:00
• 1bbeb3fe2f remove the different rtol / atol for openpilot CUDA in benchmark (#2907) chenyuandGitHub 2023-12-21 22:23:39 -05:00
• a543d8bea8 fuzz default dtypes for some test_dtype tests (#2906) chenyuandGitHub 2023-12-21 22:00:21 -05:00
• 5f3d5cfb02 catch cycles in print_tree (#2891) wozeparrotandGitHub 2023-12-21 21:40:37 -05:00
• 4432cb17bb minor cleanups / remove that op (#2905) George HotzandGitHub 2023-12-21 18:24:20 -08:00
• fd0ba33b38 onnx_ops formatting cleanup (#2904) chenyuandGitHub 2023-12-21 20:06:06 -05:00
• 5cac6338a4 apply the multitensor optimizations in lazy.py (#2901) George HotzandGitHub 2023-12-21 13:55:49 -08:00
• 5bf43c9634 reenable one onnx test failed due to dtype (#2902) chenyuandGitHub 2023-12-21 15:50:02 -05:00
• 677ae7673d use np.less and torch.lt for CMPLT (#2899) chenyuandGitHub 2023-12-21 14:37:24 -05:00
• d2e9245de8 render_locals takes a dtype (#2873) 3 people 2023-12-21 21:15:28 +02:00
• 6116039f7b don't match dtype with first input in where (#2898) chenyuandGitHub 2023-12-21 13:02:15 -05:00
• 7dc3352877 increase stable diffusion validation threshold 1e-4 -> 3e-4 (#2897) chenyuandGitHub 2023-12-21 11:45:25 -05:00
• 24e79e0f53 Move the webgpu CMPLT hack to one place (#2895) qazalandGitHub 2023-12-21 18:14:56 +02:00
• 852ef57ba4 fix readme typo geohot 2023-12-21 08:06:24 -08:00
• 193109a88c hotfix: compare on ids geohot 2023-12-20 23:45:59 -08:00
• f6c7833f9f fast compare for lazyop (#2893) George HotzandGitHub 2023-12-20 23:32:27 -08:00
• 1500aca43d remove output_type in ops_cpu and ops_torch (#2892) chenyuandGitHub 2023-12-21 02:11:27 -05:00
• 2d2c4980fe assert for elementwise dtypes in lazy (#2888) chenyuandGitHub 2023-12-21 01:42:32 -05:00
• 41b2a25be6 Fix exponential behavior in lazyops (#2890) George HotzandGitHub 2023-12-20 22:06:50 -08:00
• 8c4a0f8e15 Fix int child count (#2882) George HotzandGitHub 2023-12-20 21:06:27 -08:00
• 8a04107d30 move the op casting logic from mlops to tensor try 2 (#2887) chenyuandGitHub 2023-12-20 23:50:37 -05:00
• 7da2325dc7 get_lazyops() -> lazyops (#2884) George HotzandGitHub 2023-12-20 18:04:49 -08:00
• 64dded27f0 pad ops broke coder (#2881) George HotzandGitHub 2023-12-20 17:03:41 -08:00
• e1861ab65e remove realize from optimizer (#2880) George HotzandGitHub 2023-12-20 16:42:41 -08:00
• 1765849937 new lazy, benchmark (#2878) George HotzandGitHub 2023-12-20 14:33:21 -08:00
• dae8976889 Fix reshape merging with masks (#2877) Peter CawleyandGitHub 2023-12-20 22:00:58 +00:00
• 8fe24038d8 Revert "mulacc fusion cleanup (#2871)" (#2876) George HotzandGitHub 2023-12-20 13:26:25 -08:00
• 863c5b26ed mulacc fusion cleanup (#2871) qazalandGitHub 2023-12-20 22:39:54 +02:00
• e13b4964d7 remove the all_int(shape) check in Tensor._loadop (#2874) chenyuandGitHub 2023-12-20 15:04:50 -05:00
• 5f07ef455e update dtypes (#2872) qazalandGitHub 2023-12-20 22:04:02 +02:00
• 857c35d256 make gpt2 decode output just once at the end (#2869) chenyuandGitHub 2023-12-20 12:14:55 -05:00
• e92069fb1c remove unused symbolic.is_sym_int (#2868) chenyuandGitHub 2023-12-20 11:37:54 -05:00
• ca59054463 fix shapetracker math (#2861) George HotzandGitHub 2023-12-19 22:17:34 -08:00
• 5a739e8c20 update one skipped pad_reshape test that was fine (#2860) chenyuandGitHub 2023-12-19 23:25:52 -05:00
• 39af93ed7c minor tensor.py function cleanup (#2859) chenyuandGitHub 2023-12-19 22:39:39 -05:00
• 94f71fe238 random and empty shouldn't reshape geohot 2023-12-19 18:09:03 -08:00
• 637879af78 add direct install to readme geohot 2023-12-19 18:04:00 -08:00
• ad233d557f disable reshape merging with masks (#2858) chenyuandGitHub 2023-12-19 19:06:16 -05:00
• 1231ec5a02 run the sz.py line count at the end of linter ci (#2857) chenyuandGitHub 2023-12-19 16:33:12 -05:00
• ac6ec936cd update contributing geohot 2023-12-19 12:19:14 -08:00
• e477cc2f45 hotfix: README is ~25 ops to stop getting PRs about it geohot 2023-12-19 11:53:06 -08:00
• 42a038c83f More readable torch_load ext check (#2853) Oleg RybalkoandGitHub 2023-12-19 22:53:15 +03:00
• 172a88e719 skip slow test_indexing on METAL (#2852) chenyuandGitHub 2023-12-19 12:00:54 -05:00
• 6d7e9e0a56 hotfix convert Y_train to int before passing into index (#2850) chenyuandGitHub 2023-12-19 11:40:56 -05:00
• fec8e9060c Add simple fancy indexing exceptions (#2706) geohotstanandGitHub 2023-12-20 00:23:51 +08:00
• 417d42a363 UOps support for ImageDType (#2848) qazalandGitHub 2023-12-19 16:39:48 +02:00
• 90fb09b55c remove unused _device_extra_args geohot 2023-12-18 22:14:58 -08:00
• b2192b5400 minor improvements (#2845) George HotzandGitHub 2023-12-18 22:09:08 -08:00
• d086325b1b hotfix: failing tests geohot 2023-12-18 21:12:42 -08:00
• 07df14aa0e HIP cleanups (#2843) 4 people 2023-12-18 21:09:32 -08:00
• b6d71b131e hotfix: push broken tests geohot 2023-12-18 21:08:42 -08:00
• 9b35186bbe hotfix: don't import from tinygrad in sz.py geohot 2023-12-18 20:49:46 -08:00
• 6617dcf095 move graph to runtime, check line count with sz.py (#2842) George HotzandGitHub 2023-12-18 20:30:06 -08:00
• 15dc5bcfbd remove cstyle hip functions that deals with different types of input (#2840) chenyuandGitHub 2023-12-18 22:43:22 -05:00
• dad9253d52 minor clean up in kernels (#2832) chenyuandGitHub 2023-12-18 19:35:59 -05:00
• 954a2fef75 hotfix: add JITGRAPH and invert sints geohot 2023-12-18 16:33:22 -08:00
• 80f53245e8 shapetracker add and invert (#2828) George HotzandGitHub 2023-12-18 16:03:27 -08:00
• 73cadfbb3c Remove pytest markers (#2831) chenyuandGitHub 2023-12-18 18:53:28 -05:00
• 264fe9c93f clean up test_dtype.py (#2827) chenyuandGitHub 2023-12-18 16:06:09 -05:00
• 20ea43b6e7 dtypes.from_py to convert py types to dtypes (#2826) chenyuandGitHub 2023-12-18 14:23:31 -05:00
• 0723f26c80 dtypes.default_float and dtypes.default_int (#2824) chenyuandGitHub 2023-12-18 12:21:44 -05:00
• 8aab19ce3d Tensor.full of bool has dtypes.bool (#2823) chenyuandGitHub 2023-12-18 10:51:17 -05:00
• 220abcd8ff fix squeeze of 0-dim Tensor with negative dim (#2821) chenyuandGitHub 2023-12-17 22:02:07 -05:00
• 21ec7e09f6 minor cleanup of image (#2820) chenyuandGitHub 2023-12-17 20:53:49 -05:00
• 959d9cfed4 clean up ops_torch and ops_cpu (#2819) chenyuandGitHub 2023-12-17 19:35:19 -05:00
• f409b57854 update metal matmul and matvec for new device style (#2732) 3 people 2023-12-17 23:15:07 +02:00
• 91adb119b8 remove match_type in ops_torch and ops_cpu (#2817) chenyuandGitHub 2023-12-17 15:32:30 -05:00
• 887f3d9933 Make torch backend more usable, fix bfloat support in the llvm backend (#2765) Maksym SobolyevandGitHub 2023-12-17 11:04:26 -08:00
• 9c32474a1f Revert "Revert "Tensor.randint is Tensor.uniform with dtypes.int32 (#2801)" (#2802)" (#2814) chenyuandGitHub 2023-12-17 12:14:17 -05:00
• b4fa189c8c Revert "Revert "Make Tensor creation allow multi-dim list of int and bool (#2793)" (#2810)" (#2813) chenyuandGitHub 2023-12-17 11:48:27 -05:00
• 1fa4f161fe Update CLProgram to use unsigned long long for event profiling (#2808) Marcus AsteborgandGitHub 2023-12-16 23:48:44 -08:00
• c333bfcf69 replace a defaultdict counting with Counter in cstyle (#2809) chenyuandGitHub 2023-12-17 02:44:16 -05:00
• 4e2a92cee1 run HALF GPT2 in nvidia benchmark in addition to HALF/BEAM (#2811) chenyuandGitHub 2023-12-17 02:24:55 -05:00
• bad0ff60b7 start Qualcomm GPU driver (#2804) George HotzandGitHub 2023-12-16 23:10:50 -08:00
• 71a60762ed Revert "Make Tensor creation allow multi-dim list of int and bool (#2793)" (#2810) chenyuandGitHub 2023-12-17 02:03:52 -05:00
• 798bf813b1 Make Tensor creation allow multi-dim list of int and bool (#2793) geohotstanandGitHub 2023-12-17 14:58:10 +08:00
• 85c6250a3e support Tensor.einsum with no "->" in formula (#2807) chenyuandGitHub 2023-12-17 00:46:24 -05:00
• 157c0be509 cleanup onnx, pass one more reshape test and remove some casts (#2806) chenyuandGitHub 2023-12-16 20:40:43 -05:00