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

• 8391648822 Over 90% on CIFAR with examples/hlb_cifar10.py (#1073) 3 people 2023-07-07 09:16:22 +05:30
• c5aea13a65 Fix evaluation stage in examples/transformer.py when using CUDA (#1150) BarathandGitHub 2023-07-07 09:07:10 +08:00
• 9975f24452 Fold expand preceding reduce if the reduction is on the same axis as the expansion (#1134) Rayan HatoutandGitHub 2023-07-06 21:41:05 +01:00
• f109af3cbb Don't save parents unless needed (#1142) cheeetooandGitHub 2023-07-05 20:11:57 -05:00
• 801564f31b Remove POW llop and add SQRT llop (#1104) 3 people 2023-07-05 20:07:58 -05:00
• b7369ffcff add ptx formatter + syntax highlighter (#1128) cloud11665andGitHub 2023-07-06 02:56:09 +02:00
• d1356cac27 Fix: Jacobian tests [WIP] (#1126) Reza RezvanandGitHub 2023-07-06 00:36:22 +02:00
• d363d25ee2 fix imports for examples/transformer.py (#1136) nimlgenandGitHub 2023-07-05 18:15:13 +03:00
• c3173ff281 Add return statement to the train function (#1135) Mehmet KuzucuandGitHub 2023-07-05 17:13:38 +02:00
• 981d4980c4 feat: reword contributing (#1131) wozeparrotandGitHub 2023-07-05 01:17:47 -04:00
• 793a670187 from tensor cores + lb touchup (#1127) George HotzandGitHub 2023-07-04 15:45:20 -07:00
• 2f968f8547 ignore cloudpickle type for local mypy geohot 2023-07-04 13:51:20 -07:00
• 87d21ea979 examples: simple conv bn geohot 2023-07-04 13:50:26 -07:00
• 535224ac20 Remove float64 (#1101) Reza RezvanandGitHub 2023-07-04 17:40:51 +02:00
• b4ce23e4b8 Make cross_process use cloudpickle (#1118) Daniel HipkeandGitHub 2023-07-04 00:47:34 -07:00
• c709dec8b5 gelu: weird test was broken for metal geohot 2023-07-04 00:43:54 -07:00
• daf8e1942f sigmoid: test large postive also and add note geohot 2023-07-04 00:18:31 -07:00
• 9e6067378f Broken Sigmoid backward: Add test and mlop for Sigmoid (#1113) 3 people 2023-07-04 12:44:22 +05:30
• d58a9603ab Create COCO data directory if it doesn't exist. (#1114) Daniel HipkeandGitHub 2023-07-03 18:15:53 -07:00
• a22aad7d32 Use generators instead of lists in anys and alls (#1111) 3 people 2023-07-04 01:06:06 +02:00
• fd98f6cffa Small fix to abstractions.py so it runs on Windows without throwing an AttributeError (#1109) 3 people 2023-07-03 13:44:49 -07:00
• 651d080594 [perf] Replace more list comprehension with * (#1106) Mike OvyanandGitHub 2023-07-03 20:49:23 +03:00
• 2071e53da8 Handle broadcast flag on gemm (#1103) Frank PinnolaandGitHub 2023-07-03 01:15:07 -04:00
• cbb5c655e5 [tensor][perf] Replace list comprehension with *. (#1102) Taras TsugriiandGitHub 2023-07-02 20:34:23 -05:00
• 363fbfc2e4 do not emit loop end code for global+local loops in assembly kernel (#1100) David HouandGitHub 2023-07-02 18:33:57 -07:00
• 8ae9a054ae Refactor nn.optim (#1091) Reza RezvanandGitHub 2023-07-03 00:07:30 +02:00
• 10f1aeb144 fixed broken link (#1097) Eli FrigoandGitHub 2023-07-02 17:06:59 -05:00
• c8ddc34368 include missing queue in thneed load (#1095) Rob GrossmanandGitHub 2023-07-02 12:33:59 -07:00
• 12ce68c1ee Renamed examples/yolo to examples/vgg7_helpers because that directory contains no yolo-related code and only helper code for vgg7. This was confusing to a new user when trying to understand the examples. (#1086) nmarwell26andGitHub 2023-07-01 15:04:28 -04:00
• 2533a992e7 remove unused imports in models (#1088) Rob GrossmanandGitHub 2023-07-01 12:04:19 -07:00
• 575f75f613 hello (#1084) geohotstanandGitHub 2023-07-01 16:29:35 +08:00
• 574cbda979 Quickstart (#1015) foreign-subandGitHub 2023-06-29 22:26:58 +02:00
• 542b2d93a5 Perf/cache string ops (#1078) 3 people 2023-06-29 22:23:11 +02:00
• e234bf2298 hip matmul : add K support geohot 2023-06-28 19:54:33 +00:00
• 0e93b9642a hip matmul geohot 2023-06-28 19:21:01 +00:00
• 754e54ebb9 Fix Tensor ceil and floor for whole numbers (#1071) Jacky LeeandGitHub 2023-06-27 23:22:17 -07:00
• 1f5d45ca8c imagenet loader minor cleanups geohot 2023-06-28 05:08:09 +00:00
• 6ec0a24706 imagenet eval in 1 min 28 sec geohot 2023-06-28 04:23:26 +00:00
• 9fabdbd054 speed (#1070) George HotzandGitHub 2023-06-27 20:28:57 -07:00
• d16c16ec28 new upcast works (#1066) George HotzandGitHub 2023-06-27 19:34:53 -07:00
• 4d703be6d7 fix typo (#1065) ernieandGitHub 2023-06-27 19:56:54 +02:00
• 70c07dfea5 5k line max (#1064) George HotzandGitHub 2023-06-27 10:53:18 -07:00
• c8d87eb8d4 strip whitespace geohot 2023-06-27 10:11:43 -07:00
• 23648538fa fix folding of float4 add/mul (#1060) Rayan HatoutandGitHub 2023-06-27 04:59:29 +01:00
• a98e361da0 torch speed test, add add geohot 2023-06-26 18:55:27 -07:00
• 3e33befc1d realize hotspots (#1059) George HotzandGitHub 2023-06-26 18:31:18 -07:00
• 2977fb17f6 various touchups (#1058) George HotzandGitHub 2023-06-26 15:41:23 -07:00
• f265e8523a movement ops aren't really ops (#1056) George HotzandGitHub 2023-06-26 15:01:28 -07:00
• 65cbaa3429 no need to slice A and B twice in LLaMa complex multiplication (#1054) Rayan HatoutandGitHub 2023-06-26 22:42:58 +01:00
• 571089f10e Back off minor speed stuff for simplicity (#1053) George HotzandGitHub 2023-06-26 14:42:17 -07:00
• dedbd970aa Optimizations in lazy.py (#987) Rayan HatoutandGitHub 2023-06-26 21:55:42 +01:00
• 8bea6b6d35 perf/refactor_weakops (#1052) 3 people 2023-06-26 19:13:33 +02:00
• 8c65f9324c refactor: print formatting for llama timing (#1050) 3 people 2023-06-26 18:49:31 +02:00
• c604ef4beb symbolic.py: faster Node.sum, faster SumNode.div (#1014) 3 people 2023-06-26 18:47:17 +02:00
• 52b7105f87 Dedup params in Optimizer (#1047) Casey PrimozicandGitHub 2023-06-26 00:49:23 -07:00
• 5d3310ce56 MaskRCNN Inference (#884) 3 people 2023-06-26 04:07:51 +05:30
• 0f281e7b18 touchups geohot 2023-06-25 15:24:26 -07:00
• c8fbdeb48e test speed llama (#1046) George HotzandGitHub 2023-06-25 15:22:56 -07:00
• 5d16cc283f Docker fix (#1039) Jacky LeeandGitHub 2023-06-25 10:38:58 -07:00
• 6ff720103e Reduce tensor dot line count and fixed 1d tensor dot (#1045) Francesco CastelliandGitHub 2023-06-25 19:32:45 +02:00
• 9c6e507518 move accel into extra geohot 2023-06-23 16:38:15 -07:00
• 7f73d6a4da Fix input path in examples/compile_efficientnet.py, examples/efficientnet.py. (#1034) Yair LifshitzandGitHub 2023-06-24 01:34:33 +02:00
• 0222ee7bd2 feat: fix shell alias on readme (#1022) 兰天游andGitHub 2023-06-23 15:00:34 +08:00
• 264b1e5f48 cache gpuocelot build in cuda CI (#1032) cloud11665andGitHub 2023-06-23 02:42:12 +02:00
• 2407690d82 add cuda on cpu tests (#1020) cloud11665andGitHub 2023-06-22 23:15:50 +02:00
• e09219df0f fixed division by zero for fast kernels (#1021) Eli FrigoandGitHub 2023-06-22 16:02:53 -05:00
• 18892242b0 global -> group (#1007) George HotzandGitHub 2023-06-21 11:50:43 -07:00
• aab9ee0fca Add RDNA3 assembler UOps.CAST partial support + other fixes/improvements (#1012) Casey PrimozicandGitHub 2023-06-20 11:34:10 -07:00
• 57d3aa76a5 Windows & Ubuntu CLANG CI support (#1011) DiogoandGitHub 2023-06-19 12:33:24 -04:00
• 0d4c4f4e9e metal ci attempt (#1010) George HotzandGitHub 2023-06-19 09:23:55 -07:00
• 0ac84d5e94 exclude a few more onnx tests geohot 2023-06-18 22:34:30 -07:00
• 0fd648dff4 exclude more dumb onnx tests geohot 2023-06-18 22:29:46 -07:00
• b6102ba4ac added CUDA and PTX to env_vars.md (#1009) Pasan PereraandGitHub 2023-06-19 21:17:44 +05:30
• e829e0e718 Update CONTRIBUTING.md (#1008) Sayantan DasandGitHub 2023-06-19 10:39:03 +05:30
• d84c600e5d contibuting geohot 2023-06-18 21:48:05 -07:00
• 651d6ea457 Minor improvements + cleanup to ops_gpu.py (#1006) Casey PrimozicandGitHub 2023-06-18 21:26:40 -07:00
• 5428b5d774 good changes from tensor_cores branch (#1005) George HotzandGitHub 2023-06-18 20:28:06 -07:00
• ccb51ff5b0 "Fixed argument passing in example yolov8" (#1004) Yann HuynhandGitHub 2023-06-18 23:29:39 +02:00
• b14b7bc749 don't make HIP the default...it's slower geohot 2023-06-18 19:11:39 +00:00
• 3d63c71e27 HIP backend (#750) Alex WangandGitHub 2023-06-19 02:35:57 +08:00
• 805eef10dd Add tensorflow GEMM benchmark script (#1000) Casey PrimozicandGitHub 2023-06-18 10:57:45 -07:00
• c690eeaca9 flip mulacc to save a line (#997) George HotzandGitHub 2023-06-17 16:47:55 -07:00
• d2b837c1d9 Adds floor/ceil (#989) DiogoandGitHub 2023-06-17 13:56:21 -04:00
• 775287ed91 Add yolov8 implementation (#806) 4 people 2023-06-17 07:25:19 +05:30
• fe71282ba1 faster RDNA assembly backend (#990) George HotzandGitHub 2023-06-16 12:06:38 -07:00
• ba56ee6020 RDNA assembly backend ($1000 bounty) (#787) George HotzandGitHub 2023-06-16 09:33:18 -07:00
• dca084f227 minor == to is touchups geohot 2023-06-15 17:11:12 -07:00
• 041d96083c clang rt for msvc (#986) blakeandGitHub 2023-06-15 17:06:44 -07:00
• 039f0d372f delete ltypes (#984) George HotzandGitHub 2023-06-15 16:24:45 -07:00
• 804c45b5fc FIX: Can't pickle local object (#979) Yahya LmallasandGitHub 2023-06-14 20:32:17 +01:00
• 2d567ef688 Optimizations in tensor.py (#974) Rayan HatoutandGitHub 2023-06-14 16:44:35 +01:00
• 0629791cbd F64 support (#976) DiogoandGitHub 2023-06-14 00:31:31 -04:00
• 45bc040a63 Fix typo (#978) John MooreandGitHub 2023-06-13 18:15:45 -04:00
• 80e665bddb a couple new tests George Hotz 2023-06-13 12:36:05 -07:00
• ba4eadb04c PTX assembly support (#977) George HotzandGitHub 2023-06-13 12:31:42 -07:00
• 727416201f Shapetracker optimizations (#966) Rayan HatoutandGitHub 2023-06-13 02:13:21 +01:00
• 5f13e7c3cf cuda: fix fp16, uint8, int64, half4 codegen (#968) cloud11665andGitHub 2023-06-12 20:15:44 +02:00
• e54b6c5e7f One hot (#972) Steven AndersonandGitHub 2023-06-12 10:13:29 -07:00
• 613c74ca9f maintain input tensor dtype (#969) DiogoandGitHub 2023-06-12 13:12:47 -04:00
• 2d4370b487 Adds tril & triu support (#936) DiogoandGitHub 2023-06-10 01:13:20 -04:00