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

• 7103b716c4 merge kernel and optimizer (#2200) George HotzandGitHub 2023-11-01 15:20:01 -07:00
• 33bb650e94 use mad in opencl (#2198) 3 people 2023-11-01 10:40:08 -07:00
• c8b6a811ea no locals as opt action (#2196) George HotzandGitHub 2023-11-01 09:47:44 -07:00
• 2e9982fe2d fastvits example that's 10% faster Comma Device 2023-10-31 21:48:23 -07:00
• 8ba7ced7f9 extract const if it's const (#2193) George HotzandGitHub 2023-10-31 18:52:35 -07:00
• b245f1307e add exp2 (#2192) George HotzandGitHub 2023-10-31 17:48:42 -07:00
• e2428b63a6 external (#2191) qazalandGitHub 2023-10-31 16:57:24 -04:00
• 7e8c5f1a0f Modernize setup.py (#2187) Elias WahlandGitHub 2023-10-31 21:55:45 +01:00
• 8c07c73a9b Fix cl map buffer (#2190) nimlgenandGitHub 2023-10-31 22:02:46 +03:00
• c59ea32f90 prevent over unrolling in optimzer geohot 2023-10-31 11:45:18 -07:00
• 5aaa8a0cc1 fix shape geohot 2023-10-31 11:35:03 -07:00
• a27c9f9de5 openpilot compile2 (#2189) George HotzandGitHub 2023-10-31 11:08:58 -07:00
• be5f185ac0 Higher test coverage for dtypes (#2156) qazalandGitHub 2023-10-31 01:38:42 -04:00
• f294bdd681 fixed imports (#2185) forcefieldsovereignandGitHub 2023-10-30 22:07:17 -07:00
• 018bd29e37 Enable Multi-Output Export (#2179) Akshay KashyapandGitHub 2023-10-30 21:42:26 -04:00
• a7439af786 Fix llvm int->bool cast (#2164) qazalandGitHub 2023-10-30 18:28:23 -04:00
• 94cf652b6b don't use locals applies to GROUP also George Hotz 2023-10-30 13:56:43 -07:00
• 5cc536bcc0 don't use locals applies to LASTLOCAL George Hotz 2023-10-30 13:53:42 -07:00
• 3c88af5071 use unique table name for each disk_cache test (#2184) chenyuandGitHub 2023-10-30 16:49:49 -04:00
• 608e3ee800 fix no locals search and search both (#2171) George HotzandGitHub 2023-10-30 10:22:50 -07:00
• 194e4ad6f8 Revert "optimizer: simplify GROUP and LOCAL to have one of each (#2162)" (#2182) George HotzandGitHub 2023-10-30 10:22:26 -07:00
• 95f7183c3a Reenable global, local limiting (#2095) Ahmed HarmoucheandGitHub 2023-10-30 18:17:23 +01:00
• 8548b20b23 fix codellama params and repeat_kv (#2181) chenyuandGitHub 2023-10-30 13:16:26 -04:00
• c7f4dd6cb0 CACHELEVEL for smaller caches geohot 2023-10-28 07:26:03 -10:00
• 6c58bf3e9c in time_linearizer, allocate a scratch buffer if output buffer is also input (#2152) chenyuandGitHub 2023-10-28 13:17:41 -04:00
• 902f00b095 adding cuda TC headers (#2165) Yixiang GaoandGitHub 2023-10-27 19:25:59 -05:00
• 7f4f925385 fix hip del on compile fail (#2163) David HouandGitHub 2023-10-27 14:38:07 -07:00
• 8cf0bb9351 optimizer: simplify GROUP and LOCAL to have one of each (#2162) Francis LamandGitHub 2023-10-27 14:37:44 -07:00
• e0201922e3 Q network for pruning BEAM / uops deduping / BEAM_ESTIMATE (#2142) 4 people 2023-10-27 16:53:06 -04:00
• bc0829b677 Fix llama json loading (#2160) willandGitHub 2023-10-27 13:21:56 -07:00
• 8d41b3eb3f beam=16 makes gpt2 gpu-time < 5ms on 3090 (#2154) nimlgenandGitHub 2023-10-27 23:21:27 +03:00
• 5204864eca init cudagraph (#2153) nimlgenandGitHub 2023-10-27 23:19:50 +03:00
• 9215bccb41 Tensor.uniform set default to standard uniform (#2158) chenyuandGitHub 2023-10-27 16:15:30 -04:00
• 36ab04ae35 perf: lazyop as dataclass (#1603) 4 people 2023-10-25 23:54:30 +02:00
• 0ca0e9ee5e exclude ast with variables from beam search (#2140) chenyuandGitHub 2023-10-25 16:35:29 -04:00
• a52b420fb3 switch ocelot back to main repo (#2147) Szymon OżógandGitHub 2023-10-25 21:14:26 +02:00
• 12dd165d38 add WINO/HALF/HIP to AMD benchmark geohot 2023-10-25 13:22:45 -04:00
• bf3490cdf9 wmma: refactor tensor cores using existing local dims (#2097) 3 people 2023-10-25 10:10:46 -07:00
• c29653605e hip multigpu training (#1878) wozeparrotandGitHub 2023-10-24 17:35:53 -04:00
• 2e89fd264f Refactor hipgraph (#2141) nimlgenandGitHub 2023-10-24 22:45:56 +03:00
• e21bf776c8 fix debug=1 llama/gpt2 timings (#2143) nimlgenandGitHub 2023-10-24 22:45:00 +03:00
• 4444e6d4b3 stable diffusion < 324ms (#2129) chenyuandGitHub 2023-10-24 14:56:12 -04:00
• cea2bc7964 Add dictionary keys to reduce db size (#2131) George HotzandGitHub 2023-10-24 10:49:22 -04:00
• d5e2fdea22 remove shelve from handcode_resnet50_opt.py (#2139) chenyuandGitHub 2023-10-24 10:37:30 -04:00
• 228b310478 align cpu buffer before copy into cl buffer (#2135) imaoloandGitHub 2023-10-23 18:04:35 -07:00
• 6ee0435263 added from unaligned np test (#2134) imaoloandGitHub 2023-10-23 08:38:57 -07:00
• 3c56c181f6 string formatting 25 -> 30 to fit geohot 2023-10-22 10:57:34 -07:00
• 6dc8eb5bfd universal disk cache (#2130) George HotzandGitHub 2023-10-22 10:56:57 -07:00
• ace6b2a151 optimizer: add test for correctness of opts (#2124) Francis LamandGitHub 2023-10-22 08:02:22 -07:00
• abeba8f1fc optimization: get actions in CI (#2125) George HotzandGitHub 2023-10-20 12:22:01 -07:00
• 14625721e9 minor triton casting refactor (#2118) qazalandGitHub 2023-10-20 15:11:55 -04:00
• cb508e6923 uops graphing + phi (#2120) George HotzandGitHub 2023-10-19 22:26:28 -07:00
• bedd028061 waifu2x vgg7: testcase, auto-RGBA->RGB, function to grab pretrained models, training "fix" (#2117) 20kdcandGitHub 2023-10-20 06:07:15 +01:00
• e0b2bf46b4 Improve triton generated code quality (#2119) Szymon OżógandGitHub 2023-10-20 07:06:19 +02:00
• 36d4001b4f add test coverage for search (#2104) qazalandGitHub 2023-10-19 20:06:47 -04:00
• 7268b3c6fb make triton not write to disk (#2116) Szymon OżógandGitHub 2023-10-19 08:06:47 +02:00
• 95e17ff0d4 fix wino mask upcast calculation (#2057) David HouandGitHub 2023-10-18 16:54:48 -07:00
• 5cfec59abc hlb cifar touchups (#2113) George HotzandGitHub 2023-10-18 16:26:15 -07:00
• 5d5921d2c8 small doc env update (#2112) chenyuandGitHub 2023-10-18 17:49:25 -04:00
• 4526891db7 parallel apt (#2111) George HotzandGitHub 2023-10-18 14:49:00 -07:00
• 87b714b8cb split test_conv2d geohot 2023-10-18 14:00:50 -07:00
• 15da96f393 print test durations and add speed (#2107) George HotzandGitHub 2023-10-18 13:46:42 -07:00
• e2a1c2aaa6 force ruff reinstall geohot 2023-10-18 11:40:46 -07:00
• 0d2b3a9d33 full path for ruff geohot 2023-10-18 11:27:49 -07:00
• 8940c89d13 tests: remove 2 runners, make cache reliable (#2106) George HotzandGitHub 2023-10-18 11:10:41 -07:00
• b3afe0106b typo, src printing, and no verbose on triton (#2105) George HotzandGitHub 2023-10-18 09:44:36 -07:00
• 967a88a505 examples/waifu2x: Cleanup waifu2x vgg7 model format (now uses safetensors) (#2082) 20kdcandGitHub 2023-10-18 17:20:11 +01:00
• 881fd7c141 add mops to graph, refactor IMAGE (#2100) George HotzandGitHub 2023-10-17 21:27:51 -07:00
• 2498802b46 fix beam search for llvm, this needs tests (#2101) George HotzandGitHub 2023-10-17 20:09:42 -07:00
• 4d1e59abfd fix: only when distributed (#2102) wozeparrotandGitHub 2023-10-17 23:09:04 -04:00
• 999c95ea29 fix: hlb cifar types (#2099) Sean D'SouzaandGitHub 2023-10-17 22:23:50 -04:00
• 9b1c3cd9ca hlb_cifar: support EVAL_STEPS=1000, print when dataset is shuffled geohot 2023-10-18 01:11:08 +00:00
• 2b5ea7d9cb Fix output Float32Array size in webgpu export (#2096) Ahmed HarmoucheandGitHub 2023-10-18 00:28:19 +02:00
• 01b98b7f42 MulNode.__lt__ rule (#2086) Umut ZenginandGitHub 2023-10-17 23:18:35 +03:00
• f76fbd23e9 cleanup triton (#2092) Szymon OżógandGitHub 2023-10-17 21:49:44 +02:00
• 4bef1591f0 Disable ocelot cache + fix matvec in triton (#2010) Szymon OżógandGitHub 2023-10-17 19:33:32 +02:00
• 5ed630204b Add ONNX to CI for other backends (#2069) geohotstanandGitHub 2023-10-18 00:33:54 +08:00
• 5a4a62ecae Disable logging in early compile2 and lower kernel counts (#2090) George HotzandGitHub 2023-10-16 20:15:24 -07:00
• 442a27db8a shouldn't do anything (#2091) George HotzandGitHub 2023-10-16 18:18:34 -07:00
• 1bf4aef0f5 fix image dtype cmp (#2089) George HotzandGitHub 2023-10-16 17:52:38 -07:00
• e4846771b2 Revert "limit metal buffers and revert the 207 fix (try 2) (#2088)" geohot 2023-10-16 17:50:11 -07:00
• d0aaf7d83b Revert "Revert "Revert "openpilot kernel fix from 209 to 207 (#2006)" (#2065)"" geohot 2023-10-16 17:47:00 -07:00
• 5e24dc5a95 limit metal buffers and revert the 207 fix (try 2) (#2088) George HotzandGitHub 2023-10-16 14:52:16 -07:00
• e8fcd2f3db Revert "limit metal buffers and revert the 207 fix (#2087)" geohot 2023-10-16 14:32:22 -07:00
• 2fb10f6a19 limit metal buffers and revert the 207 fix (#2087) George HotzandGitHub 2023-10-16 14:26:32 -07:00
• a7b18ac325 try beam search on device (#2085) 3 people 2023-10-16 12:52:42 -07:00
• c36d306606 KOPT is over, BEAM is upstream (#2071) George HotzandGitHub 2023-10-16 09:46:03 -07:00
• e4660b024f mute hip warnings (#2081) nimlgenandGitHub 2023-10-16 17:09:10 +03:00
• 5472a14544 openpilot compile2 (#1977) 3 people 2023-10-15 20:39:46 -07:00
• 566660675c bugfix for cuda warning (#2078) George HotzandGitHub 2023-10-15 18:35:35 -07:00
• 0d3410d93f Stable diffusion: Make guidance modifiable (#2077) Ahmed HarmoucheandGitHub 2023-10-15 23:36:43 +02:00
• 776605f2fc O(1) VALIDHACKS (#2072) Umut ZenginandGitHub 2023-10-15 21:26:41 +03:00
• 30933d5bd0 if support (#2076) George HotzandGitHub 2023-10-15 07:17:37 -07:00
• cb9309bee6 remove temp files (#2075) nimlgenandGitHub 2023-10-15 16:45:36 +03:00
• 49bcfec383 0s in the action space (#2070) George HotzandGitHub 2023-10-14 11:22:48 -07:00
• 4124cf1df5 cleanup tensor cores, expose exclude local upcast (#2064) George HotzandGitHub 2023-10-14 09:21:03 -07:00
• 6e4a12ab68 HIPProgram __del__ (#2058) David HouandGitHub 2023-10-13 17:15:46 -07:00
• 91168a28c4 whisper: make file transcription work, add basic CI test (#2042) mmmkkaaayyandGitHub 2023-10-13 17:13:35 -07:00
• 924ecc4d6a Revert "openpilot kernel fix from 209 to 207 (#2006)" (#2065) George HotzandGitHub 2023-10-13 12:01:55 -07:00
• 63869c62fc openpilot kernel fix from 209 to 207 (#2006) Amrit SahuandGitHub 2023-10-14 00:29:15 +05:30