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

• 4cc7422769 use AM driver in bert mlperf (#9775) chenyuandGitHub 2025-04-07 23:40:27 -04:00
• db22094d35 hotfix: update softmax fusion test geohot 2025-04-08 11:22:12 +08:00
• f8fe15e64e move BoxCoder to mlperf helpers (#9773) Francis LataandGitHub 2025-04-07 20:27:06 -04:00
• bb7922b95f Vectorize Transcendental Regression Tests (#9753) Eitan TurokandGitHub 2025-04-07 20:27:39 +03:00
• 7c4a739fe4 full script for bert mi300x (#9772) chenyuandGitHub 2025-04-07 11:41:31 -04:00
• 07d1aefaf4 fast idiv (#9755) 3 people 2025-04-07 14:32:24 +02:00
• fa888ee077 minor test cleanups (#9770) nimlgenandGitHub 2025-04-07 15:29:12 +03:00
• 3069ebfad1 use BERT_LAYERS=2 in bert init (#9769) chenyuandGitHub 2025-04-07 07:46:37 -04:00
• 891322fd51 split into grouper.py (#9768) qazalandGitHub 2025-04-07 18:40:59 +08:00
• 219b8c9e8b return becomes_map in scheduler [pr] (#9766) qazalandGitHub 2025-04-07 17:36:23 +08:00
• 6306dea6e2 add a graph_rewrite pass for creating asts [pr] (#9765) qazalandGitHub 2025-04-07 16:32:11 +08:00
• 07eea567d4 reorder tensor_map and grouper parts [pr] (#9764) qazalandGitHub 2025-04-07 15:36:13 +08:00
• 8ddb1357c0 fix UPat.location after pickle (#9763) qazalandGitHub 2025-04-07 15:16:42 +08:00
• 4cd27aa0e6 hotfix: viz recenter and unlimited zoom (#9760) qazalandGitHub 2025-04-07 14:38:03 +08:00
• d0dace4306 update doc for permute to 3d tensor (#9758) chenyuandGitHub 2025-04-07 00:38:05 -04:00
• b190d85ad7 benchmark script bert softmax (#9759) chenyuandGitHub 2025-04-07 00:31:18 -04:00
• 58785181a8 AMD bf16xf32 TC (#9717) Ignacio SicaandGitHub 2025-04-07 11:41:04 +08:00
• 43e4565148 weighted linear in external_benchmark_bert_matmuls (#9757) chenyuandGitHub 2025-04-06 23:35:42 -04:00
• 28e06d2d44 minor cleanups from patternmatcher [pr] (#9756) George HotzandGitHub 2025-04-07 11:28:14 +08:00
• 1ce4912770 viz profiler ui (#9664) qazalandGitHub 2025-04-07 00:30:17 +08:00
• 8a585dc5c1 benchmark script for matmuls in bert (#9752) chenyuandGitHub 2025-04-06 07:34:25 -04:00
• 139999c6d7 map viz files + query params cleanup [pr] (#9754) qazalandGitHub 2025-04-06 16:20:00 +08:00
• 71b8890dd6 use validation dataloader inside retinanet eval (#9747) Francis LataandGitHub 2025-04-05 16:46:55 -04:00
• 5f7c79676f jit: prune independent copies (#9749) nimlgenandGitHub 2025-04-05 20:50:28 +03:00
• c2573b247c jit: rename optimize_weights -> replan_buffers_memory_layout (#9751) nimlgenandGitHub 2025-04-05 20:35:15 +03:00
• 493fb315b1 fix RDNA2 support (#9700) uuuvnandGitHub 2025-04-05 20:42:47 +05:00
• 5a04f4d4ba revert bert hparams for green and red (#9744) chenyuandGitHub 2025-04-05 07:38:01 -04:00
• 407ca54382 symbolic fold double where (#9436) chenyuandGitHub 2025-04-05 05:12:17 -04:00
• 9c2fc695b5 cond.logical_not().where(a,b) -> cond.where(b,a) (#9741) 3 people 2025-04-05 01:13:32 +02:00
• e9a3ac02a5 Remove ne from arange pattern (#9743) Sieds LyklesandGitHub 2025-04-05 00:31:13 +02:00
• 86c55414d7 ops_amd: simplify gfx version (#9742) nimlgenandGitHub 2025-04-04 22:18:11 +03:00
• 16d6aa15f1 record unittest name in process replay (#9731) qazalandGitHub 2025-04-05 01:39:48 +08:00
• 354db961c6 viz refactor to prep for profiler [pr] (#9739) qazalandGitHub 2025-04-04 17:13:14 +08:00
• fe998798fb linearizer failure test for OUT OF BOUNDS ACCESS (#9738) chenyuandGitHub 2025-04-04 03:48:43 -04:00
• 8b5a523743 fix minimum length in pattern matcher (#9736) George HotzandGitHub 2025-04-04 14:57:01 +08:00
• 640ff681c3 rename bert script to 8xMI300X (#9734) chenyuandGitHub 2025-04-03 23:36:24 -04:00
• b719aa1fb0 only check once for divisible fold lengths (#9732) George HotzandGitHub 2025-04-04 11:27:34 +08:00
• 926b0bcc57 cache folded upcast [pr] (#9733) George HotzandGitHub 2025-04-04 11:23:19 +08:00
• 8206c7281e move const multiply after REDUCE (#9730) George HotzandGitHub 2025-04-04 11:07:46 +08:00
• 6b3480ec70 update mi300x bert haparams (#9716) chenyuandGitHub 2025-04-03 22:30:00 -04:00
• cac8bcf8b5 use Ops.REDUCE (#9721) George HotzandGitHub 2025-04-04 10:14:34 +08:00
• 949459fdd6 jit: fix deallocate on unallocated buffers in free_intermediates (#9699) nimlgenandGitHub 2025-04-03 22:32:51 +07:00
• 52a8ecb15e record unittest location in process replay [pr] (#9727) qazalandGitHub 2025-04-03 20:50:09 +08:00
• ac713e04db ONNX add output shape validation (#9720) geohotstanandGitHub 2025-04-03 17:44:53 +08:00
• 7dadbf3697 insert float() in bert acc (#9726) chenyuandGitHub 2025-04-03 05:44:09 -04:00
• 79145e3d40 cleanup truncate_bf16 [pr] (#9725) chenyuandGitHub 2025-04-03 05:43:49 -04:00
• bc2d86195e increase test tolerance (#9719) Ignacio SicaandGitHub 2025-04-03 15:24:09 +08:00
• 1d25844d44 Revert "disable CI red llama 3 4 gpu beam (#9690)" (#9709) chenyuandGitHub 2025-04-03 02:34:39 -04:00
• 49dafe6d43 add gc tests [pr] (#9718) George HotzandGitHub 2025-04-03 14:08:32 +08:00
• bc91fffc5d fix gated store with index in python backend (#9703) Ignacio SicaandGitHub 2025-04-03 12:48:28 +08:00
• f2bd65ccfc delete Ops.EMPTY and Tensor._metaop (#9715) qazalandGitHub 2025-04-03 12:29:02 +08:00
• 5c7b549eab use functools.cache instead of lru_cache(None) [pr] (#9714) George HotzandGitHub 2025-04-03 11:47:13 +08:00
• bbd13191f4 cleanup tensor BIND + remove outdated comments in tensor.py [pr] (#9712) qazalandGitHub 2025-04-03 11:21:53 +08:00
• e1d7e47cca fix ONNX IsInf unintended dtype promotion (#9711) geohotstanandGitHub 2025-04-03 10:46:15 +08:00
• 11ae254dc5 construct BUFFER UOps directly when device in known [pr] (#9710) qazalandGitHub 2025-04-03 10:41:44 +08:00
• 1714fc3ba4 start work on speed [pr] (#9707) George HotzandGitHub 2025-04-03 10:39:01 +08:00
• 0f1ffc2050 hotfix: cat tests 2048 instead of 256 geohot 2025-04-03 10:37:56 +08:00
• 5bd485c027 Fix double SDMA_OP_FENCE (#9705) uuuvnandGitHub 2025-04-03 06:43:37 +05:00
• a6fec2f5ae dev_run for bert on mi300x (#9706) chenyuandGitHub 2025-04-02 21:12:55 -04:00
• d96b4983ac amd: support rdna4 in runtime again (#9702) nimlgenandGitHub 2025-04-03 01:19:23 +07:00
• 2d6d8b7355 add bf16 mfma support (#9695) Ignacio SicaandGitHub 2025-04-02 21:44:49 +08:00
• 07102624a1 Merge branch 'master' into dsp_search_merged dsp_search_merged George HotzandGitHub 2025-04-02 20:32:23 +08:00
• a6733f519f dsp: make relro sections contiguous (#9701) nimlgenandGitHub 2025-04-02 18:02:16 +07:00
• ed76dd71eb Merge branch 'master' into dsp_search_merged George HotzandGitHub 2025-04-02 18:11:25 +08:00
• ea5caefef0 gep should look at count, not vcount (#9698) George HotzandGitHub 2025-04-02 18:10:57 +08:00
• 5d6e8bd681 you can render where geohot 2025-04-02 17:37:40 +08:00
• f90656c647 cleanup dsp junk geohot 2025-04-02 17:29:15 +08:00
• 86727875f9 delete junk geohot 2025-04-02 16:50:37 +08:00
• 7e4ac744ac Merge branch 'master' into dsp_search_merged George HotzandGitHub 2025-04-02 16:45:20 +08:00
• f72a87fd0e add proper support for Ops.IGNORE to remove store masks (#9692) George HotzandGitHub 2025-04-02 16:38:01 +08:00
• 4496cc6e61 cleanup comments and flip rewrite geohot 2025-04-02 16:20:26 +08:00
• 3b8d923692 remove skip LLVM in test_div_int (#9686) chenyuandGitHub 2025-04-02 04:15:00 -04:00
• bc3bfcbad4 update install gpuocelot (#9693) chenyuandGitHub 2025-04-02 04:10:34 -04:00
• 13d3bcb6e1 fix for many geohot 2025-04-02 15:38:08 +08:00
• e78e8722dc Revert "LDS noop and spec (#9669)" (#9691) 3 people 2025-04-02 15:31:32 +08:00
• efad1ebd0d Merge branch 'master' into dsp_search_merged George HotzandGitHub 2025-04-02 15:29:42 +08:00
• 4514fd91c1 more stuff from DSP (#9689) George HotzandGitHub 2025-04-02 15:27:48 +08:00
• 6a5eacba8b disable CI red llama 3 4 gpu beam (#9690) chenyuandGitHub 2025-04-02 03:19:09 -04:00
• e20eed6208 touch ups geohot 2025-04-02 15:03:41 +08:00
• 7dc265ef93 syntax error geohot 2025-04-02 14:55:36 +08:00
• 13dec71ab0 syntax error geohot 2025-04-02 14:54:13 +08:00
• bb453613ac syntax error geohot 2025-04-02 14:53:04 +08:00
• d6c3ae186b syntax error geohot 2025-04-02 14:52:09 +08:00
• c066653428 optimize is an option geohot 2025-04-02 14:50:27 +08:00
• 66c6d35fe2 fix speed geohot 2025-04-02 14:41:37 +08:00
• 876a8be97a Debug env var breakdown (#9663) Ignacio SicaandGitHub 2025-04-02 14:34:07 +08:00
• 64e1ddf2a9 fix bad merge geohot 2025-04-02 13:21:55 +08:00
• d6013a2d50 faster geohot 2025-04-02 13:18:37 +08:00
• 1d36aa8116 Merge branch 'master' into dsp_search_merged George HotzandGitHub 2025-04-02 13:09:13 +08:00
• 6f812d3f2f fixes from the dsp branch + 12500 lines (#9683) George HotzandGitHub 2025-04-02 13:07:17 +08:00
• c20f112e9f example test use z3 to verify valid simplification (#9684) chenyuandGitHub 2025-04-02 01:05:52 -04:00
• bca0c85193 skip CI CPU test_data_parallel_resnet_train_step (#9685) chenyuandGitHub 2025-04-02 01:04:54 -04:00
• 6251ab3d90 Merge remote-tracking branch 'origin/master' into dsp_search_merged geohot 2025-04-02 12:25:01 +08:00
• bb94f13e58 add RECORD_TRACEBACKS=1 option to process replay (#9679) qazalandGitHub 2025-04-02 11:58:27 +08:00
• 3acc1b928a minor div_and_mod_folding cleanup [pr] (#9681) chenyuandGitHub 2025-04-01 23:51:36 -04:00
• c672716b38 improve vmin/vmax for IDIV (#9678) chenyuandGitHub 2025-04-01 23:16:01 -04:00
• dd51728795 fix all bugs dsp_search geohot 2025-04-02 11:07:32 +08:00
• a59b1ed970 bring speed back geohot 2025-04-02 10:21:08 +08:00
• 147fc0e648 minor l2fetch tweak geohot 2025-04-02 10:09:24 +08:00
• 17f7b226cb dsp masked stores are stores geohot 2025-04-02 09:41:56 +08:00