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

• d443e91d82 remove custom splits in Tensor.shard [pr] (#8602) chenyuandGitHub 2025-01-13 21:29:13 -05:00
• 227d96d7a3 remove unused src from metaop [pr] (#8601) chenyuandGitHub 2025-01-13 20:28:14 -05:00
• c4e33048c6 test Tensor.clone has a different lazydata [pr] (#8600) chenyuandGitHub 2025-01-13 20:13:44 -05:00
• ae2229d727 assert kernel buffer limit at compile time [pr] (#8595) qazalandGitHub 2025-01-13 16:32:07 -05:00
• c2504357af am: lock to access dev (#8594) nimlgenandGitHub 2025-01-13 23:53:13 +03:00
• 4abe631b56 fix onnx mobilenetv2-7-quantized.onnx (#8574) geohotstanandGitHub 2025-01-14 01:25:06 +08:00
• d19c1c7f03 bump 75 -> 73 for test failure geohot 2025-01-13 09:18:38 -08:00
• c25d5d3101 improve isin checks (#8589) Francis LataandGitHub 2025-01-13 12:12:31 -05:00
• 74b83c4c41 am in ci (#8532) nimlgenandGitHub 2025-01-13 19:55:17 +03:00
• d224d0ed7f nv: fix fault info (#8587) nimlgenandGitHub 2025-01-13 14:38:43 +03:00
• 586e730d32 use UOp.st for kernel reduce axes (#8499) qazalandGitHub 2025-01-13 06:24:11 -05:00
• 7562cc0399 better test for reduce swizzle + don't use double dtype [pr] (#8586) qazalandGitHub 2025-01-13 05:02:21 -05:00
• df59b072db rename to top_down_rewrite [pr] (#8583) George HotzandGitHub 2025-01-12 18:36:38 -08:00
• 994944920b simpler batch_load_train_bert [pr] (#8582) chenyuandGitHub 2025-01-12 20:25:05 -05:00
• 05e5de6a91 ugh, remove that binary blob geohot 2025-01-12 17:02:28 -08:00
• 4ac4c1415a free intermediate buffers in the jit [pr] (#8581) George HotzandGitHub 2025-01-12 15:41:41 -08:00
• d817dc10db start on test rewrite map [pr] (#8432) 3 people 2025-01-12 10:13:51 -08:00
• 2f71a00236 remove PYTHONPATH=. from mypy ci [pr] (#8578) qazalandGitHub 2025-01-12 12:52:03 -05:00
• cde18fddce fix DEBUG=2 output for copy runners [pr] (#8579) qazalandGitHub 2025-01-12 12:03:01 -05:00
• 867004fbeb use unravel in views_to_indexed_uops [pr] (#8560) 3 people 2025-01-12 15:25:55 +00:00
• 38b5ac4d4a mypy for mockgpu/cuda & dsp/run (#8575) nimlgenandGitHub 2025-01-12 18:25:39 +03:00
• def90b22f6 EVAL_BS=36 for bert [pr] (#8576) chenyuandGitHub 2025-01-12 09:43:56 -05:00
• ae241e96db fix half4 on qcom and gpu (#8573) qazalandGitHub 2025-01-12 06:23:05 -05:00
• cff1ee9038 add SINK folding from the tensor_map branch [pr] (#8562) qazalandGitHub 2025-01-12 03:39:34 -05:00
• 87cbff3ac0 always give BUFFER uops Buffers [pr] (#8572) qazalandGitHub 2025-01-11 16:17:09 -05:00
• 98c9e23560 remove global PYTHONPATH setting in CI (test.yml) [pr] (#8568) qazalandGitHub 2025-01-11 12:47:50 -05:00
• 815c505e1d fixes from adapting tvm tests (#8570) geohotstanandGitHub 2025-01-12 00:38:36 +08:00
• 79738d768c do not require PYTHONPATH=. for process replay [pr] (#8567) qazalandGitHub 2025-01-11 09:45:34 -05:00
• a70d1bf439 move print_diff to process replay [pr] (#8566) qazalandGitHub 2025-01-11 09:28:45 -05:00
• 2f0856c1e2 qcom: use hwinterface (#8565) nimlgenandGitHub 2025-01-11 17:11:23 +03:00
• 60503c8621 use CAPTURE_PROCESS_REPLAY=1 in CI [pr] (#8564) qazalandGitHub 2025-01-11 06:03:48 -05:00
• 61665a63c9 am logs to debug2 (#8563) nimlgenandGitHub 2025-01-11 13:33:18 +03:00
• c7acd40574 more aggressive onnx const creation [pr] (#8561) George HotzandGitHub 2025-01-10 17:38:32 -08:00
• 8891495996 minor arg spec check on wmma (#8525) ignaciosicaandGitHub 2025-01-10 20:42:56 -03:00
• d09897c2aa allow double copy [pr] (#8559) chenyuandGitHub 2025-01-10 18:21:01 -05:00
• 70fa65cd95 viz fixups + scheduler option [pr] (#8557) George HotzandGitHub 2025-01-10 15:09:31 -08:00
• f457cb64d6 am: do not reload fw each run (#8466) nimlgenandGitHub 2025-01-10 23:33:38 +03:00
• 337328e409 am: fini gpu after use (#8556) nimlgenandGitHub 2025-01-10 21:02:34 +03:00
• 6a7f971fa0 hotfix max(DEBUG, 2) -> max(DEBUG.value, 2) [pr] (#8553) chenyuandGitHub 2025-01-10 12:57:44 -05:00
• cd4edc5206 hotfix: pylint ignores runtime for speed geohot 2025-01-10 09:07:18 -08:00
• 92b59c9b7a test_hcq limits for mockgpu not (only) ci (#8555) nimlgenandGitHub 2025-01-10 17:37:28 +03:00
• 9833fe83d8 more work on onnx imagenet [pr] (#8552) George HotzandGitHub 2025-01-09 20:28:18 -08:00
• e172b759f0 more working (#8550) George HotzandGitHub 2025-01-09 18:40:08 -08:00
• 2cbb34535c simpler allreduce script [pr] (#8551) chenyuandGitHub 2025-01-09 21:38:13 -05:00
• 23c56817d8 update and clean up allreduce script [pr] (#8549) chenyuandGitHub 2025-01-09 19:35:28 -05:00
• 5720871903 onnx consts are const [pr] (#8548) George HotzandGitHub 2025-01-09 16:09:22 -08:00
• 88661cd96f fix checking DiskBuffer is opened [pr] (#8547) chenyuandGitHub 2025-01-09 18:58:36 -05:00
• 62447c253d viz cleanups [pr] (#8498) George HotzandGitHub 2025-01-09 15:46:48 -08:00
• 299d333806 Add QLinearConv, QLinearMatMul, QLinearAdd, QLinearGlobalAveragePool to onnx (#8478) 3 people 2025-01-10 07:08:53 +08:00
• 2fd068ffc0 delete empty op (#8544) qazalandGitHub 2025-01-09 14:10:15 -05:00
• f6eb0574f2 start tests for putting the tensor graph in a single kernel [pr] (#8542) qazalandGitHub 2025-01-09 13:33:21 -05:00
• 83a8217cbf hotfix: TRACK_MATCH_STATS=2 should not launch viz [pr] (#8543) qazalandGitHub 2025-01-09 11:10:15 -05:00
• 1efb1188d8 support pickling a realized BUFFER uop [pr] (#8541) qazalandGitHub 2025-01-09 06:37:22 -05:00
• 7595352dfc refactor buffer_view op structure [pr] (#8540) qazalandGitHub 2025-01-09 03:07:46 -05:00
• 4c5c32ff5f Small bug in _reshape_mask (#8538) eliotgoldingandGitHub 2025-01-09 03:11:24 +00:00
• aa3d612df2 add script to install amd mockgpu on macOS (#8536) nimlgenandGitHub 2025-01-09 01:29:25 +03:00
• 31fcfe764d adjust hcq test for ci macos (#8534) nimlgenandGitHub 2025-01-08 16:18:31 +03:00
• 49abe6d3a6 little more compact tensor_uop_spec [pr] (#8533) qazalandGitHub 2025-01-08 08:01:53 -05:00
• 21c7d7c71a MOCKGPU amd test on OSX (#8505) patrini32andGitHub 2025-01-08 14:27:56 +03:00
• 2f530adb04 hwiface: close fd when valid (#8530) nimlgenandGitHub 2025-01-08 10:43:59 +03:00
• 947de23cac add VIEW(DEVICE) to tensor variable [pr] (#8529) qazalandGitHub 2025-01-08 01:39:42 -05:00
• b22494b710 restrict tensor const ShapeTracker in spec [pr] (#8447) qazalandGitHub 2025-01-07 19:05:11 -05:00
• afef69a37d MOCKGPU on mac os (#8520) patrini32andGitHub 2025-01-07 20:27:43 +03:00
• ab3ac2b58d hw interface abstraction (#8524) 3 people 2025-01-07 18:18:28 +03:00
• 0e97f807e0 test fixup prereqs for delete_buffer_view [pr] (#8523) qazalandGitHub 2025-01-07 11:52:18 +02:00
• 85a4397f27 fix create_schedule_with_vars usage in allreduce benchmark [pr] (#8522) chenyuandGitHub 2025-01-07 01:30:01 -05:00
• 0061dc7447 fix benchmark allreduce and add to ci [pr] (#8521) chenyuandGitHub 2025-01-07 00:37:59 -05:00
• c69f459c96 Add checking variable dimension to onnx (#8518) 3 people 2025-01-07 13:30:35 +08:00
• 5cb9443ebb PROFILE is enabled when VIZ is enabled (#8516) nimlgenandGitHub 2025-01-06 19:47:16 +03:00
• ed618a72e7 do not use subbuffer for bitcast (#8514) qazalandGitHub 2025-01-06 18:40:46 +02:00
• 280143467b am: tune all sleep timings to match kernel (#8515) nimlgenandGitHub 2025-01-06 18:03:57 +03:00
• 547fd5078f cleanups for COPY uop implementation and spec [pr] (#8513) qazalandGitHub 2025-01-06 11:39:12 +02:00
• ed121d235c spec for CAST_BEFORE_VIEW=1 [pr] (#8512) qazalandGitHub 2025-01-06 10:43:58 +02:00
• eb7df92136 dedup COPY UOp [pr] (#8506) qazalandGitHub 2025-01-06 10:37:20 +02:00
• 76a138cdb6 simpler UOp.st [pr] (#8510) chenyuandGitHub 2025-01-05 22:08:14 -05:00
• b6be407bc6 fix handcode_opt bert [pr] (#8509) chenyuandGitHub 2025-01-05 19:14:12 -05:00
• 9229867fec Support asymmetrical pads for all pooling functions (#8109) 3 people 2025-01-06 05:01:08 +08:00
• c9c7f1be46 Remove unused R_AARCH64_CALL26 relocation (#8508) uuuvnandGitHub 2025-01-05 23:00:21 +02:00
• b4f4a3ac12 am: minor parts (#8507) nimlgenandGitHub 2025-01-05 23:05:21 +03:00
• 0e0cba2cfc move llvm_bf16_cast to the renderer [pr] (#8502) qazalandGitHub 2025-01-05 13:02:41 +02:00
• 4143f6a7d9 unused from __future__ import annotations [pr] (#8504) chenyuandGitHub 2025-01-04 23:11:01 -05:00
• 9bc317d5d2 mockcuda (#8503) nimlgenandGitHub 2025-01-05 01:23:57 +03:00
• ddad4d55da add typing to tqdm [pr] (#8500) George HotzandGitHub 2025-01-04 13:55:52 -05:00
• 036efa9157 use UOp.substitute for VIZ=1 [pr] (#8497) qazalandGitHub 2025-01-04 20:00:29 +02:00
• 615d5276b1 Suppress 'X warnings generated.' in MTLCompiler (#8489) uuuvnandGitHub 2025-01-04 17:22:37 +02:00
• 5df213d51e am: remove alloc frags logic (#8491) nimlgenandGitHub 2025-01-04 12:25:20 +03:00
• 3dfc8e1706 Share a _resolve_pool_pads function for pool ops in Tensor (#8485) geohotstanandGitHub 2025-01-04 12:54:11 +08:00
• 6c639dee5c more informative kernel opt error messages [pr] (#8487) chenyuandGitHub 2025-01-03 14:29:36 -05:00
• 5ffc50d58c Clang JIT (#8481) 3 people 2025-01-03 18:12:55 +02:00
• 12fa4340b3 pickle ContextVars in process replay [pr] (#8484) qazalandGitHub 2025-01-03 17:11:54 +02:00
• bd4d7dc4eb return becomes_map from the scheduler (#8483) qazalandGitHub 2025-01-03 16:47:21 +02:00
• c163b2c5f0 give copy a device: COPY(device, copyin) (#8482) qazalandGitHub 2025-01-03 16:34:38 +02:00
• 0d33391038 delete unused allow_buffer_view=True arg from bitcast [pr] (#8462) qazalandGitHub 2025-01-03 16:20:46 +02:00
• 5d37d33fc5 update typing.Optional to 3.10 for hcq (#8479) nimlgenandGitHub 2025-01-03 16:20:49 +03:00
• 048643e7f9 Skip test that counts Ops.LOAD on CLANG+AMX (upcasts up to float16) (#8475) uuuvnandGitHub 2025-01-03 00:17:49 +02:00
• de306c615b [fixed] onnx pool cleanup (#8474) geohotstanandGitHub 2025-01-03 05:56:10 +08:00
• 08c9d980dc use const_like in uop zero folding [pr] (#8470) qazalandGitHub 2025-01-02 19:05:09 +02:00
• 6fa38367bf Revert "onnx pool ops clean up (#8471)" (#8472) chenyuandGitHub 2025-01-02 11:04:34 -05:00
• e7c6282dd6 Fix uop.st for CLANG+AMX (#8460) uuuvnandGitHub 2025-01-02 18:01:41 +02:00
• 241db29ede onnx pool ops clean up (#8471) geohotstanandGitHub 2025-01-02 23:45:30 +08:00