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

• cf73d7e2a7 hotfix: disable slower asm gemm shape from llama seqlen 8192 (#14582) qazalandGitHub 2026-02-06 01:05:19 -05:00
• be77873974 llama: contig backward for wk / wv matmul backward (#14581) qazalandGitHub 2026-02-06 00:54:00 -05:00
• 15d3344d9e use int inputs in test_assign (#14580) chenyuandGitHub 2026-02-06 00:07:31 -05:00
• 50a166a5fa viz: cleanup amdgpu target mapping (#14579) qazalandGitHub 2026-02-05 23:51:51 -05:00
• b09dc646f5 revert some late_buffer_view change (#14578) chenyuandGitHub 2026-02-05 22:51:40 -05:00
• d41836f135 remove KERNEL special case in realize_assign [pr] (#14573) chenyuandGitHub 2026-02-05 21:55:44 -05:00
• 6cbcf98627 KernelInfo is required on get_program (#14571) George HotzandGitHub 2026-02-06 10:49:27 +08:00
• 28c56a783c add CallInfo and viz call toggle (#14570) George HotzandGitHub 2026-02-06 09:30:58 +08:00
• 771a395240 Merge branch 'master' into kernel_is_call kernel_is_call George HotzandGitHub 2026-02-06 09:15:03 +08:00
• f73468d516 fa: block skipping for fa kv bwd (#14569) wozeparrotandGitHub 2026-02-05 16:13:53 -08:00
• b7ef775677 more cleanup in create_schedule [pr] (#14566) chenyuandGitHub 2026-02-05 16:12:17 -05:00
• cee7ef7ab2 disable threads (#14555) Garret CastroandGitHub 2026-02-05 13:11:32 -08:00
• 79b7799dba clean up linearize schedule [pr] (#14565) chenyuandGitHub 2026-02-05 15:24:09 -05:00
• 41a179f542 fix test_xlm_roberta_large (#14564) chenyuandGitHub 2026-02-05 14:56:06 -05:00
• aa9dc50577 dtype decomps don't require bitshifts (#14542) sirhcmandGitHub 2026-02-05 11:42:30 -08:00
• b47397ab17 list ml_dtypes as dependency for DSP (#14562) sirhcmandGitHub 2026-02-05 11:27:50 -08:00
• 2b47a9a1b5 skip test_xlm_roberta_large (#14563) chenyuandGitHub 2026-02-05 14:00:24 -05:00
• 42c18da88a add Ops asserts in toposort sched_sink [pr] (#14561) chenyuandGitHub 2026-02-05 12:40:02 -05:00
• 483bba4f05 nv: use prof_exec_counter (#14559) nimlgenandGitHub 2026-02-05 19:00:14 +03:00
• 190042358f llama: faster bf16 matmul / rope backward (#14558) qazalandGitHub 2026-02-05 09:57:25 -05:00
• b398335f62 assembly/amd: fix saturation in python remu (#14557) 3 people 2026-02-05 18:35:57 +08:00
• c1ea6687e5 fa: simpler is faster (#14548) wozeparrotandGitHub 2026-02-05 01:13:17 -08:00
• 43e7eda4e7 grad_b uses custom gemm (#14550) 4 people 2026-02-05 14:22:27 +08:00
• d83ddc05c8 resolve_call geohot 2026-02-05 12:57:31 +08:00
• f9cfb64cd9 test asm_gemm in CI (#14551) qazalandGitHub 2026-02-04 23:32:22 -05:00
• 8e8cac4b0f don't use tag, use KernelInfo geohot 2026-02-05 12:31:14 +08:00
• 57199fd9de keep the all buffers on same device check geohot 2026-02-05 12:17:32 +08:00
• 2193d0edfa fix arg order geohot 2026-02-05 12:03:45 +08:00
• 77adccb925 use call for kernel geohot 2026-02-05 11:48:50 +08:00
• c0ca7f9c51 use more UOp.sum and UOp.prod [pr] (#14549) chenyuandGitHub 2026-02-04 22:05:20 -05:00
• e8dace41b6 clean up UOp.vars [pr] (#14547) chenyuandGitHub 2026-02-04 20:52:25 -05:00
• 232848d086 PYTHONREMU: VOP3P integer operations with constants don't cast to fp16 (#14546) sirhcmandGitHub 2026-02-04 17:10:59 -08:00
• 2966619834 feat: llama uses enable_gqa during training (#14545) wozeparrotandGitHub 2026-02-04 16:22:31 -08:00
• 664f1bf76d minor ops/jit cleanups [pr] (#14543) chenyuandGitHub 2026-02-04 17:21:34 -05:00
• 03d0fa9c3f merge as_buf into buf_uop [pr] (#14541) chenyuandGitHub 2026-02-04 16:32:23 -05:00
• 43ef24a8af remove buf_target [pr] (#14540) chenyuandGitHub 2026-02-04 15:03:47 -05:00
• 8b7343b950 clean up is_realized [pr] (#14538) chenyuandGitHub 2026-02-04 14:24:10 -05:00
• 5338ce6b74 test S_PACK in extra/assembly/amd/test/hw (#14537) sirhcmandGitHub 2026-02-04 11:17:16 -08:00
• 9052db678f remove allow_shape_mismatch in Tensor.replace (#14536) chenyuandGitHub 2026-02-04 12:38:18 -05:00
• ec2b6bbda8 hcq: update signal logic (#14531) nimlgenandGitHub 2026-02-04 19:32:56 +03:00
• 62786d488a am: mi3xx perf (#14529) nimlgenandGitHub 2026-02-04 19:32:43 +03:00
• d57d24c7d4 Buffer.as_buffer -> Buffer.as_memoryview [pr] (#14535) chenyuandGitHub 2026-02-04 11:31:11 -05:00
• 6de9da8b9c Merge branch 'master' into new_x86_backend ttomsaandGitHub 2026-02-04 16:18:47 +00:00
• 74e3d9faf3 add min x86op and neg in decomps ttomsa 2026-02-04 16:00:53 +00:00
• 024f57ecf5 jit input_buffers cleanup [pr] (#14532) chenyuandGitHub 2026-02-04 10:14:38 -05:00
• 67f91e897b UOp.is_contiguous -> UOp.has_buffer_identity [pr] (#14530) chenyuandGitHub 2026-02-04 09:21:26 -05:00
• fb9df1e031 pretty print binary (#14520) George HotzandGitHub 2026-02-04 18:04:35 +08:00
• 8c3c026d86 decomp float16 to float32 (#14417) sirhcmandGitHub 2026-02-03 22:37:47 -08:00
• ecbce5269e PYTHONREMU properly supports S_PACK_LL_B32_B16 (#14527) sirhcmandGitHub 2026-02-03 20:45:33 -08:00
• 720c9597a9 feat: llama uses is_causal on sdpa during training (#14528) wozeparrotandGitHub 2026-02-03 20:24:30 -08:00
• 9c2fc118ef relax setitem target check (#14526) chenyuandGitHub 2026-02-03 22:32:49 -05:00
• d1bfbe9ce3 isolate slow llama gemm (#14525) qazalandGitHub 2026-02-03 22:20:10 -05:00
• 2f55005ad9 qcom: sync cpu cache when from_blob (#14518) nimlgenandGitHub 2026-02-03 21:51:03 +03:00
• bbe012ac86 more ttomsa 2026-02-03 16:50:30 +00:00
• 4c3081b613 switch to PARAM ttomsa 2026-02-03 16:45:35 +00:00
• 69a27d9a5c Merge remote-tracking branch 'upstream/master' into new_x86_backend ttomsa 2026-02-03 16:41:19 +00:00
• 77a28ac3f2 x86 goes in own linearize ttomsa 2026-02-03 16:32:28 +00:00
• ee9d6a1f36 remove DEFINE_VAR in to_define_global [pr] (#14522) chenyuandGitHub 2026-02-03 10:12:33 -05:00
• af4c74bb41 delete extra cast (#14517) Nino RisteskiandGitHub 2026-02-03 14:29:04 +01:00
• 9d1e9e643e removed a duplicated remove_bufferize rule [pr] (#14519) chenyuandGitHub 2026-02-03 08:28:07 -05:00
• d59e6e7a37 move more tests to test/null, split some existing ones (#14512) George HotzandGitHub 2026-02-03 20:20:20 +08:00
• a98c53769a ASM_GEMM=1 runs the UOp gemm on non cdna (#14516) qazalandGitHub 2026-02-03 06:42:02 -05:00
• 5c1d21349e viz: profiler command line tool (#14515) qazalandGitHub 2026-02-03 05:51:25 -05:00
• 308eb13eae codegen new_ck geohot 2026-02-03 18:09:15 +08:00
• 1b1d81e3d5 allow kernel_spec in tensor for custom kernels geohot 2026-02-03 18:01:07 +08:00
• 50b7b283dc allow after on param geohot 2026-02-03 17:41:54 +08:00
• 39da624581 new custom_kernel function in tensor geohot 2026-02-03 16:28:53 +08:00
• dd2de4f838 rename all DEFINE_GLOBAL to PARAM (#14511) George HotzandGitHub 2026-02-03 15:09:38 +08:00
• dc77b3318b move files that pass with NULL=1 to test/null (#14508) George HotzandGitHub 2026-02-03 13:52:36 +08:00
• 888819ee09 call autodiff gradient (#14510) George HotzandGitHub 2026-02-03 13:51:02 +08:00
• bbcd3d67a3 fa: faster (#14453) wozeparrotandGitHub 2026-02-02 21:34:17 -08:00
• e579613b90 IR3 has aux (#14509) sirhcmandGitHub 2026-02-02 20:46:41 -08:00
• 85c7b23160 add pytest -nauto to benchmark for mac (#14458) George HotzandGitHub 2026-02-03 12:26:09 +08:00
• a5d7eb37db IR3 works on versions earlier than 3.14 (#14507) sirhcmandGitHub 2026-02-02 20:10:19 -08:00
• 33c886cafa disable copyout on NULL backend by default (#14506) George HotzandGitHub 2026-02-03 11:57:47 +08:00
• 3c5845e8a5 remove cut_store_range (#14505) chenyuandGitHub 2026-02-02 21:58:36 -05:00
• 4f2e7aed24 fix multiple REDUCE on same RANGE (#14504) chenyuandGitHub 2026-02-02 20:42:09 -05:00
• 93c41a78fa clean up NOOP [pr] (#14503) chenyuandGitHub 2026-02-02 19:46:45 -05:00
• 66d2b02f11 delete files that depends on extra.optimization.helpers (#14499) chenyuandGitHub 2026-02-02 13:33:33 -05:00
• 66634c643e test unit should pass on NULL device test_unit geohot 2026-02-03 00:11:24 +08:00
• ec0398fceb test amd gpu crashes (#14459) George HotzandGitHub 2026-02-02 23:57:47 +08:00
• 6e4238c016 amd: recovery (#14461) nimlgenandGitHub 2026-02-02 18:57:35 +03:00
• 0ae5c5e4f9 func arg order independent from op value ttomsa 2026-02-02 15:47:49 +00:00
• 1019a3d8f8 Merge branch 'master' into mac_pytest mac_pytest George HotzandGitHub 2026-02-02 23:40:03 +08:00
• e9f2e89f8f Merge branch 'master' into new_x86_backend ttomsaandGitHub 2026-02-02 15:22:49 +00:00
• 61ca19ff24 after with empty src is self [pr] (#14496) chenyuandGitHub 2026-02-02 10:19:05 -05:00
• 6e958dbfd4 assembly/amd: add RDNA4 support to emulator (#14341) George HotzandGitHub 2026-02-02 21:35:59 +08:00
• a908f447d5 remove disk special case in mstack_early_shrink [pr] (#14494) chenyuandGitHub 2026-02-02 08:34:45 -05:00
• 965940dd00 sqtt: update examples after event field change (#14493) qazalandGitHub 2026-02-02 07:39:48 -05:00
• 8a3d451ee4 failing test from application user_failure geohot 2026-02-02 16:48:24 +08:00
• 965149a46d assembly/amd: add ds perm instructions (#14486) George HotzandGitHub 2026-02-02 16:02:00 +08:00
• 1746d1f997 remove SPEC=0 context in custom_kernel tests, pyrender always skips it (#14489) qazalandGitHub 2026-02-02 02:32:01 -05:00
• d4007f36e0 remove DEFINE_GLOBAL (it is PARAM now) (#14488) George HotzandGitHub 2026-02-02 14:56:37 +08:00
• 6c487656f9 viz: kernel metadata from rodata entry (#14487) qazalandGitHub 2026-02-02 01:41:42 -05:00
• d75a1b0d5a usbgpu: use BOT interface for patch.py (#13644) Robbe DerksandGitHub 2026-02-02 04:54:46 +01:00
• 2931b52875 skip autogen if MTLCompiler is loaded (#14466) sirhcmandGitHub 2026-02-01 19:12:27 -08:00
• b25d3b1431 less sketch tests test_amd_crash geohot 2026-02-02 03:07:27 +00:00
• 9a32d6e090 add depth limit for SPEC=2 (#14485) George HotzandGitHub 2026-02-02 10:43:28 +08:00
• 368a692e1a make SPEC=2 work for everything (#14476) George HotzandGitHub 2026-02-02 10:30:56 +08:00
• ea1f1d2b9d test_assign_to_bitcast_view (#14483) chenyuandGitHub 2026-02-01 16:46:04 -05:00