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

• 68053d0510 dsp stuff / sniff ioctls from snpe (#9490) 3 people 2025-03-20 10:38:23 +08:00
• 2223b93338 add UPat.or_casted [pr] (#9513) qazalandGitHub 2025-03-20 10:08:32 +08:00
• 1839e8c9b3 place masks in INDEX for TestGatedStoreRewrite [pr] (#9512) qazalandGitHub 2025-03-20 09:46:53 +08:00
• bd731a8624 AMDCompiler refactor (no_comgr prereq) (#9497) 3 people 2025-03-20 09:44:07 +08:00
• 8c0d0a122c Add return_indices to max_pool (#9506) geohotstanandGitHub 2025-03-20 03:25:37 +08:00
• 189f62d44f add rounding to tqdm unit scale (#9507) chenyuandGitHub 2025-03-19 12:08:46 -04:00
• a5c971ff3a am: prereqs for rdna4 1/n (#9495) nimlgenandGitHub 2025-03-19 17:14:57 +08:00
• 1e5d9ad8f7 extra/gemm/max_matmul: start of custom kernels for GEMM (#6926) Francis LamandGitHub 2025-03-19 00:04:57 -07:00
• 865f23dd7b olmoe memory usage cleanups geohot 2025-03-19 12:28:18 +08:00
• 2c87a22cf2 fix prg size calculation when there are adjacent mapped ranges (#9498) 3 people 2025-03-19 11:55:03 +08:00
• 1d71436e6a use libllvm19 in ci (#9494) 3 people 2025-03-19 11:53:32 +08:00
• a95b489a55 nanoGPT train works with tiny torch backend (#9283) 3 people 2025-03-19 11:51:02 +08:00
• f8976dd2eb enable more webgpu tests (#9502) chenyuandGitHub 2025-03-18 23:03:54 -04:00
• ae688e4103 simple failing test for scheduling parallel reduce [pr] (#9501) qazalandGitHub 2025-03-19 10:52:13 +08:00
• e4dad99145 nn.state docs cleanup (#8332) 4 people 2025-03-18 22:16:40 +01:00
• 1ea4876dfa olmoe touchups (#9499) chenyuandGitHub 2025-03-18 15:25:45 -04:00
• f7506c6c25 JIT OLMoE (#9396) geohotstanandGitHub 2025-03-19 02:49:02 +08:00
• 5c56cac0a0 MI300 mfma support (#9417) Ignacio SicaandGitHub 2025-03-18 14:33:30 -03:00
• 5500887eed improve reproducibility of WebGPU CI puppeteer test (#9496) hoovedandGitHub 2025-03-18 09:27:38 -04:00
• c85dfe1500 improve verify_sink_dims verify_sink_dims_clean geohot 2025-03-18 19:38:22 +08:00
• cde4fd3be3 do not view_left assign + elementwise sources always have a shape [pr] (#9491) qazalandGitHub 2025-03-18 17:42:51 +08:00
• 117b7a16ef VALIDATE_WITH_CPU [pr] (#9488) George HotzandGitHub 2025-03-18 15:15:04 +08:00
• 935cd01f56 simple failing test for graph_rewrite children [pr] (#9489) qazalandGitHub 2025-03-18 13:07:21 +08:00
• d20494e6d7 move buffer logic to Buffer [pr] (#9487) George HotzandGitHub 2025-03-18 11:21:21 +08:00
• 3be228182f unbind Tensor variables last [pr] (#9486) qazalandGitHub 2025-03-18 09:52:01 +08:00
• b44f9c409a reorder do_realize [pr] (#9485) qazalandGitHub 2025-03-18 09:30:10 +08:00
• a82c9332d3 am: rename soc21 to soc (#9482) nimlgenandGitHub 2025-03-18 08:54:26 +08:00
• b100fc0b20 split the rule that uses context in scheduler simplifier [pr] (#9484) qazalandGitHub 2025-03-18 08:12:26 +08:00
• 5e58f4b65b Tiny backend test_ops fix part 3 (#9483) Anish UmaleandGitHub 2025-03-18 03:31:51 +05:30
• 9fcef4d009 add masked_select to tensor.py (#9468) 3 people 2025-03-17 13:05:36 -07:00
• 4f8eac59ea failed test case for threefry (#9469) chenyuandGitHub 2025-03-17 14:52:10 -04:00
• 6dd8e5ba7c refactor llvm compiler (#9403) 4 people 2025-03-18 00:13:49 +08:00
• 53d6f1e1bb Add bitonic cat sort (#9422) 3 people 2025-03-18 00:01:23 +08:00
• f53be010d7 lower bert learning rate (#9481) chenyuandGitHub 2025-03-17 10:49:56 -04:00
• e03c0aacf2 more explicit DONT_PUSH_VIEWS [pr] (#9479) qazalandGitHub 2025-03-17 20:43:21 +08:00
• 8eb9093fb8 lil geohot 2025-03-17 19:57:15 +08:00
• 45f7c08111 work geohot 2025-03-17 19:22:12 +08:00
• 3b00a778ba fix view_left for unsafe pad ops [pr] (#9478) qazalandGitHub 2025-03-17 19:02:02 +08:00
• 58fc77fdb3 improve render geohot 2025-03-17 18:50:44 +08:00
• e57258b17b prettier rendering geohot 2025-03-17 18:46:25 +08:00
• 31cd00e72f fix name get geohot 2025-03-17 18:09:39 +08:00
• b00ccc08c3 ms target geohot 2025-03-17 17:49:48 +08:00
• 94d578aec5 gep pushing geohot 2025-03-17 17:43:02 +08:00
• 45010f7eff Revert "dont do that" geohot 2025-03-17 17:26:00 +08:00
• 249141026e dont do that geohot 2025-03-17 17:15:59 +08:00
• 813f713edc merge_views for buffer ops + create valids last (#9472) qazalandGitHub 2025-03-17 17:15:44 +08:00
• a913c1aab7 multi unroll geohot 2025-03-17 17:12:45 +08:00
• bd1f71c1e2 simple failing test for extra ops in VALID [pr] (#9474) qazalandGitHub 2025-03-17 17:02:40 +08:00
• 469ec6b6b4 support tuple in beam geohot 2025-03-17 17:02:32 +08:00
• e26caf4c3a hotfix: skip test_mean_half_precision_underflow on amd ci (#9476) qazalandGitHub 2025-03-17 16:47:48 +08:00
• 1a84d504b7 Merge branch 'master' into dsp_search George HotzandGitHub 2025-03-17 16:43:07 +08:00
• 14c9f14125 dsp beam search geohot 2025-03-17 16:42:32 +08:00
• 824c5f41ac dsp work try 3 (#9475) George HotzandGitHub 2025-03-17 16:42:12 +08:00
• cc0041cb8c padding geohot 2025-03-17 16:30:29 +08:00
• e4615e0cd9 dsp work try 3 geohot 2025-03-17 16:20:46 +08:00
• 242daa4f9a ptrcat (#9473) George HotzandGitHub 2025-03-17 16:06:37 +08:00
• 52ae9af4dd Fast DSP for MobileNetV2 (try 2) (#9467) George HotzandGitHub 2025-03-17 15:10:36 +08:00
• 15ee742afa add get_children_map to uop (#9470) George HotzandGitHub 2025-03-17 14:36:13 +08:00
• d2cfbd8a4d bert lower learning rate and total steps (#9466) chenyuandGitHub 2025-03-16 17:21:20 -04:00
• 09e7708b49 minimum change for rdna4 [pr] (#9455) George HotzandGitHub 2025-03-16 13:39:24 +08:00
• be2161652b reorder into swizzler + ast_fixup [pr] (#9456) qazalandGitHub 2025-03-15 09:00:14 +01:00
• cb7a7f69c7 quantization preprocessor from DSP, should be universal (#9437) George HotzandGitHub 2025-03-15 07:49:37 +08:00
• ca5064a5b6 remove Kernel.float4_axis [pr] (#9448) chenyuandGitHub 2025-03-14 17:54:32 -04:00
• 0e591baf43 redo simple_matmul change (#9450) chenyuandGitHub 2025-03-14 17:53:52 -04:00
• b0f63d3c04 Revert "simple_matmul.py uses np to generate random (#9438)" (#9449) chenyuandGitHub 2025-03-14 17:14:22 -04:00
• 71b9bf1c58 Revert "simple_matmul.py uses np to generate random (#9438)" revert-9438-simple_matmul_np_rand chenyuandGitHub 2025-03-14 17:13:44 -04:00
• 14018050c1 simple_matmul.py uses np to generate random (#9438) Ignacio SicaandGitHub 2025-03-14 17:36:50 -03:00
• 2a50e6440d filter sink by DONT_PUSH_VIEWS + remove extra base [pr] (#9446) qazalandGitHub 2025-03-14 21:27:46 +01:00
• 3af7a08a06 ast_fixup in one graph_rewrite pass [pr] (#9444) qazalandGitHub 2025-03-14 20:14:31 +01:00
• bd4ae5ac53 am: hotfix: import modules (#9443) nimlgenandGitHub 2025-03-15 03:10:18 +08:00
• 77a8430616 am: use smu based on discovery (#9441) nimlgenandGitHub 2025-03-15 02:10:45 +08:00
• 5ff90cb261 am: less magic values (#9440) uuuvnandGitHub 2025-03-14 23:10:35 +05:00
• 459d0cd14f add arch to AMDRenderer and HIPRenderer (#9431) Ignacio SicaandGitHub 2025-03-13 13:06:27 -03:00
• 357e364ab8 am: turn off unord dispatch (#9433) nimlgenandGitHub 2025-03-13 23:59:28 +08:00
• 99b0287e4e add GROUP and GROUPTOP to test_arange (#9432) chenyuandGitHub 2025-03-13 11:28:38 -04:00
• 90ffa9bd45 swizzle without buffer ops try 2 [pr] (#9427) qazalandGitHub 2025-03-13 11:00:40 +02:00
• 4df2b6347d hotfix: bump tinybox red training CI timeout to 30 minutes (#9426) qazalandGitHub 2025-03-13 10:31:44 +02:00
• 931436204c hotfix: 12000 lines, for AMD stuff geohot 2025-03-13 10:48:14 +08:00
• bfc68d1953 add gep rules to simplify (#9419) George HotzandGitHub 2025-03-13 09:46:25 +08:00
• 0bed9b6cd2 benchmark huggingface onnx models (#8493) 3 people 2025-03-13 08:13:12 +08:00
• 4992958dae update bert beam params (#9423) chenyuandGitHub 2025-03-12 13:00:41 -04:00
• 12978f0d05 reorder contiguous/assign ast rules [pr] (#9420) qazalandGitHub 2025-03-12 13:13:27 +02:00
• 5f6d5b057d expand index isn't grouping by access size (#9418) George HotzandGitHub 2025-03-12 17:24:10 +08:00
• 815ad0b7a8 support load/store grouping in DEVECTORIZE=0 (#9409) George HotzandGitHub 2025-03-12 11:34:37 +08:00
• 4714c4f9ad torch backend multigpu - add devices and tests (#9414) Priyank PatelandGitHub 2025-03-11 20:33:11 -07:00
• 22fc0a2e36 bert sum acc in half (#9412) chenyuandGitHub 2025-03-11 23:03:15 -04:00
• f995b465b8 am: set doorbell offsets to nb (#9413) nimlgenandGitHub 2025-03-12 10:35:47 +08:00
• 95e0f069be hotfix: gitignore *.log [pr] (#9410) qazalandGitHub 2025-03-11 22:39:19 +02:00
• 78ebade125 Merge pull request #9408 from nimlgen/hcq_progress_during_wait nimlgenandGitHub 2025-03-11 19:40:23 +08:00
• e174c6c3bc new devectorizer (#9331) George HotzandGitHub 2025-03-11 18:47:56 +08:00
• 69fac5fe89 Merge pull request #9407 from tinygrad/no_const_after_sym qazalandGitHub 2025-03-11 12:24:09 +02:00
• 4d09ea4c06 hcq: reset timer on progress in singal.wait nimlgen 2025-03-11 10:02:07 +00:00
• fa69fd3afc no const/view in schedule sink after sym [pr] qazal 2025-03-11 10:58:23 +01:00
• 68f062c8be cast_vec on transcendental (#9406) George HotzandGitHub 2025-03-11 17:30:46 +08:00
• e85001b6ee SQTT profiling (#9278) 3 people 2025-03-11 10:19:56 +05:00
• 2780e2027e devectorize prereqs [pr] (#9404) George HotzandGitHub 2025-03-11 12:33:29 +08:00
• beed00eabe fix torch backend memory leak (#9395) 3 people 2025-03-10 19:48:20 -07:00
• 01e8b60911 acc_dtype -> dtype (#9402) chenyuandGitHub 2025-03-10 16:05:30 -04:00
• 59dfb234eb replace hardcoded ast with tensors in TestSwizzle [pr] (#9401) qazalandGitHub 2025-03-10 20:33:57 +02:00
• 796c3bbb23 torch: support in-place operations on views (#9371) 3 people 2025-03-10 08:29:00 -07:00