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

autogen-comgr-2

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

ci-deps

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

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

explicit-dev

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

gitea-test-ci

gl_dims

gldims

glm_flash

grad_outerworld

grok_x86_novec

group_div_rules

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-chat-template

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

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

onnx-cpu-count

outer_range

outerworld_work

pad_invalid

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

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-lp

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

view_after_const

view_in_codegen

viz_cbroadcast

warp_fun

wgpu-f16-packed

working_postopt

x86_moves

x86_numel

#1

#2

#3

#4

#5

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

• 22376e53b7 resnet mlperf logging (#4361) chenyuandGitHub 2024-05-02 00:00:04 -04:00
• f635c4d273 fix define global (#4383) George HotzandGitHub 2024-05-01 19:32:56 -07:00
• ad116dc5c6 fill in mlperf system description (#4381) chenyuandGitHub 2024-05-01 16:47:45 -04:00
• 9358b62073 rename resnet script to dev_beam.sh and dev_run.sh (#4379) chenyuandGitHub 2024-05-01 14:41:35 -04:00
• 6628e13a5f pad resnet eval data in model_train (#4374) chenyuandGitHub 2024-05-01 14:33:42 -04:00
• 105fbd7925 add 3080 support to NV George Hotz 2024-05-01 11:17:01 -07:00
• 826cccd54d fix mean underflow for half tensor (#4377) chenyuandGitHub 2024-05-01 13:38:57 -04:00
• dce7ac0160 NOCLANG=1 for tinybox green ci. (#4378) chenyuandGitHub 2024-05-01 13:31:01 -04:00
• 272bea5100 GraphRunner (#4375) George HotzandGitHub 2024-05-01 10:27:13 -07:00
• 077ea6926c remove downcast_half in sum (#4376) chenyuandGitHub 2024-05-01 11:46:44 -04:00
• bd49d2854a hotfix: skip fetch tests always geohot 2024-05-01 08:43:26 -07:00
• b683d0f496 hotfix: 100% accuracy is wrong geohot 2024-05-01 08:07:18 -07:00
• 8bcf533a84 gitignore open-images-v6TEST geohot 2024-05-01 13:55:38 +00:00
• ea06f657df fusion tests from test_opt (#4357) qazalandGitHub 2024-05-01 16:44:12 +03:00
• 995d264666 hotfix: add CNAME to put docs at docs.tinygrad.org geohot 2024-04-30 23:17:35 -07:00
• 683b7c605a pad first batch of imagenet dataloader and update eval (#4368) chenyuandGitHub 2024-05-01 00:21:52 -04:00
• 4a26718ca9 feat: tinyboxgreen (#4365) wozeparrotandGitHub 2024-04-30 19:05:37 -04:00
• 16838eae08 mlperf/resnet: update tinybox_red parameters to new best values (#4364) Francis LamandGitHub 2024-04-30 15:08:12 -07:00
• 27ee49bf30 tensor variable (#4362) George HotzandGitHub 2024-05-01 06:08:57 +09:00
• d2f89615b2 remove aql remnants in amd (#4346) nimlgenandGitHub 2024-04-30 23:36:02 +03:00
• 0d33c54d99 kernel: change PADTO check to allow up to 4x padding (#4354) Francis LamandGitHub 2024-04-30 12:29:34 -07:00
• babe87a8ae BERT: Checkpoint loading tests (#4359) Elias WahlandGitHub 2024-04-30 20:43:41 +02:00
• c12bcabb07 search: fix actions space checks to ignore TC axis and amt (#4360) Francis LamandGitHub 2024-04-30 11:02:22 -07:00
• fdc8fabae5 disable flaky mac gpt2 beam benchmark and add back cifar mac with JIT=2 (#4358) chenyuandGitHub 2024-04-30 10:41:37 -04:00
• d325be2540 update docs (#4356) George HotzandGitHub 2024-04-30 15:51:42 +08:00
• a2d81514fd just get dtype from kwargs (#4355) SohaibandGitHub 2024-04-30 00:26:14 -07:00
• a9a1fa6bbf wmma: add reduce axis choice to TC action space (#4328) Francis LamandGitHub 2024-04-29 16:15:39 -07:00
• 93abcd3113 fix function.py sum backward without downcast_half (#4353) chenyuandGitHub 2024-04-29 17:53:02 -04:00
• 18c61ce077 test/fuzz_linearizer: add --atol/rtol and change half distribution (#4352) Francis LamandGitHub 2024-04-29 12:53:59 -07:00
• 71ff68b445 dropout after eval step (#4351) Elias WahlandGitHub 2024-04-29 21:47:21 +02:00
• 27613dd881 MLPerf BERT: Main training loop (#4288) 3 people 2024-04-29 20:35:27 +02:00
• 61c97d5305 refactor ops_gpu ctypes (#4331) SohaibandGitHub 2024-04-29 10:33:34 -07:00
• cc1797673e all fusion opportunities (#4348) qazalandGitHub 2024-04-29 19:32:23 +03:00
• f363f39e83 fix dtype of const folded sum (#4349) chenyuandGitHub 2024-04-29 11:40:45 -04:00
• bf412aeb80 use tolist instead of numpy for extracting parameters in onnx (#4333) geohotstanandGitHub 2024-04-29 22:48:20 +08:00
• 774a9b0bca override assign_target in fuzz_schedule (#4342) qazalandGitHub 2024-04-29 11:04:04 +03:00
• bb849a57d1 [MLPerf] UNet3D dataloader (#4343) Francis LataandGitHub 2024-04-28 22:34:18 -04:00
• 82d0ed3cf3 cap default dataset wikipedia max_workers to 32 (#4345) chenyuandGitHub 2024-04-28 21:55:21 -04:00
• c1d8d425eb fix mean of half tensor if sum is greater than hlaf.max (#4327) chenyuandGitHub 2024-04-28 18:04:54 -04:00
• e027879475 hotfix: remove double assignment (#4340) qazalandGitHub 2024-04-28 20:41:31 +03:00
• 23445db2b9 no skipped tests in RHIP (#4337) qazalandGitHub 2024-04-28 19:23:05 +03:00
• e4befa41d7 Fix in _reshape_mask (#4332) Obada KhaliliandGitHub 2024-04-28 18:57:39 +03:00
• 664b563c91 Add insert_before to Linearizer Functions (#4320) TimmyandGitHub 2024-04-28 11:38:36 -04:00
• 3372bea322 reduce children fusion tests (#4321) qazalandGitHub 2024-04-28 18:14:02 +03:00
• f3de17912f added the download if not present missing function (#4318) Arnav MehtaandGitHub 2024-04-28 14:01:08 +05:30
• bc36940c28 fix (#4319) geohotstanandGitHub 2024-04-28 16:29:04 +08:00
• 8d1649d8c2 raise error when too many resources requested in nv (#4324) nimlgenandGitHub 2024-04-27 23:48:51 +03:00
• c6c12ba94a save schedule graph pre validation (#4317) qazalandGitHub 2024-04-27 12:06:15 +03:00
• 40264c7d1e Update index.md (#4315) Victor Ziliang PengandGitHub 2024-04-27 00:12:44 -07:00
• 24a6342950 add mem/s to external_benchmark_resnet (#4309) chenyuandGitHub 2024-04-26 20:07:17 -04:00
• 1f2642c73b kernel: fix calculation of smem size to ignore UNROLL (#4308) Francis LamandGitHub 2024-04-26 11:34:56 -07:00
• de832d26c6 disable bfloat16 from ptx tests (#4305) Szymon OżógandGitHub 2024-04-26 07:20:10 +02:00
• ec65aea32f resnet stop the script once hit target (#4303) chenyuandGitHub 2024-04-25 23:54:56 -04:00
• 1891ebb655 make ring allreduce chunks a multiple of 2^n if possible (#4302) chenyuandGitHub 2024-04-25 23:45:28 -04:00
• 1e37c4a7a1 minor llm.c improvements geohot 2024-04-26 11:15:31 +08:00
• 3ec4b745d6 JIT=2 for mac cifar benchmark (#4300) chenyuandGitHub 2024-04-25 18:33:40 -04:00
• c2dbe2a78b new split reduce heuristic try 2 (#4294) 3 people 2024-04-25 15:14:15 -07:00
• f1ebcffb87 Ptx beam fix (#4296) Szymon OżógandGitHub 2024-04-25 21:39:39 +02:00
• f9a7badace use LR=7 for resnet with BS=1536 (#4299) chenyuandGitHub 2024-04-25 15:23:10 -04:00
• 9a47ed0705 test crossing diamond assigns (#4298) qazalandGitHub 2024-04-25 21:52:05 +03:00
• 5ae252ae83 use at least float32 for optim.lr (#4297) chenyuandGitHub 2024-04-25 14:42:28 -04:00
• 6f792b727b More improvements for resnet layer bench (#4272) David HouandGitHub 2024-04-25 09:40:49 -07:00
• ac9464f47a allow specify number of beam workers (#4292) David HouandGitHub 2024-04-25 07:44:43 -07:00
• 74a1be88f5 test reduce graph permutations (#4291) qazalandGitHub 2024-04-25 11:34:44 +03:00
• 0f0627bc60 add mnist tutorial geohot 2024-04-25 16:08:32 +08:00
• d31e220cbf add mlperf-logging to setup.py mlperf (#4289) chenyuandGitHub 2024-04-24 23:34:34 -04:00
• 6b8a85939d fix lds size for amd (#4287) nimlgenandGitHub 2024-04-24 22:54:42 +03:00
• c11bad766d prepare mlperf submission (#4270) chenyuandGitHub 2024-04-24 13:19:31 -04:00
• c606a0ba6f Docs link fix (#4286) Szymon OżógandGitHub 2024-04-24 18:54:43 +02:00
• c1fbacb182 resnet benchmarks use DEFAULT_FLOAT=HALF (#4285) chenyuandGitHub 2024-04-24 12:10:57 -04:00
• 002a14088e Ptx store gate cast to bool (#4284) Szymon OżógandGitHub 2024-04-24 17:43:44 +02:00
• dbe3e1d548 or true fixes ci (#4283) George HotzandGitHub 2024-04-24 16:48:26 +04:00
• 53853e6d08 save the schedule graph in SAVE_SCHEDULE (#4248) qazalandGitHub 2024-04-24 12:08:51 +03:00
• acb32e1766 hotfix: PM4 supports timing geohot 2024-04-24 08:38:59 +00:00
• ad28fdecb1 si.inputs+outputs -> bufs (#4279) George HotzandGitHub 2024-04-24 11:12:34 +04:00
• 8401de9922 resnet benchmark return early in eval (#4278) chenyuandGitHub 2024-04-24 00:55:01 -04:00
• 38f97aa0fe rename rawbufs to bufs in ExecItem (#4274) George HotzandGitHub 2024-04-24 07:27:27 +04:00
• 60e3aa5cb1 more docs (#4271) George HotzandGitHub 2024-04-24 06:52:42 +04:00
• 6637ecc5fe use IGNORE_JIT_FIRST_BEAM to not BEAM in jit cnt=0 (#4269) chenyuandGitHub 2024-04-23 18:59:43 -04:00
• f3b4dff7c9 KFDProgram -> AMDProgram (#4268) nimlgenandGitHub 2024-04-24 00:29:50 +03:00
• 17328ded7d setitem no return value (#4266) geohotstanandGitHub 2024-04-24 04:28:14 +08:00
• 3a48773f1a BERT dataloader (#4252) Elias WahlandGitHub 2024-04-23 19:44:49 +02:00
• 69341144ba Wikipedia preprocessing script (#4229) Elias WahlandGitHub 2024-04-23 16:28:01 +02:00
• 759b4f41c3 few more KFD -> AMD (#4262) chenyuandGitHub 2024-04-23 10:15:37 -04:00
• 6c25f1abf7 Optimize ptx loops (#4263) Szymon OżógandGitHub 2024-04-23 10:20:14 +02:00
• 967638f0d5 update docs, remove corealize (#4264) George HotzandGitHub 2024-04-23 12:05:29 +04:00
• 9b7efa72ea hotfix: skip 0 line count files in sz.py geohot 2024-04-23 11:56:03 +04:00
• acf4ba5c9f method cache respects beam option (#4261) George HotzandGitHub 2024-04-23 09:00:41 +04:00
• 9a95781d51 renamed (#4260) George HotzandGitHub 2024-04-23 09:00:28 +04:00
• 2ae4f45272 WIP PM4 Support (#4110) 3 people 2024-04-23 08:31:27 +04:00
• 3f6c7ca8bf test: fix test_tensor_core_padded on CUDA and add to benchmarks (#4258) Francis LamandGitHub 2024-04-22 20:22:11 -07:00
• a90de3b574 search: add additional 7 factors to the action space (#4256) Francis LamandGitHub 2024-04-22 16:14:23 -07:00
• de2b1fb468 update adding_new_accelerators doc (#4255) chenyuandGitHub 2024-04-22 18:50:19 -04:00
• bbb0ad4800 wmma: widen TC usage in search by using PADTO on TC axes when possible (#4216) Francis LamandGitHub 2024-04-22 13:50:31 -07:00
• 9e53d6cffa hotfix: 8000 lines geohot 2024-04-22 20:58:16 +04:00
• e6227bdb15 nv driver (#4044) 3 people 2024-04-22 18:50:20 +03:00
• 77a3780005 assert reduce recompute (#4250) qazalandGitHub 2024-04-22 16:12:39 +03:00
• a9bc7c1c49 unify assign tests (#4247) qazalandGitHub 2024-04-22 11:01:15 +03:00
• 37f8be6450 resnet print epoch ops and mem in benchmark (#4244) chenyuandGitHub 2024-04-21 18:32:31 -04:00
• 7bc862767c Improves error message when CUDA module fails to load. (#4243) Micah ZoltuandGitHub 2024-04-21 23:10:14 +08:00