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

• 957e9800f1 llama + beam to mac benchmark, full cifar to nvidia benchmark (#3612) chenyuandGitHub 2024-03-04 21:35:57 -05:00
• 282bbd5acb check the input length into argfix (#3610) chenyuandGitHub 2024-03-04 19:50:17 -05:00
• 7db6dd725d multilazybuffer fix (#3609) Elias WahlandGitHub 2024-03-04 23:36:23 +01:00
• c3b8d285aa cleanup uops (#3605) chenyuandGitHub 2024-03-04 11:03:14 -05:00
• 94679322a3 simpler float4 direct store and locals support (#3592) qazalandGitHub 2024-03-04 16:28:28 +02:00
• 3db826e195 hsa in lin opts (#3602) nimlgenandGitHub 2024-03-04 17:17:32 +03:00
• 7c90005c65 search: hotfix to make sure TC behavior is all in applied_opts (#3598) Francis LamandGitHub 2024-03-03 18:44:38 -08:00
• 8e5d60a322 add more gpt2 variant in mac/nvidia benchmark (#3599) chenyuandGitHub 2024-03-03 17:55:30 -05:00
• 968d109453 apply more create_lt_node (#3597) chenyuandGitHub 2024-03-03 16:12:39 -05:00
• bc562c4747 Python div alu behavior differs slightly from others (#3596) 3 people 2024-03-03 13:48:25 -05:00
• 56d21d77b3 Fix two bugs concerning Tensor.to. (#3593) Marcin SłowikandGitHub 2024-03-03 17:48:56 +01:00
• 0082300a59 Fix symbolic negative floordiv (#3594) 3 people 2024-03-03 11:40:52 -05:00
• e09619ab6c explicitly create_lt_node when used in shapetracker _expr_view (#3561) chenyuandGitHub 2024-03-03 10:08:21 -05:00
• 640dc0fc51 hsa flush hdp (#3591) nimlgenandGitHub 2024-03-03 15:55:07 +03:00
• 660df3cff1 Add test for .softmax.argmax (#3559) 4 people 2024-03-02 23:51:52 -05:00
• ee41fafdab use operator instead of lambda in python_alu (#3590) chenyuandGitHub 2024-03-02 19:33:21 -05:00
• a89afd4ffa Directly store float4 nodes (#3564) 3 people 2024-03-03 01:58:20 +02:00
• 770707b376 hotfix: gpuocelot no rebuild geohot 2024-03-02 15:57:38 -08:00
• 74c9acddb0 simple python ALU (#3589) George HotzandGitHub 2024-03-02 15:50:58 -08:00
• 162dfb07d9 fuzz_linearizer: fix uops and add to test.yml (#3588) Francis LamandGitHub 2024-03-02 15:03:42 -08:00
• 8978488565 add sanity tests for bufs_from_lin (#3586) Jovan SardinhaandGitHub 2024-03-02 14:17:43 -08:00
• aa9b013d79 add constant folding for WHERE in uops (#3584) George HotzandGitHub 2024-03-02 10:37:14 -08:00
• 3b7e3fa2e4 fix sync in hsa graph (#3582) nimlgenandGitHub 2024-03-02 18:37:51 +03:00
• 6c36264790 Improve type hints for optimizer (#3583) Szymon OżógandGitHub 2024-03-02 16:35:44 +01:00
• 83530a585f add quick external data select test geohot 2024-03-02 05:38:32 -08:00
• 9a37273d36 consts don't have nodes in the graph (#3579) George HotzandGitHub 2024-03-02 04:19:11 -08:00
• 41f0a25b53 lazy.py: cache consts (#3577) George HotzandGitHub 2024-03-02 03:50:05 -08:00
• fb8acd1851 Don't touch UOps.DEFINE_GLOBAL (#3575) uuuvnandGitHub 2024-03-02 13:30:05 +02:00
• 50e1445e60 Revert "allow overriding weight init for Linear (#3569)" (#3576) George HotzandGitHub 2024-03-02 03:17:13 -08:00
• 2d0973a852 allow overriding weight init for Linear (#3569) David HouandGitHub 2024-03-02 03:16:04 -08:00
• 9642a8f547 search: add BEAM UPCAST/LOCAL params and loosen TC criteria during BEAM (#3563) Francis LamandGitHub 2024-03-02 03:11:25 -08:00
• ba6c041eab fix SCE ignore_index with label_smoothing (#3574) David HouandGitHub 2024-03-01 19:19:45 -08:00
• e17f1821a7 wmma: add CUDA tensor core and fix test_speed_v_torch failure (#3544) Francis LamandGitHub 2024-03-01 17:51:02 -08:00
• b3cdc11a58 label_smoothing in sparse_cat_crossentropy (#3568) David HouandGitHub 2024-03-01 17:02:46 -08:00
• 6b29c70b3d Refactor to UOpGraph class (#3566) George HotzandGitHub 2024-03-01 15:14:48 -08:00
• b7e555f6c0 run test_linearizer_failures on PYTHON backend (#3565) chenyuandGitHub 2024-03-01 17:00:18 -05:00
• 48d22067ca clean up test_linearizer_failures (#3562) chenyuandGitHub 2024-03-01 15:57:17 -05:00
• d16aa89561 don't allow MLB assigns with different axes (#3557) David HouandGitHub 2024-03-01 04:59:06 -08:00
• cfd23f398d Revert "don't allow MLB assigns with different axes (#3483)" (#3554) chenyuandGitHub 2024-02-29 23:13:07 -05:00
• f19d8bb7b4 don't allow MLB assigns with different axes (#3483) David HouandGitHub 2024-02-29 20:04:12 -08:00
• 35d998efa8 disable flaky test_conv_beam in CI (#3553) chenyuandGitHub 2024-02-29 22:59:41 -05:00
• e5385eecfc UnsyncedBatchNorm with synced trainable weights for hlb cifar (#3472) 3 people 2024-02-29 19:52:07 -08:00
• 5a6e151844 no barrier side effect (#3550) George HotzandGitHub 2024-02-29 18:10:04 -08:00
• bd9c2ced07 define var can be removed from vars to keep (#3549) George HotzandGitHub 2024-02-29 17:44:19 -08:00
• 2c19ab6561 define var (#3548) George HotzandGitHub 2024-02-29 16:43:27 -08:00
• 83cdc85790 add index to DEFINE_GLOBAL (#3542) George HotzandGitHub 2024-02-29 15:22:26 -08:00
• 978a997d1f print nvidia-smi in CI benchmark (#3546) chenyuandGitHub 2024-02-29 17:31:37 -05:00
• 5d434801fa search: add tensor core to beam search space (#3275) Francis LamandGitHub 2024-02-29 13:05:10 -08:00
• f90caa4b92 Escape table name in diskcache queries. (#3543) Marcin SłowikandGitHub 2024-02-29 22:04:21 +01:00
• 0afde98ba5 scan all gpu agents at launch (#3535) nimlgenandGitHub 2024-02-29 20:37:37 +03:00
• 2e82c5b7a4 README: ops_cpu and ops_torch have been removed (#3539) Mark McLoughlinandGitHub 2024-02-29 15:22:11 +00:00
• b05776ef3e fix addresses of dispatch packets (#3534) nimlgenandGitHub 2024-02-29 16:43:55 +03:00
• 9268a8b154 remove MULACC (#3459) geohotstanandGitHub 2024-02-29 20:40:40 +08:00
• 94fc0fd546 uop the float4 acc upcast in group_for_reduce kernels (#3466) qazalandGitHub 2024-02-29 03:33:47 +02:00
• 48918fa75a fix disktensor offset issue (#3532) George HotzandGitHub 2024-02-28 17:22:17 -08:00
• 0b1fc5888a fix 'Import Error: cannot import name compile_cuda from tinygrad.runtime.ops_cuda' error in extra/gemm/cuda_matmul.py (#3531) Caleb BunchandGitHub 2024-02-28 20:15:32 -05:00
• 275971e616 fix: align .split, .chunk and .unsqueeze with torch, add fuzz tests (#3505) David FriehsandGitHub 2024-02-29 02:06:39 +01:00
• e7cda40d52 Revert "hotfix: disable metal graph" geohot 2024-02-28 16:25:12 -08:00
• 42eb8de0d4 Revert "move all reduces to the end in lazy (#3475)" (#3529) 3 people 2024-02-28 16:24:10 -08:00
• 0c6846f9fc failed test case for disk tensor assign into dtype int64 (#3527) chenyuandGitHub 2024-02-28 17:52:21 -05:00
• d89e3c4e08 enable METAL tests now runner is M1 and no fast-math (#3523) chenyuandGitHub 2024-02-28 14:14:23 -05:00
• 1136e2a82a skipIf(not( -> skipUnless( in test_linearizer_failures (#3519) chenyuandGitHub 2024-02-28 13:48:47 -05:00
• 3541602877 hotfix: disable metal graph geohot 2024-02-28 10:33:34 -08:00
• c34d382a1e bump to macos-14 M1 (#3520) George HotzandGitHub 2024-02-28 10:28:25 -08:00
• 505ac6ac96 Revert "check buffers are seeable by other gpu before transfer (#3504)" (#3522) George HotzandGitHub 2024-02-28 10:26:27 -08:00
• db2cf48828 check buffers are seeable by other gpu before transfer (#3504) nimlgenandGitHub 2024-02-28 21:24:50 +03:00
• da32c37346 use hash as key for beam (#3516) wozeparrotandGitHub 2024-02-28 13:19:01 -05:00
• 1f5c24798b Raise exception if MTLCommandBuffer fails (#3465) uuuvnandGitHub 2024-02-28 20:14:08 +02:00
• 08ef77c721 hsa multigpu graph (#3403) nimlgenandGitHub 2024-02-28 20:40:53 +03:00
• fa88e1d0d0 cleanup lazy reduce (#3517) chenyuandGitHub 2024-02-28 08:15:01 -05:00
• 2127c1c6c2 test for the split reduce kernel (#3515) chenyuandGitHub 2024-02-27 21:29:25 -05:00
• 94b7ac7a29 no cuda compile helper (#3512) nimlgenandGitHub 2024-02-28 03:50:10 +03:00
• 88939c3347 fix Node.max can be symbolic (#3514) chenyuandGitHub 2024-02-27 17:21:31 -05:00
• 969b57f0fe enable symbolic_ops and jits test of two vars (#3513) chenyuandGitHub 2024-02-27 11:17:46 -05:00
• ea4b8e5b1f feat: don't hardcode the arch (#3511) wozeparrotandGitHub 2024-02-27 07:58:03 -05:00
• 11da65bccd test/external/fuzz_linearizer: add a FUZZ_MAX_SIZE option (#3455) Francis LamandGitHub 2024-02-27 04:34:59 -08:00
• a29cd6d464 run f64 increased precision tests on remu (#3509) qazalandGitHub 2024-02-27 01:01:07 +02:00
• b1426f3a4c cleanup SumNode mod (#3503) chenyuandGitHub 2024-02-26 11:10:55 -05:00
• 61605ccc69 Remove special case of SumNode div SumNode (#3502) chenyuandGitHub 2024-02-26 09:42:06 -05:00
• 39d75f0d58 test_linearizer_failures: add more METAL examples (#3495) Francis LamandGitHub 2024-02-26 01:19:05 -08:00
• b154089884 float64 function support for HIP (#3492) chenyuandGitHub 2024-02-24 09:46:20 -05:00
• 35aff8b0c2 properly exclude PYTHON backend and support of half (#3491) chenyuandGitHub 2024-02-24 09:22:06 -05:00
• 2fe98b64bb fix Tensor.split not passing dim to Tensor.chunk (#3490) David FriehsandGitHub 2024-02-24 13:53:11 +01:00
• b41761488d change specific string 'CLANG' to DEVICE variable in abstractions2.py (#3488) Caleb BunchandGitHub 2024-02-24 07:51:39 -05:00
• c032df520b minor symbolic type related cleaups (#3489) chenyuandGitHub 2024-02-23 16:44:43 -05:00
• 15df9406d6 fix exec_alu(UnaryOps.SQRT, <...>, (0,)) + add test (#3487) Carson RadtkeandGitHub 2024-02-23 11:28:00 -06:00
• 52567da07f jit grapher simplified (#3478) nimlgenandGitHub 2024-02-23 18:20:16 +03:00
• 2113e1eb63 move all reduces to the end in lazy (#3475) George HotzandGitHub 2024-02-23 15:49:11 +01:00
• 5cfcc2a8d7 support MLB reshaping on-axis for evenly sharded (#3484) David HouandGitHub 2024-02-23 04:51:36 -08:00
• 358a24eae6 symbolic use mod for rmod and use floordiv for rfloordiv (#3485) chenyuandGitHub 2024-02-23 01:05:13 -05:00
• 6d048a0c0b cache collector optimizations are allowed only for kernel operations (#3476) nimlgenandGitHub 2024-02-22 14:26:57 +03:00
• 7698781389 Revert "wmma: add CUDA tensor core (#3464)" (#3474) George HotzandGitHub 2024-02-22 11:58:16 +01:00
• e9cef13f0b wmma: add CUDA tensor core (#3464) Francis LamandGitHub 2024-02-22 02:57:08 -08:00
• 57678012e1 Upload correct benchmark artifact (#3471) wozeparrotandGitHub 2024-02-22 01:14:16 -05:00
• ab40c0cf93 clean up long lines in symbolic (#3469) chenyuandGitHub 2024-02-21 21:57:44 -05:00
• 7c0fc40123 enable test IMAGE=2 PYTHON=1 python3 test/test_ops.py TestOps.test_simple_conv2d (#3468) chenyuandGitHub 2024-02-21 18:30:12 -05:00
• 77d2a4c12a regenerate kernel dataset after reduce arg to axis change (#3467) chenyuandGitHub 2024-02-21 18:16:13 -05:00
• f513c37e64 support same uidx in multiple shape positions (#3205) 3 people 2024-02-21 10:37:03 -08:00
• 1eb24af63b fix softmax and log_softmax for 0d tensor (#3463) chenyuandGitHub 2024-02-21 11:30:30 -05:00
• 871ba73e65 _reduce_op is axis based now (#3462) George HotzandGitHub 2024-02-21 16:36:31 +01:00