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

• 1900acda09 [READY] ci: setup venv cache (#1475) 3 people 2023-08-21 03:43:16 +02:00
• 3fc7e984f0 __getitem__ refactoring (#1586) Umut ZenginandGitHub 2023-08-21 04:42:30 +03:00
• d627349af0 teeny changes (#1589) George HotzandGitHub 2023-08-20 13:38:38 -07:00
• 012ee7d162 not worth the speed (#1584) George HotzandGitHub 2023-08-20 10:24:58 -07:00
• 739f327d2d Shorter (#1582) George HotzandGitHub 2023-08-20 08:12:16 -07:00
• 4fbce972d7 CSE at uop level (#1483) David HouandGitHub 2023-08-19 20:40:40 -10:00
• b9feb1b743 fp16 support in stable diffusion geohot 2023-08-20 05:36:25 +00:00
• ad7d26c393 fix __launch_bounds__ and benchmark TC MATMUL (#1575) George HotzandGitHub 2023-08-19 10:54:39 -07:00
• 92754e177c cache buffer loads across multiple bufs (#1482) David HouandGitHub 2023-08-19 06:09:58 -10:00
• e464442adf WMMA for 7900XTX (#1563) George HotzandGitHub 2023-08-19 09:07:23 -07:00
• faa521bcab fix usage of arm64 regs according to CC (#1570) nimlgenandGitHub 2023-08-19 07:40:32 +03:00
• 68ebbd2954 for issue #1555, int64 and int8 in CI=1 ARM64=1 CLANG=1 (#1572) corranrandGitHub 2023-08-19 05:40:13 +01:00
• ae39cf84ab Symbolic Shape JIT main PR (#1353) chenyuandGitHub 2023-08-18 14:39:55 -07:00
• 84e6693915 fix: apt-get to apt, no recommends, clean up (#1571) 3 people 2023-08-18 22:48:59 +02:00
• 50decf0d45 train cifar using multigpu (#1529) wozeparrotandGitHub 2023-08-18 12:35:44 -04:00
• be50b2fe8f more symbolic symbolic ops (#1564) chenyuandGitHub 2023-08-18 09:21:41 -07:00
• dfec16cc83 Support arg int for CUDA kernel (#1565) chenyuandGitHub 2023-08-18 09:19:40 -07:00
• 15150d60c4 fix: small fix for lru on hip (#1567) wozeparrotandGitHub 2023-08-18 12:18:38 -04:00
• c65ad43a93 cleanup ops_gpu (#1566) wozeparrotandGitHub 2023-08-17 23:43:08 -04:00
• bd111411bf init allocator for compiled backends (#1467) 3 people 2023-08-17 20:33:32 +03:00
• a293c18d34 Gather bugfix (#1561) geohotstanandGitHub 2023-08-16 16:53:14 -07:00
• cb62911f6b PTX Reintegration and Passing Tests (#1512) Ethan SorrellandGitHub 2023-08-16 18:20:20 -05:00
• 8763037f0e Fancy indexing is fancy wow and gather thing (#1399) geohotstanandGitHub 2023-08-16 15:35:49 -07:00
• 11dd9b1741 symbolic codegen and exec (#1552) chenyuandGitHub 2023-08-16 14:43:41 -07:00
• 1e1d48b4e6 single model (#1560) George HotzandGitHub 2023-08-16 13:22:19 -07:00
• 491e85597a Run onnx commavq model (#1537) 3 people 2023-08-16 21:24:40 +02:00
• 55d95d1658 llama 70b (#1558) wozeparrotandGitHub 2023-08-16 14:36:12 -04:00
• c93e63b8b5 make TestNonFloatUOps.test_mul_bool pass on all platforms (#1557) nimlgenandGitHub 2023-08-16 21:34:09 +03:00
• 074c467020 hotfix for broken ci (#1559) wozeparrotandGitHub 2023-08-16 13:52:03 -04:00
• 962972ee68 Fix uops int32 for llvm (#1554) madt2709andGitHub 2023-08-15 23:22:32 -07:00
• 2cde667d40 Change Any to List[Optional[RawBuffer]] in JIT (#1553) Sam BaraniandGitHub 2023-08-16 02:21:33 -04:00
• fa81e282c2 fix missing dtypes in is_int,is_float,is_unsigned (#1550) nimlgenandGitHub 2023-08-16 04:22:29 +03:00
• d17ecccd78 Torch/LLVM/arm F64 support (#1551) DiogoandGitHub 2023-08-15 21:21:08 -04:00
• 913263c155 add double: c_type.double for CLANG (#1549) YiMing HanandGitHub 2023-08-15 16:19:33 -04:00
• 0b5930d406 more uops testing, who isn't passing right now... (#1522) George HotzandGitHub 2023-08-15 09:07:26 -07:00
• f8109b830c promote assembly to the main codebase (#1544) George HotzandGitHub 2023-08-14 22:47:45 -07:00
• 666ac61070 support for p2p buffer transfers (#1523) wozeparrotandGitHub 2023-08-15 01:39:57 -04:00
• 93a36c3659 Arm (#1421) 3 people 2023-08-14 22:29:30 -04:00
• a89142e46f ShapeTracker.var_vals (#1540) chenyuandGitHub 2023-08-14 18:53:37 -07:00
• a453d718a1 fix file race condition in ops_clang via pid in the filename (#1541) Pavol RusnakandGitHub 2023-08-15 03:50:10 +02:00
• 9cb2bda34f Revert "Better reshape (#1423)" (#1538) wozeparrotandGitHub 2023-08-14 13:04:54 -04:00
• cf2bf1518d Better reshape (#1423) Sieds LyklesandGitHub 2023-08-14 18:09:04 +02:00
• e00acb1eaf fix deepwalk ctx check (#1536) YiMing HanandGitHub 2023-08-14 02:03:17 -04:00
• 2fd7004980 Implementation of SoftVC VITS SVC model (#1371) JaSpa99andGitHub 2023-08-14 04:43:23 +02:00
• b6937acb7e fix casting behavior for interpreted buffers (#1525) nimlgenandGitHub 2023-08-14 05:21:37 +03:00
• 13659ac6fa examples: numpy() array returns only one value, not an array (#1534) David HeidelbergandGitHub 2023-08-13 23:33:05 +02:00
• 3e0c2d256f symbolic shapetracker (#1506) chenyuandGitHub 2023-08-12 12:22:58 -07:00
• 875da762a8 fix file race condition in ops_clang (#1458) Pavol RusnakandGitHub 2023-08-12 18:31:46 +02:00
• d3d58a37e5 Bert: use Tensor.scaled_dot_product_attention (#1528) JaSpa99andGitHub 2023-08-12 17:46:04 +02:00
• 330fb7b1a3 Print more meaningfull hip error messages (#1530) Szymon OżógandGitHub 2023-08-12 16:16:20 +02:00
• 29d5801387 distributed collectives (#1519) wozeparrotandGitHub 2023-08-11 13:22:07 -04:00
• 2e85fce068 Transformer: use Tensor.scaled_dot_product_attention (#1520) Jacky LeeandGitHub 2023-08-11 09:00:37 -07:00
• 38fe84d92b cleanup mlops (#1521) George HotzandGitHub 2023-08-10 19:53:28 -07:00
• 47f18f4d60 [New] SD: Refactor AttnBlock, CrossAttention, CLIPAttention to share code (#1516) (#1518) 3 people 2023-08-10 15:04:18 -07:00
• 7e7c9001e9 distributed world (#1481) wozeparrotandGitHub 2023-08-10 13:00:51 -04:00
• e3c6c0c6db add GPT2 example (#1511) (#1514) 3 people 2023-08-10 09:09:47 -07:00
• c82bd59b85 Revert "SD: Refactor AttnBlock, CrossAttention, CLIPAttention to share code (#1513)" (#1515) George HotzandGitHub 2023-08-10 09:08:51 -07:00
• 85e02311a2 SD: Refactor AttnBlock, CrossAttention, CLIPAttention to share code (#1513) Jacky LeeandGitHub 2023-08-10 08:52:33 -07:00
• 07b79f210f llvmir support for bool <-> float casting (#1492) geohotstanandGitHub 2023-08-09 10:12:52 -07:00
• 351684395c dont run on fork (#1510) wozeparrotandGitHub 2023-08-09 13:06:45 -04:00
• 88e2e0c8a3 Revert "don't try to run benchmark on forks" (#1508) wozeparrotandGitHub 2023-08-09 12:59:49 -04:00
• 65b65b760b don't try to run benchmark on forks (#1507) wozeparrotandGitHub 2023-08-09 12:59:19 -04:00
• c417cd3c97 fast HIP gemm -> 100 TFLOPS (#1476) George HotzandGitHub 2023-08-09 06:54:15 -07:00
• 1766f0c0cf use ConstOp for valid.max == 0 (#1501) 3 people 2023-08-09 00:01:59 -07:00
• ef5f648e2f Tensor.scaled_dot_product_attention to match torch, used in LLaMA, and tested (#1502) Jacky LeeandGitHub 2023-08-08 23:27:13 -07:00
• dabfd7569a use allclose instead of equals in test_jit (#1504) nimlgenandGitHub 2023-08-09 08:22:17 +03:00
• 827d13e64e correct patch JIT llama chat (#1500) chenyuandGitHub 2023-08-08 16:52:09 -07:00
• 7c2ea85bb0 Raise memory limit for CIFAR test (#1499) Yixiang GaoandGitHub 2023-08-08 18:40:56 -05:00
• 293a10204b Add tinygrad.renderer to packages in setup.py (#1497) Thiago Franco de MoraesandGitHub 2023-08-08 19:51:49 -03:00
• 0415a48cfc patch JIT llama chat mode (#1496) chenyuandGitHub 2023-08-08 15:15:56 -07:00
• 6480a1a180 CIFAR 94.03% (#1340) Yixiang GaoandGitHub 2023-08-08 17:13:24 -05:00
• aa83a9e910 ci: fix gpuocelot build cache (#1474) 3 people 2023-08-08 23:00:04 +02:00
• d24f936501 just cmplt (#1493) George HotzandGitHub 2023-08-08 13:58:10 -07:00
• e2cf0f322e [READY] ci: missing n=auto (#1486) 3 people 2023-08-08 16:37:24 +02:00
• 0ce7511110 fix: is not use with a literal (#1487) 3 people 2023-08-08 16:35:30 +02:00
• 932dad1a2b fix cast bool->float in llvmir (#1480) nimlgenandGitHub 2023-08-08 07:30:51 +03:00
• 046fd7437a use fake buffer for external_test_speed_llama.py (#1478) nimlgenandGitHub 2023-08-08 05:05:44 +03:00
• 5fdd248617 don't download cifar (#1472) George HotzandGitHub 2023-08-06 21:38:59 -07:00
• d78fb8f4ed add stable diffusion and llama (#1471) George HotzandGitHub 2023-08-06 21:31:51 -07:00
• 24933ab551 Actually flip local_max in CUDA (#1462) 3 people 2023-08-06 20:35:25 +03:00
• d7d1011f1e Add WEBGPU tests to CI (#1463) DiogoandGitHub 2023-08-06 13:32:01 -04:00
• 486a9dbfd9 speed v torch (#1464) George HotzandGitHub 2023-08-06 09:32:33 -07:00
• 2ab282bfec run on update_benchmark too (#1460) George HotzandGitHub 2023-08-06 08:58:37 -07:00
• 3d41674b42 Fixed regression (#1447) 3 people 2023-08-06 17:55:58 +03:00
• d67e248d9b simple bitcast 2 (#1445) George HotzandGitHub 2023-08-06 00:30:50 -07:00
• 943b227cb1 only on push to master geohot 2023-08-06 00:10:07 -07:00
• 2274e3e757 Fix benchmark (#1454) George HotzandGitHub 2023-08-05 23:44:36 -07:00
• bf21aec81f do benchmarking (#1451) George HotzandGitHub 2023-08-05 23:35:01 -07:00
• 1ba8ae62a1 Match Torch speed for sum reduction (#1387) 3 people 2023-08-06 08:27:33 +03:00
• 09ede08b23 simplify Node.sum aggregating (#1449) chenyuandGitHub 2023-08-05 22:19:36 -07:00
• 7fa730b506 external model benchmark test geohot 2023-08-05 22:10:48 -07:00
• cb5dcc7b57 remove view_from_shape (#1448) chenyuandGitHub 2023-08-05 20:39:13 -07:00
• e2af95c2f8 moved global_max and local_max to LinearizerOptions also added assert for max bufs (#1446) DiogoandGitHub 2023-08-05 21:23:18 -04:00
• 7b8d06c9f1 test uops (#1444) George HotzandGitHub 2023-08-05 12:35:56 -07:00
• 84c430355e fix backends for new style (#1443) George HotzandGitHub 2023-08-05 11:07:04 -07:00
• 67781fcf5d fix fail fast in CI geohot 2023-08-05 10:24:24 -07:00
• bd7f4b1249 move renamer to linearizer (#1442) George HotzandGitHub 2023-08-05 08:53:25 -07:00
• 669b406ec6 correct children count with lazycache (#1429) nimlgenandGitHub 2023-08-05 10:30:16 +03:00
• 97a6029cf7 Corrected a few misspelled words (#1435) FelixandGitHub 2023-08-05 01:51:08 +02:00
• 043d5f2cb5 Fix NOUNROLL (#1439) Adrian KretzandGitHub 2023-08-05 01:50:19 +02:00