Skip to content

Sync: tinygrad/tinygrad:master into sunnypilot/tinygrad:master#3

Merged
Discountchubbs merged 956 commits into
masterfrom
tinygrad-sync
Apr 26, 2026
Merged

Sync: tinygrad/tinygrad:master into sunnypilot/tinygrad:master#3
Discountchubbs merged 956 commits into
masterfrom
tinygrad-sync

Conversation

@Discountchubbs
Copy link
Copy Markdown

@Discountchubbs Discountchubbs commented Apr 26, 2026

No description provided.

nimlgen and others added 30 commits April 5, 2026 18:01
@nimlgen
am: large allocs aligned to 2mb to use 2mb pages (tinygrad#15609)
604cdbf
@nimlgen
hcq: add write/poll_bit commands (tinygrad#15610)
5e134aa 
* hcq: add write/poll_bit commands

* x
@nimlgen
hcq: support non for signal_t and compute_t (tinygrad#15611)
e0988db 
* hcq: support non for signal_t and compute_t

* revert

* x
@nimlgen
mlx: init runtime (tinygrad#15612)
e3986a6 
* mlx: init

* x

* swap
@nimlgen
hotfix: fix bert (tinygrad#15613)
6a334ce
@Andrew-most-likely
validate lr, momentum, weight_decay in optimizers (tinygrad#15576)
e39cfe6
@13Perrius
remove redundant iteration and toposort in _deepwalk (tinygrad#15532)
ff0c941
@valtterivalo
add gpu_family detection to Metal, target MSL 4.0 on macOS 26+ (tinyg…
86c4431 
…rad#15079)

use supportsFamily API to detect GPU generation instead of parsing
ICB debug description strings. also adds metal4.0 compiler target.
@nimlgen
mlx: cleaner (tinygrad#15617)
e270047 
* mlx: cleaner

* x
@chenyuxyz
support int operand for shifts (tinygrad#15618)
01b49c8 
matches torch/jax, also symbolic rule to remove mask
@chenyuxyz
lower fuzz_symbolic_symbolic_div timeout (tinygrad#15619)
a444be1 
mitigate timeout crash due to high total time
@chenyuxyz
update shifts in torch backend (tinygrad#15622)
6e30a5f
@chenyuxyz
support shift by Tensor (tinygrad#15623)
1483f7e 
* support shift by Tensor

* use mixin
@chenyuxyz
more activations to mixin (tinygrad#15624)
66ec188
@chenyuxyz
shared _normalize_indices for getitem (tinygrad#15625)
2f7d085 
* shared _normalize_indices for getitem

* list
@Qazalin
viz: reenable tests (tinygrad#15626)
8ba5830
@sirhcm
interface in DEV (tinygrad#15620)
19e9649
@wozeparrot
llama: unify scripts (tinygrad#15628)
810d7c0
@geohot
USB driver for custom ASM firmware (tinygrad#15597)
2b01ca5 
* USB driver for custom ASM firmware

* timeout

* fix mypy

* pcie mem read

* flip in f/w

* one tx

* litle endian

* autodetect custom

* mock bypass

* lint

* clean
@geohot
improvements to apps.llm (tinygrad#15631)
d3de63d
@Qazalin
viz: speed up profiler first render (tinygrad#15632)
d29f0ef 
* viz: speed up profiler first render

* better comment
@nimlgen
mlx: graph (tinygrad#15621)
b78b384 
* Dx

* Dx

* simpler

* mypy

* x

* f

* Dx

* x

* c

* x
@Qazalin
update llama profile.sh (tinygrad#15633)
890286e 
* update llama profile.sh

* BENCHMARK 5
@chenyuxyz
move lerp to mixin (tinygrad#15634)
9c6e925 
last function of math function section
@Qazalin
viz: delete redundant things (tinygrad#15637)
a508b8f 
* delete that

* remove

* delete graph config
@b1tg
llm: buffer SSE chunks to fix parse errors from split reads (tinygrad…
bf37635 
…#15641)
@geohot
llm: change the default port to 8000 so you can remember it (match vLLM)
f930579
@sirhcm
fix: comma 4 does not have pcie (tinygrad#15642)
bcf6931
@wozeparrot
llama: move custom_kernel into flat_llama (tinygrad#15643)
70dbd35
@Qazalin
viz: add # of bytes to sdma (tinygrad#15639)
dc6a51e 
* viz: add # of bytes to sdma

* update test_viz
eitanturok and others added 24 commits April 24, 2026 20:07
@eitanturok
no uv.lock (tinygrad#15893)
48d7ab2
@chenyuxyz
update Tensor.allclose to return Tensor (tinygrad#15904)
7a1adfd 
matches jax
@nimlgen
jit: no jit_cache in graphrunner (tinygrad#15907)
4010aa4 
* jit: no jit_cache in graphrunner

* m
@chenyuxyz
sort argsort topk allclose to mixin (tinygrad#15910)
03a7604
@nimlgen
remove last jit_cahce (tinygrad#15911)
56a9f1e 
* remove last jit_cahce

* linter
@nimlgen
remove linear_to_schedule from tests (tinygrad#15912)
f275195 
* remove linear_to_schedule from tests

* x
@chenyuxyz
scatter to mixin (tinygrad#15917)
2f9fdb4
@chenyuxyz
nll_loss to mixin (tinygrad#15918)
b501ba3
@nimlgen
schedule() -> schedule_linear() in tests (batch 1) (tinygrad#15915)
d337801 
* schedule_with_vars -> linear_with_vars in tests

* tests batch 1

* batch 2

* estimate_uop

* simpler

* rm
@wozeparrot
llama: fused ce loss (tinygrad#15920)
4b908b6
@sirhcm
amd: fallback to llvm when comgr is not available (tinygrad#15914)
57fbaa3
@sirhcm
nv: fix shader local memory for NAK (tinygrad#15921)
8b2826e
@germiBest
webgpu: fix weight lookup in export_model after compile_net key change (
1fdcb13 
tinygrad#15919)

* fix lookup site in export_model_webgpu after refactoring

webgpu (sd): fix export_model weight lookup after compile_net changes

fix lookup site in export_model_webgpu after refactoring

* add regression test
@nimlgen
remove schedule() from tests batch 2 (tinygrad#15923)
3c8a2db 
* remove schedule() from tests batch 2

* batch 4
@nimlgen
remove schedule batch 3 (tinygrad#15924)
d2ab6ea 
* remove shcedule batch 3

* batch 6

* batch 7
@nimlgen
remove schedule batch 4 (tinygrad#15927)
a5e9ea7 
* remove schedule batch 4

* fini
@nimlgen
remove schedule from extra/docs/examples (tinygrad#15929)
768106a 
* remove schedule from extra/docs/examples

* f
@Qazalin
viz/cli: unify profile and rewrites, -s ALL default (tinygrad#15931)
9a23de7 
* work

* workg

* better

* cleanup

* better defaults

* --ls

* better

* work

* update llama

* update
@nimlgen
remove old schedule (tinygrad#15930)
e0ff6cc 
* remove old schedule

* tests

* r

* x
@chenyuxyz
remove unused UOp.shard_size [pr] (tinygrad#15933)
e27444a
@nimlgen
remove execitem (tinygrad#15932)
bb65235 
* remove execitem

* f

* x
@nimlgen
remove some runners (tinygrad#15934)
ac3494a 
* remove runners

* mypy
@chenyuxyz
remove unused QCOMTextureInfo, QueueType [pr] (tinygrad#15935)
e9983e3
@Discountchubbs
Merge remote-tracking branch 'tinygrad/master' into HEAD
62ff96b
@Discountchubbs Discountchubbs self-assigned this Apr 26, 2026
@github-actions
Copy link
Copy Markdown

github-actions Bot commented Apr 26, 2026

Changes

Name                                         Lines    Diff    Tokens/Line    Diff
-----------------------------------------  -------  ------  -------------  ------
tinygrad/uop/ops.py                           1349    +172           18.3    -0.4
tinygrad/viz/js/index.js                      1048    +136            5.3    +0.2
tinygrad/nn/onnx.py                           1046      +0           18.2    -0.0
tinygrad/tensor.py                             893    -598           20.1    -0.9
tinygrad/runtime/ops_amd.py                    841     +26           18.1    +0.0
tinygrad/viz/serve.py                          656    +118           17.2    +0.3
tinygrad/runtime/ops_nv.py                     634      +9           18.5    -0.2
tinygrad/renderer/amd/sqtt.py                  567    +567            9.0    +9.0
tinygrad/runtime/support/am/ip.py              518     +29           19.9    -0.5
tinygrad/runtime/support/nv/ip.py              483      +4           18.1    +0.1
tinygrad/helpers.py                            481     +51           16.9    +0.3
tinygrad/renderer/cstyle.py                    463     +10           20.0    -0.1
tinygrad/renderer/amd/generate.py              421    +421           14.8   +14.8
tinygrad/mixin/__init__.py                     419    +415           22.3   +16.0
tinygrad/schedule/rangeify.py                  398     -14           20.3    +0.7
tinygrad/uop/decompositions.py                 388     +22           25.2    -0.2
tinygrad/runtime/support/hcq.py                380     +38           17.4    -0.1
tinygrad/renderer/amd/dsl.py                   351    +351           14.0   +14.0
tinygrad/runtime/support/usb.py                340    +108           18.1    +0.3
tinygrad/runtime/support/system.py             338     +69           20.3    -0.7
tinygrad/llm/model.py                          333    +333           17.3   +17.3
tinygrad/uop/symbolic.py                       324     +36           29.4    -0.2
tinygrad/device.py                             323      -8           13.7    +0.1
tinygrad/dtype.py                              317      -7           15.1    +1.3
tinygrad/runtime/ops_qcom.py                   311     -12           19.1    -0.1
tinygrad/codegen/opt/postrange.py              290     -12           17.4    -0.5
tinygrad/codegen/late/devectorizer.py          276     +25           21.0    +0.7
tinygrad/mixin/elementwise.py                  274    +274           14.2   +14.2
tinygrad/renderer/llvmir.py                    258      -1           21.3    -0.0
tinygrad/runtime/ops_dsp.py                    255      -1           17.3    +0.1
tinygrad/engine/jit.py                         253     -33           16.6    +0.8
tinygrad/runtime/support/autogen.py            247      +6           18.3    +0.1
tinygrad/mixin/movement.py                     246    +100           20.6    +1.2
tinygrad/renderer/nir.py                       241      -3           25.1    +0.5
tinygrad/runtime/support/am/amdev.py           240      +6           17.4    +0.0
tinygrad/runtime/graph/hcq.py                  218     +63           20.0    -1.0
tinygrad/renderer/ptx.py                       211      -3           25.8    -0.1
tinygrad/runtime/ops_python.py                 203     -12           18.3    +0.2
tinygrad/viz/cli.py                            199    +199           18.1   +18.1
tinygrad/llm/cli.py                            199    +199           15.5   +15.5
tinygrad/runtime/support/memory.py             196     +10           18.1    +0.2
tinygrad/engine/realize.py                     192     +13           19.8    +2.6
tinygrad/schedule/indexing.py                  185      -9           15.3    -0.4
tinygrad/runtime/support/mlx/mlxdev.py         182    +182           20.6   +20.6
tinygrad/nn/state.py                           178     -68           16.1    -2.5
tinygrad/nn/__init__.py                        176     +24           17.9    -1.0
tinygrad/uop/spec.py                           174      +9           26.1    -0.5
tinygrad/codegen/opt/search.py                 170      +2           15.7    +0.1
tinygrad/runtime/ops_webgpu.py                 166      +0           15.0    -0.0
tinygrad/runtime/ops_metal.py                  149      +6           13.5    -0.2
tinygrad/runtime/support/nv/nvdev.py           141      +2           19.9    -0.1
tinygrad/llm/gguf.py                           133    +133           26.4   +26.4
tinygrad/codegen/opt/heuristic.py              133      +1           16.1    +0.1
tinygrad/renderer/__init__.py                  130      +0           14.5    +0.4
tinygrad/callify.py                            125    +125           17.9   +17.9
tinygrad/schedule/multi.py                     124     -39           22.3    +0.2
tinygrad/codegen/opt/tc.py                     122      +2           25.3    +0.2
tinygrad/schedule/__init__.py                  120    +120           13.7   +13.7
tinygrad/runtime/ops_disk.py                   120     +11           14.5    -0.2
tinygrad/runtime/ops_cuda.py                   115      -2           15.9    +0.1
tinygrad/runtime/support/c.py                  114     -48           16.4    +1.4
tinygrad/runtime/ops_cl.py                     114      -5           18.2    -0.8
tinygrad/runtime/ops_cpu.py                    113      +2           15.9    -0.1
tinygrad/codegen/late/expander.py              113      +7           19.5    +0.1
tinygrad/codegen/__init__.py                   113      +7           17.9    +1.7
tinygrad/runtime/support/compiler_amd.py       112      +0           13.0    +0.3
tinygrad/codegen/simplify.py                   112      +3           23.3    +1.4
tinygrad/nn/optim.py                           111      +5           19.3    +0.4
tinygrad/gradient.py                           104     +30           24.5    -0.8
tinygrad/runtime/support/compiler_mesa.py      103      +3           16.4    -0.6
tinygrad/viz/js/worker.js                      102     +17            4.6    -0.0
tinygrad/runtime/ops_tinyfs.py                 101      +0           11.6    +0.0
tinygrad/renderer/wgsl.py                       96      +2           25.6    +0.0
tinygrad/runtime/graph/metal.py                 87      +1           13.5    -1.1
tinygrad/codegen/gpudims.py                     87      +3           17.6    -0.0
tinygrad/renderer/amd/elf.py                    86     +86           13.2   +13.2
tinygrad/uop/divandmod.py                       85     +10           21.4    -0.7
tinygrad/runtime/ops_rdma.py                    84     +84           19.1   +19.1
tinygrad/runtime/support/amd.py                 83      +0           21.4    -0.2
tinygrad/runtime/support/compiler_cpu.py        82     +10           15.2    +0.2
tinygrad/runtime/support/compiler_cuda.py       81      +0           17.9    +0.4
tinygrad/function.py                            75     +75           13.6   +13.6
tinygrad/runtime/support/objc.py                61      -1           12.9    -0.1
tinygrad/runtime/support/elf.py                 60      +0           19.1    +0.1
tinygrad/uop/validate.py                        60      +2           26.0    -0.5
tinygrad/renderer/amd/__init__.py               59     +59           15.1   +15.1
tinygrad/runtime/graph/cuda.py                  56      -2           18.0    -1.1
tinygrad/runtime/ops_hip.py                     55      +0           18.8    +0.3
tinygrad/uop/__init__.py                        55      +2           17.2    -0.2
tinygrad/mixin/reduce.py                        51     +51           23.8   +23.8
tinygrad/schedule/memory.py                     50     +50           17.3   +17.3
tinygrad/schedule/allreduce.py                  44     +44           20.5   +20.5
tinygrad/runtime/ops_null.py                    37      -2           15.3    +1.6
tinygrad/runtime/support/compiler_qcom.py       30     +30           15.5   +15.5
tinygrad/mixin/dtype.py                         22     +22            9.5    +9.5
tinygrad/__init__.py                            12      +1            6.7    -0.1
tinygrad/mixin/creation.py                      11     +11           11.6   +11.6
tinygrad/runtime/ops_npy.py                      9      +0           16.6    +0.1
tinygrad/llm/__main__.py                         2      +2            8.0    +8.0
tinygrad/engine/memory.py                        0     -51            0.0   -20.2
tinygrad/mixin/math.py                           0    -182            0.0   -12.6
tinygrad/engine/schedule.py                      0    -168            0.0   -15.9
tinygrad/apps/llm.py                             0    -297            0.0   -14.3


total lines changes: +3468

@Discountchubbs Discountchubbs changed the title Sync: tinygrad/tinygrad: master into sunnypilot/tinygrad:master Sync: tinygrad/tinygrad:master into sunnypilot/tinygrad:master Apr 26, 2026
@Discountchubbs Discountchubbs merged commit 8164a9c into master Apr 26, 2026
50 of 54 checks passed
@Discountchubbs Discountchubbs deleted the tinygrad-sync branch April 26, 2026 06:35
@Discountchubbs Discountchubbs mentioned this pull request Apr 27, 2026
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

@Discountchubbs Discountchubbs

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.