github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-04-07 03:00:26 -04:00
Code Issues Packages Projects Releases Wiki Activity
3,638 Commits 241 Branches 14 Tags
969b57f0fe2e4e267de48b841d953f8e4e3e8a96
Commit Graph

3638 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
chenyu
969b57f0fe enable symbolic_ops and jits test of two vars (#3513) 2024-02-27 11:17:46 -05:00
wozeparrot
ea4b8e5b1f feat: don't hardcode the arch (#3511) 2024-02-27 07:58:03 -05:00
Francis Lam
11da65bccd test/external/fuzz_linearizer: add a FUZZ_MAX_SIZE option (#3455) 
* test/external/fuzz_linearizer: add a FUZZ_MAX_SIZE option

this allows us to limit the size of the kernel and reduce running
times by avoiding ones that take a long time

* fix spacing and re-order to put parameters together
 2024-02-27 07:34:59 -05:00
qazal
a29cd6d464 run f64 increased precision tests on remu (#3509) 
* run the test in CI

* temp: use the pre-release

* Revert "temp: use the pre-release"

This reverts commit 28e8571421.
 2024-02-26 18:01:07 -05:00
chenyu
b1426f3a4c cleanup SumNode mod (#3503) 2024-02-26 11:10:55 -05:00
chenyu
61605ccc69 Remove special case of SumNode div SumNode (#3502) 2024-02-26 09:42:06 -05:00
Francis Lam
39d75f0d58 test_linearizer_failures: add more METAL examples (#3495) 
these were obtained from running fuzz_linearizer on METAL
 2024-02-26 10:19:05 +01:00
chenyu
b154089884 float64 function support for HIP (#3492) 
* float64 function support for HIP

* not CI
 2024-02-24 09:46:20 -05:00
chenyu
35aff8b0c2 properly exclude PYTHON backend and support of half (#3491) 
should be able to run in CI with python 3.12
 2024-02-24 09:22:06 -05:00
David Friehs
2fe98b64bb fix Tensor.split not passing dim to Tensor.chunk (#3490) 2024-02-24 07:53:11 -05:00
Caleb Bunch
b41761488d change specific string 'CLANG' to DEVICE variable in abstractions2.py (#3488) 2024-02-24 07:51:39 -05:00
chenyu
c032df520b minor symbolic type related cleaups (#3489) 2024-02-23 16:44:43 -05:00
Carson Radtke
15df9406d6 fix exec_alu(UnaryOps.SQRT, <...>, (0,)) + add test (#3487) 
* fix exec_alu(UnaryOps.SQRT, <...>, (0,)) + add test

* sqrt(0) != nan

* fix tabs
 2024-02-23 18:28:00 +01:00
nimlgen
52567da07f jit grapher simplified (#3478) 2024-02-23 16:20:16 +01:00
George Hotz
2113e1eb63 move all reduces to the end in lazy (#3475) 
* move all reduces to the end in lazy

* apply as reshape, not permute
 2024-02-23 15:49:11 +01:00
David Hou
5cfcc2a8d7 support MLB reshaping on-axis for evenly sharded (#3484) 
* support MLB reshaping on-axis for evenly sharded

* update test

* not -> !=
 2024-02-23 07:51:36 -05:00
chenyu
358a24eae6 symbolic use mod for rmod and use floordiv for rfloordiv (#3485) 2024-02-23 01:05:13 -05:00
nimlgen
6d048a0c0b cache collector optimizations are allowed only for kernel operations (#3476) 2024-02-22 12:26:57 +01:00
George Hotz
7698781389 Revert "wmma: add CUDA tensor core (#3464)" (#3474) 
This reverts commit e9cef13f0b.
 2024-02-22 11:58:16 +01:00
Francis Lam
e9cef13f0b wmma: add CUDA tensor core (#3464) 2024-02-22 11:57:08 +01:00
wozeparrot
57678012e1 Upload correct benchmark artifact (#3471) 
* fix: correct filename

* fix: why is this .py?
 2024-02-22 01:14:16 -05:00
chenyu
ab40c0cf93 clean up long lines in symbolic (#3469) 2024-02-21 21:57:44 -05:00
chenyu
7c0fc40123 enable test IMAGE=2 PYTHON=1 python3 test/test_ops.py TestOps.test_simple_conv2d (#3468) 2024-02-21 18:30:12 -05:00
chenyu
77d2a4c12a regenerate kernel dataset after reduce arg to axis change (#3467) 
```
./extra/optimization/generate_dataset.sh
gzip /tmp/sops
mv /tmp/sops.gz extra/datasets/
```
 2024-02-21 18:16:13 -05:00
David Hou
f513c37e64 support same uidx in multiple shape positions (#3205) 
* support same uidx in multiple shape positions

* rename var

* update comment

* add contiguous index check to global_store too

* update comment

* small change

* is this better?

* smh

* smaller change?

* get rid of more changes

* get rid of more changes

* is this even making anything better

* comment

* fix test

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
 2024-02-21 19:37:03 +01:00
chenyu
1eb24af63b fix softmax and log_softmax for 0d tensor (#3463) 
matched torch to take axis \in [-1, 0] and used axis=None internally
 2024-02-21 11:30:30 -05:00
George Hotz
871ba73e65 _reduce_op is axis based now (#3462) 
* _reduce_op is axis based now

* axis_

* update lin failures

* disable that

* fix shape
 2024-02-21 16:36:31 +01:00
George Hotz
22a90cbb15 change frontend reduce API to use axis (#3460) 
* change frontend API to axis

* switch lazy to also take axis input
 2024-02-21 12:26:17 +01:00
chenyu
6c1063ba39 add mypy --strict-equality to pre-commit (#3458) 
matched ci mypy behavior
 2024-02-21 03:41:05 -05:00
chenyu
02683a8659 gate the cast before movements in lazy (#3452) 
it made gpt2 slower (2ms -> 2.5ms on 3090, 7ms -> 8ms on M1 Max with BEAM=2).
disabled it in gpt2 benchmark before understanding the full issue
 2024-02-20 09:36:22 -05:00
chenyu
0d326a48b8 fix LtNode simplification when lhs and rhs contain same variables (#3451) 
* fix LtNode simplification when lhs and rhs contain same variables

`(Variable("a", 1, 5) < Variable("a", 1, 5))` should eval to `NumNode(0)`

* fix with less perf impact
 2024-02-20 09:06:55 -05:00
George Hotz
1b6e890ef2 uops flop counter (#3373) 
* factor out winograd functions

* test counter

* uops flop counter

* more correct

* ish

* correct

* cleanup

* tests for uops flop counter

* tests still fail

* fix symbolic uops flop cnt

* fix symbolic uops flop cnt

* hmm, it's an alu

* uops alu resolve

* relax that
 2024-02-20 09:36:30 +01:00
Patrick Tsai
9dd64b1f5f Fix python cast uint/int overflow (#3448) 
* Fix numpy uint/int overflow

* lol

* Works

* Update

* Move overflow test to float64/float32

* One line

* Update

* One more

---------

Co-authored-by: Patrick Tsai <patosai@users.noreply.github.com>
 2024-02-20 09:20:43 +01:00
qazal
7864fb69d1 delete MovementOps (#3434) 
* delete MovementOps

* keep extra/to_movement_ops.py
 2024-02-19 23:21:44 +01:00
nimlgen
015d414786 fix gpu page fault by ensuring code memory persistence during execution (#3435) 
* fix pf for exec image memory

* no new noqa: E501
 2024-02-19 13:40:53 +01:00
Daniel Yeh
0a4029c519 fix path to models folder (#3442) 
Co-authored-by: Chen-Chen Yeh <ge96noj@mytum.de>
 2024-02-19 13:35:57 +01:00
Patrick Tsai
ac9d94a068 Cast correctly in python emulator (dtype tests pass) (#3446) 
* Cast correctly in python emulator

* Update test yml and fix lint

* make ruff pass

* mypy passes

---------

Co-authored-by: Patrick Tsai <patosai@users.noreply.github.com>
 2024-02-19 13:34:02 +01:00
chenyu
ddec76e9c4 remove unused LtNode.__floordiv__ (#3445) 2024-02-18 22:12:54 -05:00
chenyu
86efdf0b34 remove create_rednode (#3444) 
handle Node collapsing into NumNode similar to OpNode
 2024-02-18 21:08:19 -05:00
chenyu
2da734920e use __getnewargs__ to fix unpickling Variable (#3441) 
it's recommended to use __getnewargs__ to update the args of classes that use __new__ when unpickling.
It's preferred because it does not change the __new__ behavior.
 2024-02-18 10:28:37 -05:00
nimlgen
5647148937 fix hip invalid ordinal (#3440) 2024-02-18 08:31:44 -05:00
chenyu
8c0e85fdaf limit symbolic substitute var_vals to have NumNode or Variable (#3438) 
this can greatly reduce the posiible output types of substitute
 2024-02-18 01:29:44 -05:00
George Hotz
6b4f734dc1 hotfix: better copy stats 2024-02-16 16:52:39 +01:00
George Hotz
c7fda10aa0 hotfix: disk doesn't sync 2024-02-16 16:46:48 +01:00
chenyu
230fc33d5b limit sint to be Union[int, Variable, MulNode, SumNode] (#3430) 
* limit sint to be Union[int, Variable, MulNode, SumNode]

these are the only allowed nodes in a Tensor shape

* stride can be sint
 2024-02-16 10:05:46 -05:00
George Hotz
fe97a85014 the compiler is a driver (#3427) 2024-02-16 10:18:09 +01:00
zku
2d702ca073 If feasible, do not truncate float64 down to float32 in cstyle renderer (#3420) 
* do not truncate float64 precision

* use l suffix to try avoid overload confusion

* long line, ruff bloats the function otherwise

* fmt

* remove long double suffix (l), it's sufficient to have the float32 (f) suffix to avoid function overload ambigouity; add test showcasing rtol=1e-12 precision increase, the test fails without the renderer changes

* use more reasonable test values, same as test_int_to_float_unary_func

* disable test for CUDACPU, does not support half and segfaults on some operations per dtypes_alu test

* disable test for HIP, renderer does not support f64 precision

* do not use noqa E501, break up condition
 2024-02-16 10:08:59 +01:00
chenyu
30f26279c5 add back "CPU" in test_onnx_backend supports_device (#3426) 
the onnx tests were all skipped.
 2024-02-16 00:49:30 -05:00
xarkes
28a8b72024 Remove Interpreted device & remaining CPU/TORCH ref (#3423) 
* Remove Interpreted device & remaining CPU/TORCH ref

* Oops

* supports_device was useful

* Fix doc wording

---------

Co-authored-by: chenyu <chenyu@fastmail.com>
 2024-02-16 00:30:21 -05:00
chenyu
6efa68f97b remove use of TORCH in pre-commit (#3424) 
it's silently using DEFAULT after removing TORCH
 2024-02-15 19:38:37 -05:00