George Hotz
617614beb7
add mi350x support to mmapeak ( #12784 )
2025-10-19 16:11:07 +08:00
qazal
c8ef4b60f6
viz: share match tracing and TINY device profiler ( #12783 )
...
* set a default name for the traces
* set profile_matches + renames
* profile_matches test
* traces 4 steps total
2025-10-19 14:30:07 +08:00
chenyu
350a4754a9
Update openpilot models ( #12780 )
...
* Update openpilot models
* Update slower model
* fix that
---------
Co-authored-by: Bruce Wayne <harald.the.engineer@gmail.com >
2025-10-18 20:32:35 -04:00
chenyu
30ff84d050
update test_conv2d_ceildiv_edge_case ( #12779 )
2025-10-18 16:43:32 -04:00
nimlgen
442218266d
qcom: fix profiler ( #12778 )
...
* qcom: fix profiler
* this way
2025-10-19 01:27:59 +08:00
Harald Schäfer
addc54b96c
Simplify openpilot compile3.py ( #12748 )
...
* Simpler compile3
* tests
* remove default args
* onnx file is still fp16
* self-test FP16 too
* allow test disable
* absurd tolerance
* Just do latest
* Try simplest
* use later models
* kernel count not relevant if speed is good
* dead improts
* Revert "dead improts"
This reverts commit f68c2cd15d0955ca4ee0
2025-10-18 10:12:22 -04:00
nimlgen
037f6e8fa0
qcom: ioctl for 7xx ( #12777 )
2025-10-18 20:33:14 +08:00
wozeparrot
82f10cfe2e
feat: assert on bufferview math ( #12772 )
2025-10-17 14:20:08 -07:00
chenyu
fcdf4ab37e
remove a contiguous in LARS ( #12770 )
2025-10-17 17:07:30 -04:00
nimlgen
910d698b78
system: cleanup page sizes ( #12771 )
...
* system: cleanup page sizes
* ooops
2025-10-18 02:06:42 +08:00
George Hotz
062a6d68d7
test flash attention backward ( #12762 )
...
* test flash attention backward
* TODO: fix pcontig
* end ranges
* render colors
* very big
* multiout at every level
* reset ending ranges
* fix tests
* ugh
2025-10-17 23:15:59 +08:00
George Hotz
33025b99f6
small changes from fa backward ( #12769 )
2025-10-17 22:41:18 +08:00
chenyu
e0d0d4372d
fix shape of m and v in onnx Adam with FUSE_OPTIM ( #12768 )
...
value is still slightly off but that's not onnx specific
2025-10-17 10:32:41 -04:00
qazal
bd662bea67
viz: light up program runs ( #12764 )
...
* basics work
* fix the color
* light up program events
* swap a with p
* better
2025-10-17 19:33:18 +08:00
George Hotz
c9a3464f76
those decimals never mattered ( #12760 )
...
* those decimals never mattered
* this
* improve debug
* real substitute fixes pcontig
* locals are different buffers
2025-10-17 17:16:24 +08:00
qazal
0160f034d6
viz: show display name for copy runners ( #12761 )
...
* viz: show display name for copy runners
* more u32
2025-10-17 16:59:51 +08:00
qazal
253d32b065
viz: add metadata to buffer user list ( #12758 )
...
* simple failing test
* encodings
* test passing
* key is deduped
2025-10-17 16:28:54 +08:00
George Hotz
935a60db72
bring back partial contig and flash attention ( #12756 )
...
* bring back partial contig and flash attention
* why not 2
* work
* that
* fix pcontig
2025-10-17 16:19:05 +08:00
Sieds Lykles
f6bc620169
UOp.prod and UOp.sum methods ( #12755 )
2025-10-17 10:02:01 +02:00
Sieds Lykles
d1bb5c0426
slightly flatter symbolic ( #12757 )
2025-10-17 09:58:45 +02:00
qazal
5417e4b099
viz helper cleanups ( #12754 )
2025-10-17 15:20:24 +08:00
qazal
3196a7aae3
viz: pre reqs for lighting up programs ( #12753 )
2025-10-17 15:03:21 +08:00
qazal
dfb8f9fc9e
viz: annotate buffer mutability in the memory graph ( #12750 )
2025-10-17 11:53:02 +08:00
Sieds Lykles
79c2f1ae26
remove reduce_rangless and replace with reduce_unparented ( #12749 )
2025-10-17 04:46:05 +02:00
chenyu
9561803cb0
fix assert in test_schedule ( #12745 )
...
* fix assert in test_schedule
updated kernel counts and some old tests
* fix
2025-10-16 15:39:50 -04:00
chenyu
285534ce64
delete DONT_REALIZE_EXPAND and DONT_GROUP_REDUCES ( #12744 )
...
does nothing now
2025-10-16 14:11:33 -04:00
chenyu
98239f1156
few shapetracker cleanups ( #12741 )
2025-10-16 12:43:27 -04:00
chenyu
53478c741d
relax ASSERT_MIN_STEP_TIME for space lab policy ( #12742 )
2025-10-16 11:40:36 -04:00
geohotstan
5d209ee7ec
onnx helper intermediate node output validation ( #12740 )
...
* start
* update comments
* good
* add comments and better printing
* done
2025-10-16 11:17:47 -04:00
Christopher Milan
bce2bc0465
Revert "use RTLD_GLOBAL on macos" ( #12738 )
...
This reverts commit 89fe3e574d
2025-10-16 10:07:21 -04:00
chenyu
f34f26bca0
fix gpt2 with benchmark ( #12736 )
...
`CPU=1 python3 examples/gpt2.py --benchmark 128` works now
2025-10-16 09:55:20 -04:00
Sieds Lykles
55db1b0e0e
reduce where that is cut from two sides ( #12733 )
...
* better rule
* correct pattern
* shorten line
2025-10-16 15:25:15 +02:00
nimlgen
cf9baeea61
Revert "nv: check if jitlink is avail ( #12731 )" ( #12735 )
...
This reverts commit a069a45d14
2025-10-16 20:41:49 +08:00
George Hotz
8be7844b2e
use apply uop for assign to fix assign metadata ( #12732 )
...
* use apply uop for assign
* fix metadata for assign
* fix backward metadata
* those aren't real tests
2025-10-16 20:34:12 +08:00
nimlgen
3aa2277b8f
nv: usb4 ( #12696 )
...
* hackish
* prog
* match
* l
* simpler
* refactor
* not osx
* apple things
* tiny changes
* fix mask
* match fix
* nn
2025-10-16 20:11:19 +08:00
nimlgen
a069a45d14
nv: check if jitlink is avail ( #12731 )
2025-10-16 19:58:50 +08:00
George Hotz
a498ec9c18
cleanup names of postrange + fast FUSE_OPTIM ( #12730 )
...
* cleanup names of postrange
* make FUSE_OPTIM not slow
* delete junk in def r
2025-10-16 19:38:31 +08:00
Sieds Lykles
8f740e07ff
no broadcasting/vectors in reduce collapse ( #12729 )
2025-10-16 13:22:57 +02:00
qazal
533f18b22c
viz: add trace data for inflight buffers ( #12728 )
...
* viz: add trace data for inflight buffers
* add test_inflight_buf
* temp stores the keys
* update tests / use Tensor.ones
2025-10-16 19:15:03 +08:00
George Hotz
af4479c169
faster stable diffusion load ( #12725 )
...
* faster stable diffusion load
* failing tests
2025-10-16 18:31:59 +08:00
nimlgen
e7c057d5dc
system: alloc_sysmem return view ( #12724 )
...
* system: alloc_sysmem return view
* e
2025-10-16 17:55:01 +08:00
nimlgen
b86a33a312
ptx: support bw ( #12722 )
2025-10-16 15:38:08 +08:00
nimlgen
b8cd66c7a2
nv: support all gb20x and small bar ( #12721 )
2025-10-16 15:37:54 +08:00
George Hotz
1d1e1d9d88
delete the ShapeTracker ( #12720 )
...
* delete the ShapeTracker
* fix tests
* fix more
* fix gc test
2025-10-16 15:36:22 +08:00
George Hotz
592e86f6f5
remove UOp.st ( #12716 )
...
* remove UOp.st
* fix tests
* torch backend disable
2025-10-16 14:44:09 +08:00
wozeparrot
cc2dfe22f5
tinyfs: fetch file utility ( #12719 )
2025-10-15 23:38:56 -07:00
nimlgen
3ed543f956
system: reorder funcs + barrier on macos ( #12714 )
2025-10-16 14:38:01 +08:00
qazal
b77bdbbc62
viz: count unpickle in server startup time ( #12715 )
...
* viz: count unpickle in server startup time
* type checking
2025-10-16 13:07:46 +08:00
George Hotz
7c19db00f1
remove st from jit/split_reduceop ( #12713 )
...
* remove st from jit
* fix by merging reshapes
* no st usage in rangeify
* hmm, stop early works
* fix speed regressions
2025-10-16 12:50:58 +08:00
qazal
069177c1be
trace buffer producer and consumers ( #12639 )
...
* trace buffer producer and consumers
* work
* generic colored util
* fix batched
* basic clicking works
* generic javascript that works for producer and consumers
* keep focused shape
* idle time
* timings for producer and consumers dedup
* from sd test
* tiny cleanups
* timeline
* work
* up to here
* assert
* list it
* work
2025-10-16 11:11:31 +08:00
