nimlgen
f381ad3b29
fix
2026-08-17 05:27:40 -07:00
nimlgen
f96f3a32f6
fst
2026-08-17 01:45:40 -07:00
nimlgen
5824a210bd
x
2026-08-17 00:06:44 -07:00
nimlgen
36fb8803d0
Dx
2026-08-16 23:06:43 -07:00
nimlgen
9213223c82
revrt
2026-08-16 22:52:31 -07:00
nimlgen and GitHub
38655e8123
Merge branch 'master' into hcq2_one_submitter
2026-08-17 08:42:51 +03:00
chenyu and GitHub
e25f86721d
more torch backend fixups ( #17562 )
2026-08-16 21:31:53 -04:00
chenyu and GitHub
138fb4a783
delete dead DType.scalar [PR] ( #17561 )
2026-08-16 21:12:17 -04:00
chenyu and GitHub
bfd4048abf
no dtype in vconst_like [PR] ( #17560 )
2026-08-16 21:06:38 -04:00
chenyu and GitHub
057a18a07c
fix emulated long cast to double ( #17559 )
2026-08-16 20:52:42 -04:00
chenyu and GitHub
c30bf116b7
few torch_backend fix ( #17558 )
...
* few torch_backend fix
* fix
2026-08-16 20:07:25 -04:00
nimlgen
d435280c05
Dx
2026-08-16 13:16:25 -07:00
nimlgen
5ff2028a87
x
2026-08-16 08:04:02 -07:00
nimlgen
2fcb40d400
x
2026-08-16 14:02:23 +00:00
nimlgen
3c6f47b5f6
simpler
2026-08-16 13:55:26 +00:00
nimlgen
8bdcc1b6d1
simpler
2026-08-16 13:31:12 +00:00
nimlgen
52c07bf381
x
2026-08-16 13:17:23 +00:00
nimlgen
90823835c8
x
2026-08-16 13:04:44 +00:00
nimlgen
8063095c71
x
2026-08-16 12:15:05 +00:00
nimlgen
0a2d1e7689
hcq2: c submitter
2026-08-16 11:35:30 +00:00
nimlgen and GitHub
e7bf2a811d
iface in device ( #17554 )
...
* iface in device
* drop
* move
* sorry
2026-08-16 12:09:24 +03:00
George Hotz and GitHub
e688e07758
add max_shape/max_numel to mixins + pad_to ( #17553 )
2026-08-15 20:05:10 -07:00
nimlgen and GitHub
97022960ae
device: fix remap ( #17549 )
2026-08-16 01:01:36 +03:00
chenyu and GitHub
417563ca20
fix webgpu is_nan [pr] ( #17551 )
2026-08-15 16:31:19 -04:00
chenyu and GitHub
5ca87f1bac
fix cast to weak twice [pr] ( #17548 )
...
also no gradient for weak target
2026-08-15 12:39:54 -04:00
chenyu and GitHub
26cbadd69a
no pm_fold_cast_const in full_rewrite_to_sink [PR] ( #17547 )
2026-08-15 10:18:59 -04:00
chenyu and GitHub
fae893753b
no pm_fold_cast_const in get_kernel_graph [pr] ( #17546 )
...
* no pm_fold_cast_const in get_kernel_graph [pr]
* maybe
2026-08-15 09:57:05 -04:00
nimlgen and GitHub
c17849a1f8
disk: neable polling ( #17538 )
2026-08-15 15:03:28 +03:00
chenyu and GitHub
539a03343a
no casted const from sub and div [pr] ( #17543 )
2026-08-15 07:47:40 -04:00
qazal and GitHub
a57569349c
renumber invalids before callify ( #17542 )
...
* renumber invalids before callify
* change
* Revert "change"
This reverts commit 6f4df1541e
2026-08-15 18:00:54 +09:00
qazal and GitHub
5c43a89fb1
precompile_backward tests for sched_cache ( #17544 )
...
* work
* back
* work
* keep +
2026-08-15 15:25:31 +09:00
qazal and GitHub
e6f5bb9c09
simple test for Invalid clone cache miss regression ( #17541 )
...
* simple test for Invalid clone cache miss regression
* xfail
* _
2026-08-15 11:00:21 +09:00
chenyu and GitHub
4b0525e594
no pm_fold_cast_const in UOp.simplify and hcq2 [pr] ( #17540 )
2026-08-14 21:39:47 -04:00
chenyu and GitHub
64ccbde3bb
clean up STACK with a const [PR] ( #17539 )
2026-08-14 21:07:01 -04:00
chenyu and GitHub
6ea665ed66
remove pm_fold_cast_const from dtype decomp [pr] ( #17536 )
...
* remove pm_fold_cast_const from dtype decomp [pr]
* fix
2026-08-14 16:24:51 -04:00
nimlgen and GitHub
0725acc392
reenable hcq2 ci ( #17532 )
2026-08-14 23:13:18 +03:00
wozeparrot and GitHub
4a1f32977c
gptoss: default GROUPED_MOE=1 ( #17537 )
2026-08-14 12:24:37 -07:00
chenyu and GitHub
13c381b0c0
remove pm_fold_cast_const from initial symbolic [pr] ( #17535 )
...
interestingly it gives more accurate numerics when composing const like log10
2026-08-14 14:14:40 -04:00
chenyu and GitHub
ac7067ac60
fix deconstruct_function for python 3.11 ( #17534 )
2026-08-14 13:38:34 -04:00
chenyu and GitHub
80169c6758
remove where push cast to branches from sym [pr] ( #17533 )
...
* remove where push cast to branches from sym [pr]
not really needed and one less place that generates casted weak const when it's not needed
* fix
2026-08-14 12:57:56 -04:00
nimlgen and GitHub
adacaa3e17
hcq2 fix hangs ( #17529 )
2026-08-14 16:23:59 +03:00
chenyu and GitHub
89ab344c42
fix assign into bitcast with no explicit realize ( #17531 )
2026-08-14 09:23:01 -04:00
chenyu and GitHub
25c3bd027b
remove pm_fold_cast_const in simplify_merge_adjacent [pr] ( #17530 )
2026-08-14 09:10:05 -04:00
nimlgen and GitHub
6b35220622
cpu hcq2 ( #17503 )
...
* cpu hcq2
* temp
* slop
* test with backpressure
* x
* x
* x
* x
* x
* x
* Dx
* save reverts
* um?
* x
* x
* call from py
* x?
* x
* submitters gone
* x
* x
* z
* Dx
* Dx
* x
* x
* fixes
* repl
* x
* f
* for now keep hcqbuffer
2026-08-14 15:06:03 +03:00
George Hotz and GitHub
b1859805b1
remove Ops.BIND ( #17511 )
...
* remove Ops.BIND
* param arg
* simplify that
* simplify
* cleaner
* props, not functions
* param and buffer can share
2026-08-13 23:52:16 -07:00
qazal and GitHub
faba071b1d
don't enter CALL body in assign fixups ( #17527 )
...
* fix python time regression in mxfp4
* this saves even more time
* s_nop test
* itertools count
* cleanup
2026-08-14 15:05:04 +09:00
qazal and GitHub
81dc8ec232
Revert "amd: fix ALL2ALL speed on amdgpu (gpt) ( #17518 )" ( #17526 )
...
This reverts commit cc6d33bde7
2026-08-14 10:39:14 +09:00
geohot
95ca5081fe
hotfix: update extra/runbook_digitalocean_mi350x
2026-08-13 17:26:39 -07:00
sirhcm and GitHub
303d1677b3
qcomcl: use qemu for compilation ( #17524 )
2026-08-13 20:16:02 -04:00
George Hotz and GitHub
673c6463f9
disable HCQ2 for AMD CI ( #17523 )
2026-08-13 13:14:40 -07:00
nimlgen
geohot
sirhcmandGitHub