chenyu and GitHub
13ca9bd8a6
remove dtypes.index again ( #17104 )
...
also reverted some dtype change, the split made things needlessly complicated
2026-07-20 20:30:04 -04:00
nimlgen and GitHub
2b96d64496
hcq2: tiny opts and fixes ( #17092 )
2026-07-20 18:46:52 +03:00
qazal and GitHub
abba2aebda
llama: correct fused qkv shape assert ( #17086 )
2026-07-20 15:51:21 +09:00
qazal and GitHub
1cf8f2f68c
llama: inplace amax update ( #17064 )
...
* llama: inplace amax update
* remove amax_out return
* work
* fit
* work
* work
* keep
* diff cleanup
2026-07-20 15:05:41 +09:00
chenyu and GitHub
9970a0aad0
fix Tensor << Tensor for x86 ( #17082 )
...
* fix Tensor << Tensor for x86
* torch
2026-07-19 12:31:05 -04:00
nimlgen and GitHub
232529ce88
hcq2: simpler sync ( #17069 )
...
* x
* y
* n
2026-07-18 16:27:44 +03:00
qazal and GitHub
7dd3422c63
llama: replace two stage amax with atomics ( #17063 )
...
* atomic amax in c kernels
* quantize fp8 UOp kernel
* diff
2026-07-17 19:27:10 +09:00
wozeparrot and GitHub
a836c3822a
gptoss: 3d mx block scale ( #17062 )
2026-07-16 23:30:24 -07:00
George Hotz and GitHub
61e104bdfb
use UOp.wmma everywhere ( #17045 )
...
* use UOp.wmma everywhere
* fix
2026-07-16 10:40:48 -07:00
nimlgen and GitHub
e33e96415f
hcq2: tiny cleanupg ( #17042 )
2026-07-16 16:14:54 +03:00
George Hotz and GitHub
8b0dd870ce
use wmma helper ( #17038 )
2026-07-15 23:25:17 -07:00
nimlgen and GitHub
4257939e50
remove copyin/copyout from Buffer ( #17020 )
...
* remove copyin/copyout from Buffer
* x
* x
* x
* x
2026-07-14 19:47:22 +03:00
qazal and GitHub
939f28d571
fused qkv rope custom kernel ( #17021 )
...
* work
* fused qkv_norm
* work
* speed
* not that yet
* test cleanup
* just clone
* remove .realize()
* cleanup tests
2026-07-15 01:08:42 +09:00
nimlgen and GitHub
e4bdc529c4
hcq2 ci ( #17008 )
...
* hcq2 ci
* x
2026-07-13 19:29:08 +03:00
qazal and GitHub
62ad646d1c
llama: gemm/fa backward speedups (gpt 5.6) ( #17007 )
...
* fp8 atb gemm speedup
* work
* revert
* fa bw faster
2026-07-14 00:27:00 +09:00
nimlgen and GitHub
4d2becddf8
hcq2: spec=2 ( #17006 )
...
* hcq2: spec=2
* hcq: isolate HCQ spec rules
* chq
* move
2026-07-13 18:13:33 +03:00
George Hotz and GitHub
03ecad9486
full removal of dtype.vec ( #16996 )
...
* full removal of dtype.vec
* fix typo
2026-07-12 09:25:50 -07:00
qazal and GitHub
cae6696d75
llama: split current and next amax state ( #16993 )
2026-07-12 18:52:33 +09:00
qazal and GitHub
9ea7de020b
llama: fuse grad scale in gemm epilogue ( #16992 )
2026-07-12 18:15:01 +09:00
qazal and GitHub
73b38854b4
llama: aT_b fp8 gemm ( #16991 )
2026-07-12 16:57:10 +09:00
chenyu and GitHub
047a467bf9
delete UOp._stack and UOp.vectorize [PR] ( #16988 )
2026-07-11 15:26:55 -04:00
nimlgen and GitHub
928af24b74
hcq2: mini speed ups ( #16984 )
2026-07-11 15:48:07 +03:00
qazal and GitHub
75a4bfddc9
fp8 gemm tests including fused scales ( #16980 )
...
* fp8 gemm tests matching fused scales
* work
* diff
2026-07-11 12:33:21 +09:00
nimlgen and GitHub
52e060fd7a
hcq2: link cache ( #16965 )
...
* x
* x
* mini
* cleaner
2026-07-10 23:47:16 +03:00
chenyu and GitHub
ba2c68b1ed
start dtypes.index [PR] ( #16949 )
2026-07-09 15:12:23 -04:00
nimlgen and GitHub
c9f6b2e42b
hcq2: prereq for linker ( #16945 )
2026-07-09 17:21:21 +03:00
qazal and GitHub
6dbc35b8e0
fix am_smi for non interactive shell ( #16942 )
2026-07-09 13:32:59 +09:00
wozeparrot and GitHub
2971343a60
fa: support D=64 and attn sink ( #16935 )
2026-07-08 20:03:47 -07:00
nimlgen and GitHub
52b84adc2a
hcq2: tiny changes ( #16934 )
2026-07-08 23:17:28 +03:00
George Hotz and GitHub
fdffc6c0c8
remove dtypes base ( #16931 )
...
* remove dtypes base
* find/replace bug
* vcount is junk too
2026-07-08 11:48:11 -07:00
George Hotz and GitHub
63cb1369cb
remove ImageDType (glm) ( #16918 )
...
* remove ImageDType (glm)
* gpt work
* better
* is_image_shape
* cleanups
* remove crap from index
* revert crap python change
* image index with float, this needs to be fixed
* simp
2026-07-08 11:27:52 -07:00
chenyu and GitHub
e0c70b6b82
mixin/op.py [PR] ( #16926 )
...
* mixin/op.py [PR]
* that
2026-07-07 21:45:52 -04:00
George Hotz and GitHub
2fc7e5341b
final removal of PtrDType (glm) ( #16913 )
...
* final removal of PtrDType (glm)
* junk
2026-07-07 13:37:33 -07:00
nimlgen and GitHub
91cefdb52a
hcq2: rewrite patch split ( #16907 )
...
* x
* new patch splitter
* style
* x
* x
2026-07-07 20:43:52 +03:00
George Hotz and GitHub
6f1a983493
Ops.GROUP has no shape, like sink ( #16906 )
...
* Ops.GROUP has no shape, like sink
* readme
2026-07-07 08:20:38 -07:00
George Hotz and GitHub
6e44176cfe
fix flash attention example ( #16900 )
...
* fix spec for amd_copy_matmul
* remove shaped wmma
* fix flash attention example?
* dead code
2026-07-06 20:51:32 -07:00
George Hotz and GitHub
d94ad4444e
fix spec for amd_copy_matmul + remove SHAPED_WMMA ( #16897 )
...
* fix spec for amd_copy_matmul
* remove shaped wmma
2026-07-06 19:00:10 -07:00
George Hotz and GitHub
40d112d4d6
remove ptr=True from index ( #16898 )
2026-07-06 18:34:18 -07:00
George Hotz and GitHub
9512dc30f4
remove vec from const ( #16889 )
...
* remove vec from const
* reject in spec
* remove all vector dtypes
* not needed
* remove that
* clean up invalid
2026-07-06 12:23:07 -07:00
nimlgen and GitHub
47f9677f67
hcq2: shuffle passes ( #16886 )
2026-07-06 21:56:55 +03:00
wozeparrot and GitHub
35522af0df
gptoss: use quantize_mxfp8 ( #16881 )
2026-07-05 17:48:24 -07:00
nimlgen and GitHub
62cde59cb5
hcq2: tiny changes ( #16880 )
2026-07-06 00:24:55 +03:00
nimlgen and GitHub
17e4f9b6f9
hcq2: simpler ( #16878 )
2026-07-05 18:26:19 +03:00
George Hotz and GitHub
a145fdce4b
remove the .gep method ( #16867 )
2026-07-04 14:07:09 -07:00
nimlgen and GitHub
fdf434062b
hcq2: simpler ( #16862 )
...
* x
* x
* x
2026-07-04 22:07:37 +03:00
qazal and GitHub
07f7383d29
llama: remove unused bf16 assembly gemm ( #16859 )
...
* only hk bf16 gemm
* rm asm gemm
* more cleanup
* half isn't supported in asm gemm anymore
* more test edits
* unused
* remove TestMagicGu
* uop gemm is still tested
* minimal diff
2026-07-04 18:41:54 +09:00
nimlgen and GitHub
d73bca617b
hcq2: fix deps tracking ( #16841 )
2026-07-03 15:51:25 +03:00
nimlgen and GitHub
c2c784e028
hcq2: rewrite ( #16775 )
...
* hcq2: rewrite
* rw
* mid
* jit
* tmp
* x
* x
* revert
* x
* x
2026-07-03 00:00:08 +03:00
chenyu and GitHub
2e62dd308d
move fs_load and fs_store to nn ( #16821 )
...
similar to safe_load and safe_save
2026-07-01 22:01:15 -04:00
chenyu and GitHub
fd7fd9cdca
move GETADDR device to its arg ( #16808 )
2026-07-01 09:14:07 +03:00