Repository navigation

Cuda 12 sm86 - #12

Open

li0nr wants to merge 99 commits into

tinygrad:masterfrom

li0nr:cuda-12-sm86

Commits on Aug 25, 2026
- adding support for new arch sm_50
  li0nr
  committed
  9a00e19
- correctly parsing bfe/bfi immediate.
  li0nr
  committed
  1c64719
Commits on Aug 26, 2026
- addign sm_86!
  li0nr
  committed
  d3a6ddf
- cvt to bf16
  li0nr
  committed
  2561103
- supporting mov packed in parser.
  li0nr
  committed
  7b59cc9
Commits on Aug 27, 2026
- supporting setp.f16
  li0nr
  committed
  a9d55a4
- cvt f32 to f16.
  li0nr
  committed
  ac233be
- cvt.f32.bf16
  li0nr
  committed
  3af2b61
- .b32 destination support
  li0nr
  committed
  beabd31
- fma.rn.bf16
  li0nr
  committed
  01ae9e8
- cvt bulk fixes f16.
  li0nr
  committed
  f8afa7f
Commits on Aug 28, 2026
- supporting setp.f16
  li0nr
  committed
  f335dc8
- fma f16
  li0nr
  committed
  ea5eb41
- mul f16
  li0nr
  committed
  b629f8a
- add.f16, sub.f16, and neg.f16
  li0nr
  committed
  8fc35f1
Commits on Aug 29, 2026
- mma in ocelot
  li0nr
  committed
  ab51d99
- neg bf16
  li0nr
  committed
  232c2cb
Commits on Aug 30, 2026
- b32 fallback f32.
  li0nr
  committed
  6def197
Commits on Sep 13, 2026
- mov b16, f16 dest, vector operand
  li0nr
  committed
  38a5884
- max f16
  li0nr
  committed
  0a29a4f
- ex2 f16
  li0nr
  committed
  7c4f020
- mma more shapes
  li0nr
  committed
  059bf51
- mma tf32, more good
  li0nr
  committed
  b672a13
- Fns added.
  li0nr
  committed
  afd87e3
- szext added.
  li0nr
  committed
  5ad6354
- bmsk added.
  li0nr
  committed
  97a1014
- lop3 added.
  li0nr
  committed
  527f256
- dp2a/dp4a.
  li0nr
  committed
  ac1cf2d
- tanh added.
  li0nr
  committed
  5cec974
- max/min f32.
  li0nr
  committed
  4ed317c
- add/sub fp and rounding.
  li0nr
  committed
  f0ec0e8
- mul fp and roudning.
  li0nr
  committed
  3803fe8
- fma roudning.
  li0nr
  committed
  ec82a24
- mad using fma.
  li0nr
  committed
  c714ace
- div added.
  li0nr
  committed
  3500d4f
- rem fixes.
  li0nr
  committed
  ff112da
- sad fixes.
  li0nr
  committed
  0d3c2a0
- and tests.
  li0nr
  committed
  7189920
- or tests.
  li0nr
  committed
  53995a7
- xor added.
  li0nr
  committed
  f08c004
- not tests.
  li0nr
  committed
  1f02a5b
- shl tests.
  li0nr
  committed
  e0adc9f
- shl adjusted ub.
  li0nr
  committed
  cc09fc1
- shl valid.
  li0nr
  committed
  46170e4
- shr fixes.
  li0nr
  committed
  6aaa382
- shf fixes.
  li0nr
  committed
  9dfd0e2
- rcp adjustments.
  li0nr
  committed
  0e65208
- sqrt adjustments.
  li0nr
  committed
  8d51d2b
- rsqrt.
  li0nr
  committed
  0fc3832
- cos, sin rounding fixes.
  li0nr
  committed
  2c2fb50
- ex2.
  li0nr
  committed
  42d4f90
- testp.
  li0nr
  committed
  38576e2
Commits on Sep 16, 2026
- cvta param.
  li0nr
  committed
  98c7e9f
- tests correct copysign.
  li0nr
  committed
  6cbc423
- set fixes.
  li0nr
  committed
  d288c13
- setp.
  li0nr
  committed
  5039c8a
- slct.
  li0nr
  committed
  84a1a81
- cvt.
  li0nr
  committed
  7ff4629
- cvt more stuff per spec.
  li0nr
  committed
  46adbfe
- isspacep fixes.
  li0nr
  committed
  c81830e
- bfe.
  li0nr
  committed
  dafb4bc
- bfind, bfi fixes.
  li0nr
  committed
  ed1b12c
- isspacep.
  li0nr
  committed
  48155b6
- issapce cta.
  li0nr
  committed
  332802b
Commits on Sep 17, 2026
- isspace good.
  li0nr
  committed
  6ff17d0
- cvta fixes.
  li0nr
  committed
  6e4d16d
- neg.
  li0nr
  committed
  145ce39
- min.
  li0nr
  committed
  0a41bec
- max.
  li0nr
  committed
  b286ffd
- popc.
  li0nr
  committed
  ae3cf01
- add half.
  li0nr
  committed
  411b5a8
- sub half.
  li0nr
  committed
  a73ad6a
- mul half.
  li0nr
  committed
  3e35427
- fma half.
  li0nr
  committed
  d531149
- neg half.
  li0nr
  committed
  d1523ad
- abs half.
  li0nr
  committed
  e8965c9
Commits on Sep 18, 2026
- max/min half.
  li0nr
  committed
  82532eb
Commits on Sep 20, 2026
- max/min half, more things.
  li0nr
  committed
  3121fab
Commits on Sep 28, 2026
- tanh for halfs.
  li0nr
  committed
  56fd0e1
- isspace fixes.
  li0nr
  committed
  4b309db
- add fixes.
  li0nr
  committed
  1c41e02
- arithmetic half fixes.
  li0nr
  committed
  093131d
Commits on Sep 29, 2026
- fma rounding.
  li0nr
  committed
  04cb697
- astra review fixes.
  li0nr
  committed
  49a6403
- ld.lu
  li0nr
  committed
  d659c36
- st parser
  li0nr
  committed
  c3cdb71
- parser fixes, agnet found it
  li0nr
  committed
  7e26140
- prefetch text stuff.
  li0nr
  committed
  a5ed9e6
Commits on Sep 30, 2026
- red and atom stuff.
  li0nr
  committed
  6b9891d
- just parse memory inst.
  li0nr
  committed
  4bbc116
Commits on Oct 1, 2026
- more memory parsing.
  li0nr
  committed
  f96a232
- more memory parsing - mmio parsing
  li0nr
  committed
  a5f2388
Commits on Oct 3, 2026
- m16n8k16, mma.
  li0nr
  committed
  e90faa3
Commits on Oct 5, 2026
- more mma stuff.
  li0nr
  committed
  c12dbd7
Commits on Oct 6, 2026
- more mma m8n8k4
  li0nr
  committed
  4696e54
- more mma m8n8k4 f32
  li0nr
  committed
  169381c
- more mma m8n8k32
  li0nr
  committed
  0e86454
- more mma m16n8k64
  li0nr
  committed
  ac67da0
Commits on Oct 7, 2026
- mma dense done.
  li0nr
  committed
  ed2ba13

Commits

Commits on Aug 25, 2026

Commits on Aug 26, 2026

Commits on Aug 27, 2026

Commits on Aug 28, 2026

Commits on Aug 29, 2026

Commits on Aug 30, 2026

Commits on Sep 13, 2026

Commits on Sep 16, 2026

Commits on Sep 17, 2026

Commits on Sep 18, 2026

Commits on Sep 20, 2026

Commits on Sep 28, 2026

Commits on Sep 29, 2026

Commits on Sep 30, 2026

Commits on Oct 1, 2026

Commits on Oct 3, 2026

Commits on Oct 5, 2026

Commits on Oct 6, 2026

Commits on Oct 7, 2026