1
0
Fork 0
Commit Graph

17 Commits (fdd7f282cbd2925f0a00e80e72857024439145bf)

Author SHA1 Message Date
nimlgen 8844a0a822
llvm jitted (#1652) 2023-08-28 20:22:44 -07:00
George Hotz 82623697a8
Move asm renderer (#1648)
* teeny changes

* teeny updates

* move to renderer
2023-08-23 10:06:43 -07:00
George Hotz db8344ab83
add noalias to llvm (#1622) 2023-08-22 09:26:01 -07:00
Roelof van Dijk 109100656f
refactor: no len if it is not needed (#1598)
Co-authored-by: Roelof van Dijk <roelof.van.dijk@vitestro.com>
2023-08-21 14:06:32 -07:00
George Hotz ad7d26c393
fix __launch_bounds__ and benchmark TC MATMUL (#1575)
* fix

* benchmark matmul
2023-08-19 10:54:39 -07:00
George Hotz e464442adf
WMMA for 7900XTX (#1563)
* go

* hip no LRU

* work

* works

* 16 TFLOPS

* 29 TFLOPS

* 30 TFLOPS

* never mind, it's 60 TFLOPS

* fix metal WMMA

* put hip alloc back
2023-08-19 09:07:23 -07:00
chenyu 11dd9b1741
symbolic codegen and exec (#1552)
* symbolic codegen and exec

* fix and add test

* no sketchy

* merge_dicts type

* dtypes._arg_int32
2023-08-16 14:43:41 -07:00
nimlgen c93e63b8b5
make TestNonFloatUOps.test_mul_bool pass on all platforms (#1557) 2023-08-16 11:34:09 -07:00
madt2709 962972ee68
Fix uops int32 for llvm (#1554)
* fix-uops-int32-llvm

* fix tests

* Ignore mypy error
2023-08-15 23:22:32 -07:00
Diogo d17ecccd78
Torch/LLVM/arm F64 support (#1551) 2023-08-15 21:21:08 -04:00
George Hotz 0b5930d406
more uops testing, who isn't passing right now... (#1522)
* more uops

* llvm refactor

* update test uops

* rest of the nodes

* ors and ands
2023-08-15 09:07:26 -07:00
geohotstan 07b79f210f
llvmir support for bool <-> float casting (#1492) 2023-08-09 13:12:52 -04:00
David Hou 1766f0c0cf
use ConstOp for valid.max == 0 (#1501)
* use ConstOp for valid.max == 0

* don't render valid for invalid load cache key

* Update linearizer.py

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
2023-08-09 00:01:59 -07:00
George Hotz d24f936501
just cmplt (#1493)
* just cmplt

* fix maximum

* don't save, there's no backward

* ugh, no slot either

* eq is a scam
2023-08-08 13:58:10 -07:00
nimlgen 932dad1a2b
fix cast bool->float in llvmir (#1480)
Closes #1479
2023-08-07 21:30:51 -07:00
Diogo e2af95c2f8
moved global_max and local_max to LinearizerOptions also added assert for max bufs (#1446) 2023-08-05 18:23:18 -07:00
George Hotz 84c430355e
fix backends for new style (#1443)
* fix backends for new style

* fix method cache

* fix fakeless

* llvm blacklist

* fix kernel optimizer
2023-08-05 11:07:04 -07:00