George Hotz cb500466c2 assembly/amd: amd_asm_matmul (#13989) ...

* amd_asm_matmul

* dsl transform

* asm roundtrip

* fixed

* less

* better

* more

* simpler

* simplify

* lil

* simpler

* compact

* work

* cleanups

* simplify

* simpler

* cleanup

* name the regs

* simp

* big simp

* big simp

* simp

* acc grid

* fast

* stuff

* fast

* simpler

* owrks

* save vgprs

* save vgprs

* Compact

* less VGPRs

* after

* SQTT support

* fastest

* faster

* lil faster

* tile regs

* faster

* readable

* one more

* simpler

* lil simpler

* NO_GLOBAL skips early globals

* stock kernel

* cleanups

* cleanups

* one b reg

* safe reg changes

* acc is compact now

* remove confusing stuff

* sregs

* lds cleanups

* vopd

2026-01-07 20:11:05 -08:00

..

amd_seb

kernel4 and 5 in uops (#11411)

2025-07-28 19:35:48 -07:00

asm

RDNA3 fp16 assembly gemm 85 TFLOPS (#13990)

2026-01-03 18:34:23 +09:00

max_kernels

extra/gemm/max_matmul: start of custom kernels for GEMM (#6926)

2025-03-19 15:04:57 +08:00

.gitignore

mi350x 1tflop bf16 gemm in extra (#13702)

2025-12-28 21:45:42 +09:00

amd_asm_matmul.py

assembly/amd: amd_asm_matmul (#13989)

2026-01-07 20:11:05 -08:00

amd_matmul.py

remove ScheduleItem and merge it with ExecItem (#13759)

2025-12-19 17:04:24 -04:00

amd_uop_matmul.py

RDNA3 fp16 assembly gemm 85 TFLOPS (#13990)

2026-01-03 18:34:23 +09:00

amx.py

rename allocator methods to not conflict [pr] (#7788)

2024-11-20 00:10:29 +08:00

cuda_matmul.py

rename allocator methods to not conflict [pr] (#7788)

2024-11-20 00:10:29 +08:00

fuzz_matmul.py

acc_dtype -> dtype (#9402)

2025-03-10 16:05:30 -04:00

gemm.c

only 62 gflops (#2629)

2023-12-05 13:28:24 -08:00

gemm.py

only 62 gflops (#2629)

2023-12-05 13:28:24 -08:00

halide_gemm.py

add halide example (#10980)

2025-06-26 16:14:57 -07:00

hip_matmul.py

rename allocator methods to not conflict [pr] (#7788)

2024-11-20 00:10:29 +08:00

intel_xmx.py

ops_gpu -> ops_cl (#12103)

2025-09-10 15:15:48 -04:00

max_matmul.py

fix: make max_matmul run again (#13085)

2025-11-03 18:09:09 -08:00

metal_conv.py

create engine folder and move code (#3948)

2024-03-26 20:38:03 -07:00

metal_matmul.py

compile fixes (#10442)

2025-06-06 18:38:37 -04:00

metal_matvec.py

compile fixes (#10442)

2025-06-06 18:38:37 -04:00

metal_uop_matmul.py

matmul example on metal showing off tensor core (#13033)

2025-10-31 19:40:36 +08:00

mi350x_uop_matmul_2.py

more mi350x matmul work (#13138)

2025-11-13 09:09:28 -08:00

mi350x_uop_matmul.py

index slicing + allclose (#13071)

2025-11-03 13:01:48 +08:00

real_pmatmul.py

pmatmul example + GB/s bugfix [run_process_replay] (#5974)

2024-08-07 22:32:11 -07:00

simple_conv.py

acc_dtype -> dtype (#9402)

2025-03-10 16:05:30 -04:00

simple_matmul.py

remove ScheduleItem and merge it with ExecItem (#13759)

2025-12-19 17:04:24 -04:00

simple_matvec.py

acc_dtype -> dtype (#9402)

2025-03-10 16:05:30 -04:00

tinygrad_nv_matmul.py

remove ScheduleItem and merge it with ExecItem (#13759)

2025-12-19 17:04:24 -04:00

torch_gemm.py

work from benchmarking tinybox red v2 (#13264)

2025-11-13 16:38:40 -08:00

triton_nv_matmul.py

remove ScheduleItem and merge it with ExecItem (#13759)

2025-12-19 17:04:24 -04:00

tvm_gemm.py

move opt under codegen (#11569)

2025-08-07 14:19:17 -07:00