metal: TRI, FILL, EXPM1, SOFTPLUS (#16623)

* feat(wip): Port initial TRI impl from pervious work

The kernel does not work and is not optimized, but the
code compiles and runs, so this will be the starting point
now that the core op has been merged.

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* fix: Remove argument for constant val override

This was added in the original draft, but later removed. With this, the
kernel now passes tests.

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* feat: Move the ttype conditional to templating to avoid conditional in kernel

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* fix: Type fixes

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* feat: Add softplus for metal

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* feat: Add EXPM1 for metal

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* feat: Add FILL for metal

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* refactor: Branchless version of tri using _ggml_vec_tri_cmp as a mask

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* fix: Remove unused arguments

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

* refactor: Use select instead of branch for softplus non-vec

Branch: ggml-cumsum-tri

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>

---------

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Gabe Goodhart
2025-12-04 19:12:19 +02:00
committed by GitHub
co-authored by Georgi Gerganov
parent 9d0229967a
commit bde188d60f
7 changed files with 265 additions and 0 deletions
+5
View File
@@ -818,6 +818,8 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
case GGML_UNARY_OP_HARDSWISH:
case GGML_UNARY_OP_HARDSIGMOID:
case GGML_UNARY_OP_EXP:
case GGML_UNARY_OP_SOFTPLUS:
case GGML_UNARY_OP_EXPM1:
return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
default:
return false;
@@ -850,6 +852,7 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
case GGML_OP_ACC:
case GGML_OP_REPEAT:
case GGML_OP_SCALE:
case GGML_OP_FILL:
case GGML_OP_CONV_TRANSPOSE_1D:
return true;
case GGML_OP_CONV_TRANSPOSE_2D:
@@ -867,6 +870,8 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
case GGML_OP_SUM:
return has_simdgroup_reduction && ggml_is_contiguous(op->src[0]);
case GGML_OP_TRI:
return ggml_is_contiguous_rows(op->src[0]);
case GGML_OP_SUM_ROWS:
case GGML_OP_CUMSUM:
case GGML_OP_MEAN: