fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-05 00:58:05 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

etnaviv/ml: Specify which of the input tensors need transposing.

Browse source 
In preparation for operations that have more than one input that may
need transposing, such as Add.

Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32498>
This commit is contained in:
Tomeu Vizoso 2024-10-24 09:50:10 +02:00 • committed by Marge Bot
parent d47deba142
commit 10bd5c23c6
3 changed files with 16 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/gallium/drivers/etnaviv/etnaviv_ml.c
View file

@ -268,7 +268,7 @@ lower_operations(struct etna_ml_subgraph *subgraph,
if (needs_transpose(poperations, count, poperation)) {
ML_DBG("Adding transpose for convolution operation.\n");
struct etna_operation *operation = calloc(1, sizeof(*operation));
etna_ml_lower_transpose(subgraph, poperation, operation, &input_tensor);
etna_ml_lower_transpose(subgraph, poperation->input_tensors[0], operation, &input_tensor);
list_addtail(&operation->link, etna_operations);
}

27
src/gallium/drivers/etnaviv/etnaviv_ml_tp.c
View file

@ -536,7 +536,7 @@ create_reshuffle_config(struct etna_ml_subgraph *subgraph, const struct etna_ope
}
static inline uint8_t
etna_tensor_zero_point(struct pipe_tensor *tensor)
etna_tensor_zero_point(const struct pipe_tensor *tensor)
{
if (tensor->is_signed) {
assert(tensor->zero_point >= -128 && tensor->zero_point <= 127);
@ -549,30 +549,31 @@ etna_tensor_zero_point(struct pipe_tensor *tensor)
void
etna_ml_lower_transpose(struct etna_ml_subgraph *subgraph,
const struct pipe_ml_operation *first_operation,
const struct pipe_tensor *input_tensor,
struct etna_operation *operation,
unsigned *output_tensor)
{
operation->type = ETNA_JOB_TYPE_TP;
operation->tp_type = ETNA_ML_TP_TRANSPOSE;
operation->input_tensors[0] = first_operation->input_tensors[0]->index;
operation->input_width = first_operation->input_tensors[0]->dims[1];
operation->input_height = first_operation->input_tensors[0]->dims[2];
operation->input_channels = first_operation->input_tensors[0]->dims[3];
operation->input_zero_point = etna_tensor_zero_point(first_operation->input_tensors[0]);
operation->input_scale = first_operation->input_tensors[0]->scale;
operation->input_tensors[0] = input_tensor->index;
operation->input_count = 1;
operation->input_width = input_tensor->dims[1];
operation->input_height = input_tensor->dims[2];
operation->input_channels = input_tensor->dims[3];
operation->input_zero_point = etna_tensor_zero_point(input_tensor);
operation->input_scale = input_tensor->scale;
operation->input_tensor_size = operation->input_width *
operation->input_height *
operation->input_channels;
*output_tensor = etna_ml_allocate_tensor(subgraph);
operation->output_tensor = *output_tensor;
operation->output_width = first_operation->input_tensors[0]->dims[1];
operation->output_height = first_operation->input_tensors[0]->dims[2];
operation->output_channels = first_operation->input_tensors[0]->dims[3];
operation->output_zero_point = etna_tensor_zero_point(first_operation->input_tensors[0]);
operation->output_scale = first_operation->input_tensors[0]->scale;
operation->output_width = operation->input_width;
operation->output_height = operation->input_height;
operation->output_channels = operation->input_channels;
operation->output_zero_point = operation->input_zero_point;
operation->output_scale = operation->input_scale;
}
void

2
src/gallium/drivers/etnaviv/etnaviv_ml_tp.h
View file

@ -7,7 +7,7 @@
void
etna_ml_lower_transpose(struct etna_ml_subgraph *subgraph,
const struct pipe_ml_operation *first_operation,
const struct pipe_tensor *input_tensor,
struct etna_operation *operation,
unsigned *output_tensor);