CANN: add support for partial RoPE and Vision mode (#17543)

* cann: add support for partial RoPE and Vision mode

Add support for two important RoPE variants: partial rotation (rope_dims < ne0)
and Vision mode rotation.

1. Support for partial RoPE (rope_dims < ne0):
   - Split tensor into head (first rope_dims dimensions) and tail portions
   - Apply rotation only to head portion using RotaryPositionEmbedding operator
   - Copy unrotated tail portion directly from source to destination
   - Handle both contiguous and non-contiguous tensor layouts

2. Support for Vision mode (GGML_ROPE_TYPE_VISION):
   - Set rope_dims = ne0 for Vision mode to rotate entire tensor
   - Vision mode pairs dimension i with dimension i+n_dims (where n_dims = ne0/2)
   - No tail handling needed since entire tensor is rotated

Implementation details:
   - Use has_tail flag to determine execution path: head/tail splitting when
     rope_dims < ne0, or full tensor rotation when rope_dims == ne0
   - Support both F32 and F16 data types with intermediate F32 conversion
   - Copy non-contiguous tensors to contiguous buffers before calling
     RotaryPositionEmbedding operator for compatibility
   - Improve cache invalidation logic to include rope_dims and indep_sects
     parameters

These enhancements enable CANN backend to handle various RoPE configurations
used in modern vision-language models and models with partial rotation.

* cann: fix review comment
This commit is contained in:
Chenguang Li
2025-12-09 17:53:23 +08:00
committed by GitHub
parent 0cdce38a97
commit ca709e427b
3 changed files with 162 additions and 72 deletions
+6 -8
View File
@@ -2308,7 +2308,7 @@ static enum ggml_status ggml_backend_cann_graph_compute(ggml_backend_t backend,
bool cann_graph_update_required = false;
#ifdef USE_ACL_GRAPH
bool use_cann_graph = true;
bool use_cann_graph = true;
static bool prefill_use_graph = parse_bool(get_env("GGML_CANN_PREFILL_USE_GRAPH").value_or(""));
if (!prefill_use_graph) {
@@ -2338,7 +2338,7 @@ static enum ggml_status ggml_backend_cann_graph_compute(ggml_backend_t backend,
}
}
#else
bool use_cann_graph = false;
bool use_cann_graph = false;
#endif // USE_ACL_GRAPH
evaluate_and_capture_cann_graph(cann_ctx, cgraph, use_cann_graph, cann_graph_update_required);
@@ -2474,16 +2474,14 @@ static bool ggml_backend_cann_supports_op(ggml_backend_dev_t dev, const ggml_ten
}
case GGML_OP_ROPE:
{
// TODO: with ops-test v == 1
// TODO: n_dims <= ne0
if (op->src[0]->ne[0] != op->op_params[1]) {
return false;
}
if (op->src[0]->ne[0] > 896) {
return false;
}
#ifdef ASCEND_310P
// TODO: Support rope_dim < ne00(dim)
if (op->src[0]->ne[0] != op->op_params[1]) {
return false;
}
if (!ggml_is_contiguous(op->src[0])) {
return false;
}