Commit 8bed7cfb9 for llama.cpp
commit 8bed7cfb999cebc03d2cc5b78911026a8d57549a
Author: shaofeiqi <shaoqi@qti.qualcomm.com>
Date: Wed Nov 4 19:18:24 2026 -0800
opencl: support non-contig sigmoid (#31687)
Co-authored-by: Li He <lih@qti.qualcomm.com>
diff --git a/ggml/src/ggml-opencl/ggml-opencl.cpp b/ggml/src/ggml-opencl/ggml-opencl.cpp
index 8aa013edb..e5b382c7a 100644
--- a/ggml/src/ggml-opencl/ggml-opencl.cpp
+++ b/ggml/src/ggml-opencl/ggml-opencl.cpp
@@ -853,6 +853,7 @@ struct ggml_backend_opencl_context {
cl_kernel kernel_gelu_quick, kernel_gelu_quick_4;
cl_kernel kernel_relu;
cl_kernel kernel_sigmoid_f32, kernel_sigmoid_f16;
+ cl_kernel kernel_sigmoid_f32_nc, kernel_sigmoid_f16_nc;
cl_kernel kernel_tri;
cl_kernel kernel_fill;
cl_kernel kernel_clamp, kernel_clamp_nc;
@@ -4310,6 +4311,8 @@ static void load_cl_kernels(ggml_backend_opencl_context *backend_ctx) {
CL_CHECK((backend_ctx->kernel_sigmoid_f32 = clCreateKernel(backend_ctx->program_sigmoid, "kernel_sigmoid_f32", &err), err));
CL_CHECK((backend_ctx->kernel_sigmoid_f16 = clCreateKernel(backend_ctx->program_sigmoid, "kernel_sigmoid_f16", &err), err));
+ CL_CHECK((backend_ctx->kernel_sigmoid_f32_nc = clCreateKernel(backend_ctx->program_sigmoid, "kernel_sigmoid_f32_nc", &err), err));
+ CL_CHECK((backend_ctx->kernel_sigmoid_f16_nc = clCreateKernel(backend_ctx->program_sigmoid, "kernel_sigmoid_f16_nc", &err), err));
GGML_LOG_CONT(".");
}
@@ -11759,7 +11762,8 @@ static bool ggml_opencl_supports_op(ggml_backend_dev_t dev, const struct ggml_te
case GGML_UNARY_OP_GELU_QUICK:
return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32;
case GGML_UNARY_OP_SIGMOID:
- return ggml_is_contiguous(op->src[0]);
+ return op->src[0]->type == op->type &&
+ (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_BF16);
case GGML_UNARY_OP_TANH:
case GGML_UNARY_OP_NEG:
case GGML_UNARY_OP_EXP:
@@ -19464,6 +19468,33 @@ static void ggml_cl_sigmoid(ggml_backend_t backend, const ggml_tensor * src0, co
cl_ulong offset0 = extra0->offset + src0->view_offs;
cl_ulong offsetd = extrad->offset + dst->view_offs;
+ if (!ggml_is_contiguous(src0) || !ggml_is_contiguous(dst)) {
+ GGML_TENSOR_LOCALS(int, ne0, src0, ne);
+ GGML_TENSOR_LOCALS(cl_ulong, nb0, src0, nb);
+ GGML_TENSOR_LOCALS(cl_ulong, nb, dst, nb);
+
+ cl_kernel kernel = src0->type == GGML_TYPE_F32 ? backend_ctx->kernel_sigmoid_f32_nc : backend_ctx->kernel_sigmoid_f16_nc;
+
+ CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device));
+ CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0));
+ CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device));
+ CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd));
+ CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne00));
+ CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &nb00));
+ CL_CHECK(clSetKernelArg(kernel, 6, sizeof(cl_ulong), &nb01));
+ CL_CHECK(clSetKernelArg(kernel, 7, sizeof(cl_ulong), &nb02));
+ CL_CHECK(clSetKernelArg(kernel, 8, sizeof(cl_ulong), &nb03));
+ CL_CHECK(clSetKernelArg(kernel, 9, sizeof(cl_ulong), &nb0));
+ CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb1));
+ CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb2));
+ CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb3));
+
+ size_t global_work_size[] = {(size_t)ne01 * 64, (size_t)ne02, (size_t)ne03};
+ size_t local_work_size[] = {64, 1, 1};
+ backend_ctx->enqueue_ndrange_kernel(kernel, 3, global_work_size, local_work_size, dst);
+ return;
+ }
+
cl_kernel kernel;
if (src0->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) {
kernel = backend_ctx->kernel_sigmoid_f32;
diff --git a/ggml/src/ggml-opencl/kernels/sigmoid.cl b/ggml/src/ggml-opencl/kernels/sigmoid.cl
index e3f669dde..cfe9f7295 100644
--- a/ggml/src/ggml-opencl/kernels/sigmoid.cl
+++ b/ggml/src/ggml-opencl/kernels/sigmoid.cl
@@ -27,3 +27,63 @@ kernel void kernel_sigmoid_f16(
dst[get_global_id(0)] = 1.0f / (1.0f + exp(-src0[get_global_id(0)]));
}
+
+kernel void kernel_sigmoid_f32_nc(
+ global const char * src0,
+ ulong offset0,
+ global char * dst,
+ ulong offsetd,
+ int ne00,
+ ulong nb00,
+ ulong nb01,
+ ulong nb02,
+ ulong nb03,
+ ulong nb0,
+ ulong nb1,
+ ulong nb2,
+ ulong nb3
+) {
+ src0 = src0 + offset0;
+ dst = dst + offsetd;
+
+ const int i3 = get_group_id(2);
+ const int i2 = get_group_id(1);
+ const int i1 = get_group_id(0);
+
+ for (int i0 = get_local_id(0); i0 < ne00; i0 += get_local_size(0)) {
+ global const float * x = (global const float *)(src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
+ global float * y = (global float *)(dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
+
+ *y = 1.0f / (1.0f + exp(-*x));
+ }
+}
+
+kernel void kernel_sigmoid_f16_nc(
+ global const char * src0,
+ ulong offset0,
+ global char * dst,
+ ulong offsetd,
+ int ne00,
+ ulong nb00,
+ ulong nb01,
+ ulong nb02,
+ ulong nb03,
+ ulong nb0,
+ ulong nb1,
+ ulong nb2,
+ ulong nb3
+) {
+ src0 = src0 + offset0;
+ dst = dst + offsetd;
+
+ const int i3 = get_group_id(2);
+ const int i2 = get_group_id(1);
+ const int i1 = get_group_id(0);
+
+ for (int i0 = get_local_id(0); i0 < ne00; i0 += get_local_size(0)) {
+ global const half * x = (global const half *)(src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
+ global half * y = (global half *)(dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
+
+ *y = 1.0f / (1.0f + exp(-(float)*x));
+ }
+}