mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-22 14:53:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

GGML update to ec98e2002 (#13451)

Browse Source 
* Revert "add support for NVIDIA Nemotron 3 Nano"

This reverts commit e7d2ae9d69421012e9a8765c06a3fdf0e45b12f3.

* GGML update to 380b4c984

Remove MaskBatchPadding as GGML_KQ_MASK_PAD is no longer present (no
padding required)

* update to c45f89d55

* ec98e2002

solar pro needed more adjusting - needs verification

* review comments
This commit is contained in:
Daniel Hiltgen
2025-12-17 13:13:55 -08:00
committed by GitHub
parent 1c094038bc
commit 49a9c9ba6a
127 changed files with 8128 additions and 6710 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
llama/patches/0022-ggml-Enable-resetting-backend-devices.patch
View File

@@ -16,7 +16,7 @@ unused then it can be reset to free these data structures.
6 files changed, 32 insertions(+), 2 deletions(-)
diff --git a/ggml/include/ggml-backend.h b/ggml/include/ggml-backend.h
index b3b5b356a..69223c488 100644
index dbbb61d9c..92ca32a4b 100644
--- a/ggml/include/ggml-backend.h
+++ b/ggml/include/ggml-backend.h
@@ -178,6 +178,7 @@ extern "C" {
@@ -43,10 +43,10 @@ index 7bdf9d81f..21b35ac5c 100644
struct ggml_backend_device {
diff --git a/ggml/src/ggml-backend.cpp b/ggml/src/ggml-backend.cpp
index 4092dfe8a..a1a19fe51 100644
index 7746e8b92..189e97170 100644
--- a/ggml/src/ggml-backend.cpp
+++ b/ggml/src/ggml-backend.cpp
@@ -526,6 +526,14 @@ ggml_backend_t ggml_backend_dev_init(ggml_backend_dev_t device, const char * par
@@ -532,6 +532,14 @@ ggml_backend_t ggml_backend_dev_init(ggml_backend_dev_t device, const char * par
return device->iface.init_backend(device, params);
}
@@ -62,7 +62,7 @@ index 4092dfe8a..a1a19fe51 100644
GGML_ASSERT(device);
return device->iface.get_buffer_type(device);
diff --git a/ggml/src/ggml-cuda/ggml-cuda.cu b/ggml/src/ggml-cuda/ggml-cuda.cu
index ede1d089a..ec63cadab 100644
index eeaae3fe4..6852d2e20 100644
--- a/ggml/src/ggml-cuda/ggml-cuda.cu
+++ b/ggml/src/ggml-cuda/ggml-cuda.cu
@@ -113,6 +113,11 @@ int ggml_cuda_get_device() {
@@ -89,7 +89,7 @@ index ede1d089a..ec63cadab 100644
bool host_buffer = getenv("GGML_CUDA_NO_PINNED") == nullptr;
#ifdef GGML_CUDA_NO_PEER_COPY
@@ -4907,6 +4915,11 @@ static void ggml_backend_cuda_device_event_synchronize(ggml_backend_dev_t dev, g
@@ -4908,6 +4916,11 @@ static void ggml_backend_cuda_device_event_synchronize(ggml_backend_dev_t dev, g
CUDA_CHECK(cudaEventSynchronize((cudaEvent_t)event->context));
}
@@ -101,7 +101,7 @@ index ede1d089a..ec63cadab 100644
static const ggml_backend_device_i ggml_backend_cuda_device_interface = {
/* .get_name = */ ggml_backend_cuda_device_get_name,
/* .get_description = */ ggml_backend_cuda_device_get_description,
@@ -4923,6 +4936,7 @@ static const ggml_backend_device_i ggml_backend_cuda_device_interface = {
@@ -4924,6 +4937,7 @@ static const ggml_backend_device_i ggml_backend_cuda_device_interface = {
/* .event_new = */ ggml_backend_cuda_device_event_new,
/* .event_free = */ ggml_backend_cuda_device_event_free,
/* .event_synchronize = */ ggml_backend_cuda_device_event_synchronize,
@@ -110,10 +110,10 @@ index ede1d089a..ec63cadab 100644
// backend reg
diff --git a/ggml/src/ggml-cuda/vendors/hip.h b/ggml/src/ggml-cuda/vendors/hip.h
index b7d6edf7f..b987d7aeb 100644
index 951a88d56..4e162258d 100644
--- a/ggml/src/ggml-cuda/vendors/hip.h
+++ b/ggml/src/ggml-cuda/vendors/hip.h
@@ -45,6 +45,7 @@
@@ -49,6 +49,7 @@
#define cudaDeviceDisablePeerAccess hipDeviceDisablePeerAccess
#define cudaDeviceEnablePeerAccess hipDeviceEnablePeerAccess
#define cudaDeviceProp hipDeviceProp_t
@@ -122,10 +122,10 @@ index b7d6edf7f..b987d7aeb 100644
#define cudaError_t hipError_t
#define cudaErrorPeerAccessAlreadyEnabled hipErrorPeerAccessAlreadyEnabled
diff --git a/src/llama.cpp b/src/llama.cpp
index ab2e9868a..74c49e651 100644
index f69964b6d..759152b76 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -270,10 +270,12 @@ static struct llama_model * llama_model_load_from_file_impl(
@@ -921,10 +921,12 @@ static struct llama_model * llama_model_load_from_file_impl(
for (auto * dev : model->devices) {
ggml_backend_dev_props props;
ggml_backend_dev_get_props(dev, &props);