Skip to content

Fix support for ROCm 7 #11

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Anbeeld merged 1 commit into from
May 13, 2026
+17 −14
Merged

Fix support for ROCm 7 #11

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 9 additions & 9 deletions ggml/src/ggml-cuda/cross-ring-interleave.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -178,7 +178,7 @@ extern "C" bool dflash_cross_ring_gpu_write_d2d(
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (attr.type != cudaMemoryTypeDevice || attr.device != ring->device) {
return false;
}
Expand Down Expand Up @@ -253,7 +253,7 @@ extern "C" bool dflash_rebuild_conv_state(
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (r_attr.type != cudaMemoryTypeDevice || qkv_attr.type != cudaMemoryTypeDevice ||
r_attr.device != qkv_attr.device) {
return false;
Expand Down Expand Up @@ -288,7 +288,7 @@ extern "C" bool dflash_cuda_copy_d2d(void * dst, const void * src, size_t size)
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (dst_attr.type != cudaMemoryTypeDevice || src_attr.type != cudaMemoryTypeDevice ||
dst_attr.device != src_attr.device) {
return false;
Expand All @@ -315,7 +315,7 @@ extern "C" bool dflash_cuda_prepare_ptr(const void * ptr) {
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (attr.type != cudaMemoryTypeDevice) {
return false;
}
Expand Down Expand Up @@ -345,7 +345,7 @@ extern "C" bool dflash_cuda_synchronize_ptr(const void * ptr) {
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (attr.type != cudaMemoryTypeDevice) {
return false;
}
Expand Down Expand Up @@ -418,7 +418,7 @@ extern "C" void dflash_cross_ring_gpu_set_tensor(
cudaGetLastError();
}

#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
const bool dst_is_device = dst_err == cudaSuccess && dst_attr.type == cudaMemoryTypeDevice;
const bool src_is_device = src_err == cudaSuccess && src_attr.type == cudaMemoryTypeDevice;
#else
Expand Down Expand Up @@ -456,7 +456,7 @@ extern "C" bool dflash_kv_cache_write_d2d(
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (ring_attr.type != cudaMemoryTypeDevice || src_attr.type != cudaMemoryTypeDevice ||
ring_attr.device != src_attr.device) {
return false;
Expand Down Expand Up @@ -528,7 +528,7 @@ extern "C" bool dflash_kv_cache_append_d2d(
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (cache_attr.type != cudaMemoryTypeDevice || src_attr.type != cudaMemoryTypeDevice ||
cache_attr.device != src_attr.device) {
return false;
Expand Down Expand Up @@ -649,7 +649,7 @@ extern "C" bool dflash_kv_cache_interleave(
cudaGetLastError();
return false;
}
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
if (ring_attr.type != cudaMemoryTypeDevice || stage_attr.type != cudaMemoryTypeDevice ||
ring_attr.device != stage_attr.device) {
return false;
Expand Down
2 changes: 1 addition & 1 deletion ggml/src/ggml-cuda/ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3147,7 +3147,7 @@ static void ggml_cuda_log_nonlocal_src_buffer(
cudaError_t err = cudaPointerGetAttributes(&attr, data);
if (err == cudaSuccess) {
ptr_device = attr.device;
#if CUDART_VERSION >= 10000
#if CUDART_VERSION >= 10000 || defined(GGML_USE_HIP)
switch (attr.type) {
#else
switch (attr.memoryType) {
Expand Down
5 changes: 5 additions & 0 deletions ggml/src/ggml-cuda/vendors/hip.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -99,6 +99,11 @@
#define cudaMemGetInfo hipMemGetInfo
#define cudaOccupancyMaxPotentialBlockSize hipOccupancyMaxPotentialBlockSize
#define cudaSetDevice hipSetDevice
#define cudaPointerAttributes hipPointerAttribute_t
#define cudaPointerGetAttributes hipPointerGetAttributes
#define cudaMemoryTypeDevice hipMemoryTypeDevice
#define cudaMemoryTypeHost hipMemoryTypeHost
#define cudaMemoryTypeManaged hipMemoryTypeManaged
#define cuDeviceGet hipDeviceGet
#define CUdevice hipDevice_t
#define CUdeviceptr hipDeviceptr_t
Expand Down
6 changes: 2 additions & 4 deletions ggml/src/ggml-hip/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,10 +80,8 @@ else()
../ggml-cuda/template-instances/fattn-vec-instance-q4_0-q4_0.cu
../ggml-cuda/template-instances/fattn-vec-instance-q8_0-q8_0.cu
../ggml-cuda/template-instances/fattn-vec-instance-bf16-bf16.cu)
if (WIN32)
file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec-instance-*turbo*.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
endif()
file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec-instance-*turbo*.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
endif()

ggml_add_backend_library(ggml-hip
Expand Down