intel/llvm
1
0
Fork 0
mirror of https://github.com/intel/llvm.git synced 2026-01-20 01:58:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Libomptarget][NFC] Clang format the AMDGPU plugin

Browse Source 
Summary:
A previous patch did not format the plugin again after making changes.
Ensure that libomptarget stays formatted.
This commit is contained in:
Joseph Huber
2022-08-03 15:17:30 -04:00
parent f0bb276794
commit b3335e8ed7
1 changed files with 13 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
openmp/libomptarget/plugins/amdgpu/src/rtl.cpp
View File

@@ -1131,7 +1131,7 @@ int32_t dataRetrieve(int32_t DeviceId, void *HstPtr, void *TgtPtr, int64_t Size,
(long long unsigned)(Elf64_Addr)HstPtr);
Err = DeviceInfo().freesignalpoolMemcpyD2H(HstPtr, TgtPtr, (size_t)Size,
DeviceId);
DeviceId);
if (Err != HSA_STATUS_SUCCESS) {
DP("Error when copying data from device to host. Pointers: "
@@ -1495,8 +1495,8 @@ int32_t runRegionLocked(int32_t DeviceId, void *TgtEntryPtr, void **TgtArgs,
// under a multiple reader lock, not a writer lock.
static pthread_mutex_t HostcallInitLock = PTHREAD_MUTEX_INITIALIZER;
pthread_mutex_lock(&HostcallInitLock);
uint64_t Buffer = hostrpc_assign_buffer(DeviceInfo().HSAAgents[DeviceId],
Queue, DeviceId);
uint64_t Buffer = hostrpc_assign_buffer(
DeviceInfo().HSAAgents[DeviceId], Queue, DeviceId);
pthread_mutex_unlock(&HostcallInitLock);
if (!Buffer) {
DP("hostrpc_assign_buffer failed, gpu would dereference null and "
@@ -1858,7 +1858,7 @@ struct DeviceEnvironment {
}
return DeviceInfo().freesignalpoolMemcpyH2D(StatePtr, &HostDeviceEnv,
StatePtrSize, DeviceId);
StatePtrSize, DeviceId);
}
}
return HSA_STATUS_SUCCESS;
@@ -2114,7 +2114,8 @@ int32_t __tgt_rtl_init_device(int DeviceId) {
if (print_kernel_trace & STARTUP_DETAILS)
DP("Device#%-2d CU's: %2d %s\n", DeviceId,
DeviceInfo().ComputeUnits[DeviceId], DeviceInfo().GPUName[DeviceId].c_str());
DeviceInfo().ComputeUnits[DeviceId],
DeviceInfo().GPUName[DeviceId].c_str());
// Query attributes to determine number of threads/block and blocks/grid.
uint16_t WorkgroupMaxDim[3];
@@ -2126,7 +2127,8 @@ int32_t __tgt_rtl_init_device(int DeviceId) {
RTLDeviceInfoTy::DefaultNumTeams);
} else if (WorkgroupMaxDim[0] <= RTLDeviceInfoTy::HardTeamLimit) {
DeviceInfo().GroupsPerDevice[DeviceId] = WorkgroupMaxDim[0];
DP("Using %d ROCm blocks per grid\n", DeviceInfo().GroupsPerDevice[DeviceId]);
DP("Using %d ROCm blocks per grid\n",
DeviceInfo().GroupsPerDevice[DeviceId]);
} else {
DeviceInfo().GroupsPerDevice[DeviceId] = RTLDeviceInfoTy::HardTeamLimit;
DP("Max ROCm blocks per grid %d exceeds the hard team limit %d, capping "
@@ -2281,8 +2283,9 @@ __tgt_target_table *__tgt_rtl_load_binary_locked(int32_t DeviceId,
return NULL;
{
auto Env = DeviceEnvironment(DeviceId, DeviceInfo().NumberOfDevices,
DeviceInfo().Env.DynamicMemSize, Image, ImgSize);
auto Env =
DeviceEnvironment(DeviceId, DeviceInfo().NumberOfDevices,
DeviceInfo().Env.DynamicMemSize, Image, ImgSize);
auto &KernelInfo = DeviceInfo().KernelInfoTable[DeviceId];
auto &SymbolInfo = DeviceInfo().SymbolInfoTable[DeviceId];
@@ -2439,8 +2442,8 @@ __tgt_target_table *__tgt_rtl_load_binary_locked(int32_t DeviceId,
// If unified memory is present any target link variables
// can access host addresses directly. There is no longer a
// need for device copies.
Err = DeviceInfo().freesignalpoolMemcpyH2D(Varptr, E->addr, sizeof(void *),
DeviceId);
Err = DeviceInfo().freesignalpoolMemcpyH2D(Varptr, E->addr,
sizeof(void *), DeviceId);
if (Err != HSA_STATUS_SUCCESS)
DP("Error when copying USM\n");
DP("Copy linked variable host address (" DPxMOD ")"