IntenseWebs/tuxclocker
3
0
Fork 0
mirror of https://github.com/Lurkki14/tuxclocker.git synced 2024-11-21 15:57:25 -06:00
Code Issues Packages Projects Releases Wiki Activity

use clock offsets for all performance levels when available

Browse Source 
Fixes clock offset setting failing on some NVIDIA GPUs
This commit is contained in:
Jussi Kuokkanen 2023-10-31 16:07:16 +02:00
parent 72f68a8a02
commit 08016b35c8
1 changed files with 77 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

112
src/plugins/Nvidia.cpp
View File

@ -117,26 +117,37 @@ std::vector<TreeNode<DeviceNode>> getMemClockWrite(NvidiaGPUData data) {
if (!data.maxPerfState.has_value())
return {};
auto maxPerfState = data.maxPerfState.value();
// TODO: getting current offset and available range doesn't work properly through NVML
// but setting does
NVCTRLAttributeValidValuesRec values;
if (!XNVCTRLQueryValidTargetAttributeValues(data.dpy, NV_CTRL_TARGET_TYPE_GPU, data.index,
maxPerfState, NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET, &values))
return {};
// Transfer rate -> clock speed
Range<int> range{values.u.range.min / 2, values.u.range.max / 2};
auto getFunc = [=]() -> std::optional<AssignmentArgument> {
auto maxPerfState = data.maxPerfState.value();
auto getFunc = [=](int attribute) -> std::optional<AssignmentArgument> {
int value;
if (!XNVCTRLQueryTargetAttribute(data.dpy, NV_CTRL_TARGET_TYPE_GPU, data.index,
maxPerfState, NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET, &value))
maxPerfState, attribute, &value))
return std::nullopt;
return value / 2;
};
auto setFunc = [=](AssignmentArgument a) -> std::optional<AssignmentError> {
// Check which attribute works (same as nvidia-settings code)
std::optional<int> attribute;
if (getFunc(NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET).has_value())
attribute = NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET;
if (getFunc(NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET_ALL_PERFORMANCE_LEVELS).has_value())
attribute = NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET_ALL_PERFORMANCE_LEVELS;
if (!attribute.has_value())
return {};
NVCTRLAttributeValidValuesRec values;
if (!XNVCTRLQueryValidTargetAttributeValues(
data.dpy, NV_CTRL_TARGET_TYPE_GPU, data.index, maxPerfState, *attribute, &values)) {
return {};
}
// Transfer rate -> clock speed
Range<int> range{values.u.range.min / 2, values.u.range.max / 2};
// Seems the ..ALL_PERFORMANCE_LEVELS attribute doesn't work with the NVML function
auto setFuncNVML = [=](AssignmentArgument a) -> std::optional<AssignmentError> {
if (!std::holds_alternative<int>(a))
return AssignmentError::InvalidType;
auto target = std::get<int>(a);
@ -148,15 +159,34 @@ std::vector<TreeNode<DeviceNode>> getMemClockWrite(NvidiaGPUData data) {
return fromNVMLRet(ret);
};
Assignable a{setFunc, range, getFunc, _("MHz")};
auto setFuncXNVCTRL = [=](AssignmentArgument a) -> std::optional<AssignmentError> {
if (!std::holds_alternative<int>(a))
return AssignmentError::InvalidType;
auto target = std::get<int>(a);
if (target < range.min || target > range.max)
return AssignmentError::OutOfRange;
if (getFunc().has_value())
return {DeviceNode{
.name = _("Memory Clock Offset"),
.interface = a,
.hash = md5(data.uuid + "Memory Clock Offset"),
}};
return {};
// Clock speed -> transfer rate
auto value = target * 2;
if (!XNVCTRLSetTargetAttributeAndGetStatus(data.dpy, NV_CTRL_TARGET_TYPE_GPU,
data.index, maxPerfState, *attribute, value))
return AssignmentError::UnknownError;
return std::nullopt;
};
auto getFuncWithAttribute = [=]() -> std::optional<AssignmentArgument> {
return getFunc(*attribute);
};
auto a = (*attribute == NV_CTRL_GPU_MEM_TRANSFER_RATE_OFFSET)
? Assignable{setFuncNVML, range, getFuncWithAttribute, _("MHz")}
: Assignable{setFuncXNVCTRL, range, getFuncWithAttribute, _("MHz")};
return {DeviceNode{
.name = _("Memory Clock Offset"),
.interface = a,
.hash = md5(data.uuid + "Memory Clock Offset"),
}};
}
std::vector<TreeNode<DeviceNode>> getCoreClockWrite(NvidiaGPUData data) {
@ -164,25 +194,33 @@ std::vector<TreeNode<DeviceNode>> getCoreClockWrite(NvidiaGPUData data) {
return {};
auto maxPerfState = data.maxPerfState.value();
// TODO: NVML has functions for core clock as well but they're borked as of writing
NVCTRLAttributeValidValuesRec values;
if (!XNVCTRLQueryValidTargetAttributeValues(data.dpy, NV_CTRL_TARGET_TYPE_GPU, data.index,
maxPerfState, NV_CTRL_GPU_NVCLOCK_OFFSET, &values)) {
std::cout << "b" << maxPerfState << "\n";
return {};
}
// Transfer rate -> clock speed
Range<int> range{values.u.range.min, values.u.range.max};
auto getFunc = [=]() -> std::optional<AssignmentArgument> {
auto getFunc = [=](int attribute) -> std::optional<AssignmentArgument> {
int value;
if (!XNVCTRLQueryTargetAttribute(data.dpy, NV_CTRL_TARGET_TYPE_GPU, data.index,
maxPerfState, NV_CTRL_GPU_NVCLOCK_OFFSET, &value))
maxPerfState, attribute, &value))
return std::nullopt;
return value;
};
// Check which attribute works (same as nvidia-settings code)
std::optional<int> attribute;
if (getFunc(NV_CTRL_GPU_NVCLOCK_OFFSET).has_value())
attribute = NV_CTRL_GPU_NVCLOCK_OFFSET;
if (getFunc(NV_CTRL_GPU_NVCLOCK_OFFSET_ALL_PERFORMANCE_LEVELS).has_value())
attribute = NV_CTRL_GPU_NVCLOCK_OFFSET_ALL_PERFORMANCE_LEVELS;
if (!attribute.has_value())
return {};
// TODO: NVML has functions for core clock as well but they're borked as of writing
NVCTRLAttributeValidValuesRec values;
if (!XNVCTRLQueryValidTargetAttributeValues(
data.dpy, NV_CTRL_TARGET_TYPE_GPU, data.index, maxPerfState, *attribute, &values)) {
return {};
}
Range<int> range{values.u.range.min, values.u.range.max};
auto setFunc = [=](AssignmentArgument a) -> std::optional<AssignmentError> {
if (!std::holds_alternative<int>(a))
return AssignmentError::InvalidType;
@ -191,12 +229,16 @@ std::vector<TreeNode<DeviceNode>> getCoreClockWrite(NvidiaGPUData data) {
return AssignmentError::OutOfRange;
if (!XNVCTRLSetTargetAttributeAndGetStatus(data.dpy, NV_CTRL_TARGET_TYPE_GPU,
data.index, maxPerfState, NV_CTRL_GPU_NVCLOCK_OFFSET, target))
data.index, maxPerfState, *attribute, target))
return AssignmentError::UnknownError;
return std::nullopt;
};
Assignable a{setFunc, range, getFunc, _("MHz")};
auto getFuncWithAttribute = [=]() -> std::optional<AssignmentArgument> {
return getFunc(*attribute);
};
Assignable a{setFunc, range, getFuncWithAttribute, _("MHz")};
return {DeviceNode{
.name = _("Core Clock Offset"),