thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

Updated to enable for opmult to work

Browse source
This commit is contained in:
Alejandro Saucedo 2021-02-26 18:58:19 +00:00
parent fb617d1722
commit 3304767f2c
11 changed files with 49 additions and 39 deletions
Download patch file Download diff file Expand all files Collapse all files

9
test/TestAsyncOperations.cpp
View file

@ -80,8 +80,11 @@ TEST(TestAsyncOperations, TestManagerParallelExecution)
std::vector<std::shared_ptr<kp::Tensor>> inputsAsyncB;
std::vector<std::shared_ptr<kp::Algorithm>> algosAsync;
for (uint32_t i = 0; i < numParallel; i++) {
inputsAsyncB.push_back(mgr.tensor(data));
algosAsync.push_back(mgr.algorithm({inputsAsyncB[i]}, spirv));
}
std::vector<std::shared_ptr<kp::Sequence>> sqs;
@ -93,7 +96,7 @@ TEST(TestAsyncOperations, TestManagerParallelExecution)
auto startAsync = std::chrono::high_resolution_clock::now();
for (uint32_t i = 0; i < numParallel; i++) {
sqs[i]->evalAsync<kp::OpAlgoDispatch>(algorithms[i]);
sqs[i]->evalAsync<kp::OpAlgoDispatch>(algosAsync[i]);
}
for (uint32_t i = 0; i < numParallel; i++) {
@ -157,8 +160,8 @@ TEST(TestAsyncOperations, TestManagerAsyncExecution)
sq1->eval<kp::OpTensorSyncLocal>({ tensorA, tensorB });
std::shared_ptr<kp::Algorithm> algo1 = mgr.algorithm({tensorA});
std::shared_ptr<kp::Algorithm> algo2 = mgr.algorithm({tensorB});
std::shared_ptr<kp::Algorithm> algo1 = mgr.algorithm({tensorA}, spirv);
std::shared_ptr<kp::Algorithm> algo2 = mgr.algorithm({tensorB}, spirv);
sq1->evalAsync<kp::OpAlgoDispatch>(algo1);
sq2->evalAsync<kp::OpAlgoDispatch>(algo2);

6
test/TestDestroy.cpp
View file

@ -74,7 +74,7 @@ TEST(TestDestroy, TestDestroyTensorVector)
mgr.sequence()
->record<kp::OpTensorSyncDevice>(algo->getTensors())
->record<kp::OpAlgoDispatch>(algo)
->record<kp::OpTensorSyncDevice>(algo->getTensors())
->record<kp::OpTensorSyncLocal>(algo->getTensors())
->eval();
tensorA->destroy();
@ -111,12 +111,14 @@ TEST(TestDestroy, TestDestroySequenceSingle)
tensorA = mgr.tensor({0, 0, 0});
mgr.sequence()
sq = mgr.sequence()
->record<kp::OpTensorSyncDevice>({tensorA})
->record<kp::OpAlgoDispatch>(mgr.algorithm({tensorA}, spirv))
->record<kp::OpTensorSyncLocal>({tensorA})
->eval();
sq->destroy();
EXPECT_FALSE(sq->isInit());
}
}

12
test/TestLogisticRegression.cpp
View file

@ -32,13 +32,15 @@ TEST(TestLogisticRegression, TestMainLogisticRegression)
wIn, wOutI, wOutJ,
bIn, bOut, lOut };
mgr.sequence()->eval<kp::OpTensorSyncDevice>(params);
std::vector<uint32_t> spirv = std::vector<uint32_t>(
(uint32_t*)kp::shader_data::shaders_glsl_logisticregression_comp_spv,
(uint32_t*)(kp::shader_data::shaders_glsl_logisticregression_comp_spv +
kp::shader_data::shaders_glsl_logisticregression_comp_spv_len));
(uint32_t*)kp::shader_data::test_shaders_glsl_test_logistic_regression_comp_spv,
(uint32_t*)(kp::shader_data::test_shaders_glsl_test_logistic_regression_comp_spv +
kp::shader_data::test_shaders_glsl_test_logistic_regression_comp_spv_len));
std::shared_ptr<kp::Algorithm> algorithm =
mgr.algorithm(params, spirv, kp::Workgroup(), kp::Constants({5.0}));
mgr.algorithm(params, spirv, kp::Workgroup({5}), kp::Constants({5.0}));
std::shared_ptr<kp::Sequence> sq =
mgr.sequence()
@ -104,6 +106,8 @@ TEST(TestLogisticRegression, TestMainLogisticRegressionManualCopy)
wIn, wOutI, wOutJ,
bIn, bOut, lOut };
mgr.sequence()->record<kp::OpTensorSyncDevice>(params)->eval();
std::vector<uint32_t> spirv = std::vector<uint32_t>(
(uint32_t*)kp::shader_data::shaders_glsl_logisticregression_comp_spv,
(uint32_t*)(kp::shader_data::shaders_glsl_logisticregression_comp_spv +