Skip to content
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.

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

Online compiler tests failing with new GPU driver #16406

Open
sarnex opened this issue Dec 17, 2024 · 2 comments
Open

Online compiler tests failing with new GPU driver #16406

sarnex opened this issue Dec 17, 2024 · 2 comments
Labels
bug Something isn't working

Comments

@sarnex
Copy link
Contributor

sarnex commented Dec 17, 2024

Describe the bug

https://github.com/intel/llvm/actions/runs/12381671746/job/34561005375?pr=16312

Based on the error it seems we need to change something to work with the new driver API.

024-12-17T21:52:13.5615438Z # RUN: at line 4
2024-12-17T21:52:13.5615705Z env ONEAPI_DEVICE_SELECTOR=level_zero:gpu  /__w/llvm/llvm/build-e2e/OnlineCompiler/Output/online_compiler_L0.cpp.tmp.out
2024-12-17T21:52:13.5616168Z # executed command: env ONEAPI_DEVICE_SELECTOR=level_zero:gpu /__w/llvm/llvm/build-e2e/OnlineCompiler/Output/online_compiler_L0.cpp.tmp.out
2024-12-17T21:52:13.5616480Z # .---command stdout------------
2024-12-17T21:52:13.5616625Z # | Test case1
2024-12-17T21:52:13.5616741Z # | IL size = 1024
2024-12-17T21:52:13.5616857Z # | Run the kernel now:
2024-12-17T21:52:13.5616979Z # | 0*2 + 100 = 100
2024-12-17T21:52:13.5617092Z # | 1*2 + 100 = 102
2024-12-17T21:52:13.5617199Z # | 2*2 + 100 = 104
2024-12-17T21:52:13.5617309Z # | 3*2 + 100 = 106
2024-12-17T21:52:13.5617425Z # | Test case2
2024-12-17T21:52:13.5617533Z # | IL size = 1024
2024-12-17T21:52:13.5617655Z # | Run the kernel now:
2024-12-17T21:52:13.5617773Z # | 0*2 + 100 = 100
2024-12-17T21:52:13.5617884Z # | 1*2 + 100 = 102
2024-12-17T21:52:13.5617995Z # | 2*2 + 100 = 104
2024-12-17T21:52:13.5618098Z # | 3*2 + 100 = 106
2024-12-17T21:52:13.5618211Z # | Test case3
2024-12-17T21:52:13.5618381Z # | Compilation to IL failed: ocloc reported compilation errors: {
2024-12-17T21:52:13.5618644Z # | AdaptorCM: incompatible clangFEWrapper interface: expected = 11, loaded = 10
2024-12-17T21:52:13.5618879Z # | Build failed with error code: -11
2024-12-17T21:52:13.5619098Z # | Command was: ocloc -q -spv_only -device tgllp -64 -options "-cmc" -file main.cl
2024-12-17T21:52:13.5619299Z # | 
2024-12-17T21:52:13.5619403Z # | }
2024-12-17T21:52:13.5619511Z # `-----------------------------
2024-12-17T21:52:13.5619650Z # .---command stderr------------
2024-12-17T21:52:13.5619904Z # | ZE_LOADER_DEBUG_TRACE:zeInitDrivers called first, but not supported by driver, returning uninitialized.
2024-12-17T21:52:13.5620161Z # `-----------------------------
2024-12-17T21:52:13.5620305Z # error: command failed with exit status: 1





2024-12-17T21:52:13.5698802Z env ONEAPI_DEVICE_SELECTOR=opencl:gpu  /__w/llvm/llvm/build-e2e/OnlineCompiler/Output/online_compiler_OpenCL.cpp.tmp.out
2024-12-17T21:52:13.5699332Z # executed command: env ONEAPI_DEVICE_SELECTOR=opencl:gpu /__w/llvm/llvm/build-e2e/OnlineCompiler/Output/online_compiler_OpenCL.cpp.tmp.out
2024-12-17T21:52:13.5699672Z # .---command stdout------------
2024-12-17T21:52:13.5699815Z # | Test case1
2024-12-17T21:52:13.5699936Z # | IL size = 1024
2024-12-17T21:52:13.5700059Z # | Run the kernel now:
2024-12-17T21:52:13.5700186Z # | 0*2 + 100 = 100
2024-12-17T21:52:13.5700303Z # | 1*2 + 100 = 102
2024-12-17T21:52:13.5700412Z # | 2*2 + 100 = 104
2024-12-17T21:52:13.5700511Z # | 3*2 + 100 = 106
2024-12-17T21:52:13.5700623Z # | Test case2
2024-12-17T21:52:13.5700733Z # | IL size = 1024
2024-12-17T21:52:13.5700855Z # | Run the kernel now:
2024-12-17T21:52:13.5700974Z # | 0*2 + 100 = 100
2024-12-17T21:52:13.5701078Z # | 1*2 + 100 = 102
2024-12-17T21:52:13.5701185Z # | 2*2 + 100 = 104
2024-12-17T21:52:13.5701292Z # | 3*2 + 100 = 106
2024-12-17T21:52:13.5701396Z # | Test case3
2024-12-17T21:52:13.5701569Z # | Compilation to IL failed: ocloc reported compilation errors: {
2024-12-17T21:52:13.5701838Z # | AdaptorCM: incompatible clangFEWrapper interface: expected = 11, loaded = 10
2024-12-17T21:52:13.5702064Z # | Build failed with error code: -11
2024-12-17T21:52:13.5702286Z # | Command was: ocloc -q -spv_only -device tgllp -64 -options "-cmc" -file main.cl
2024-12-17T21:52:13.5702491Z # | 
2024-12-17T21:52:13.5702589Z # | }
2024-12-17T21:52:13.5702696Z # `-----------------------------
2024-12-17T21:52:13.5702847Z # error: command failed with exit status: 1
2024-12-17T21:52:13.5702955Z 

To reproduce

No response

Environment

No response

Additional context

No response

@sarnex sarnex added the bug Something isn't working label Dec 17, 2024
@sarnex
Copy link
Contributor Author

sarnex commented Dec 23, 2024

BTW this test fails even when run on OCL CPU/FPGA, it seems it manually calls the GPU driver in every mode

FYI @cperkinsintel

@uditagarwal97
Copy link
Contributor

Online Compiler tests passed on PVC + dev igc: https://github.com/intel/llvm/actions/runs/12471647091/job/34809310703?pr=15308#step:24:2993

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

2 participants