-
Notifications
You must be signed in to change notification settings - Fork 957
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
which case can report "No configurations found." #1879
Comments
@feixuedudiao can you verify your OpenCL driver version? adding @kealan-barbieri, @echeresh for input |
This is due to limited support for gen9 platforms in the optimized conv implementation. I submitted a small PR to provide a workaround for such cases. |
@yehudaorel. Thanks. The driver version is "31.0.101.2125". |
@kealan-barbieri. Thank you, but how can i get this version code? |
@kealan-barbieri, @feixuedudiao, I'd like to reiterate that support of Intel(R) UHD Graphics 630 is discontinued. |
@vpirogov Thanks. I want to apply oneDNN at HD 630 or lower cpu with fp16 or bf16, which version can support? |
oneDNN v3.3.6 is the last version with HD 630 support. |
@vpirogov Ok ,thanks, i try it . |
@vpirogov @kealan-barbieri thank you to help me. But i test this problem, when the data type for f16, the dims of src ,weight, bias and dst are {1,4,224,384} ,{32, 4, 3, 3},{32} and {1,32,112,192}, the kind is gpu, it can not crash at auto conv_prim = convolution_forward(conv_pd). i test version with v3.3.6 and v.3.3.4. The code that is listed below:
/// @example convolution.cpp #include #include "example_utils.hpp" using namespace dnnl; using tag = memory::format_tag; void convolution_example(dnnl::engine::kind engine_kind) {
*/
/*
*/
*/
auto conv_pd = convolution_forward::primitive_desc(engine,
/*
} int main(int argc, char **argv) {
} |
Summary
Run the example of convolution with the above code :
// Create execution dnnl::engine.
int isGpu = 1;
engine::kind cpu_kind = engine::kind::cpu;
engine::kind gpu_kind = engine::kind::gpu;
Environment
windows 10 cpu is i7-10700 @2.90GHz Compiler vs2022 with DNN_GPU_RUNTIME=ocl
Version
v3.3.6
Problem
It reported "No configurations found" when the code is running "auto conv_prim = convolution_forward(conv_pd);",
????????. . .onednn_verbose,info,oneDNN v3.3.6 (commit N/A)
onednn_verbose,info,cpu,runtime:sequential,nthr:1
onednn_verbose,info,cpu,isa:Intel AVX2
onednn_verbose,info,gpu,runtime:OpenCL
onednn_verbose,info,gpu,engine,0,name:Intel(R) UHD Graphics 630,driver_version:31.0.101,binary_kernels:enabled
onednn_verbose,primitive,info,template:operation,engine,primitive,implementation,prop_kind,memory_descriptors,attributes,auxiliary,problem_desc,exec_time
onednn_verbose,primitive,exec,gpu,reorder,jit:ir,undef,src_f32::blocked:abcd::f0 dst_f16::blocked:aBcd16b::f0,,,1x32x56x96,0.7471
onednn_verbose,primitive,exec,gpu,reorder,jit:ir,undef,src_f32::blocked:abcd::f0 dst_f16:p:blocked:ABcd16b8a::f0,,,4x32x3x3,0.6235
onednn_verbose,primitive,exec,gpu,reorder,jit:ir,undef,src_f32::blocked:a::f0 dst_f16::blocked:a::f0,,,4,0.5051
Assertion !params_gen_.is_empty() failed at F:\Share\OneDNN\v3.3.6\oneDNN-3.3.6_no_verbose\src\gpu\jit\conv\tiler.cpp:2373
No configurations found.
Note
I test the data type with bf16, it not crashed.
The text was updated successfully, but these errors were encountered: