[TVM Discuss] [Application] Competitive gemm/matmul example?

2020-06-05 Thread Iz Beltagy via TVM Discuss
Thanks, @jknight for the updates. This is great. > Let us know if you have additional ideas as well and we’ll see what we can do > on our side! If I have to mention something, it would be supporting fp16 using `half2` which usually leads to compute savings, not just memory savings. ---

[TVM Discuss] [Questions] Tvm.contrib.graph_runtime.create error

2020-06-05 Thread Yanming Wang via TVM Discuss
Then increasing the memory of your machine will likely solve this issue. --- [Visit Topic](https://discuss.tvm.ai/t/tvm-contrib-graph-runtime-create-error/6868/4) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](htt

[TVM Discuss] [Questions] Tvm.contrib.graph_runtime.create error

2020-06-05 Thread Heliqi via TVM Discuss
I load a very large model --- [Visit Topic](https://discuss.tvm.ai/t/tvm-contrib-graph-runtime-create-error/6868/3) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](https://discuss.tvm.ai/email/unsubscribe/3c005

[TVM Discuss] [Questions] Tvm.contrib.graph_runtime.create error

2020-06-05 Thread Yanming Wang via TVM Discuss
I saw this error (OSError: [WinError - 529697949] Windows Error 0xe06d7363) when I was trying to load a very large model (~600MB) using 32-bit TVM on Windows. Are you using 32-bit TVM or compiling a large model? --- [Visit Topic](https://discuss.tvm.ai/t/tvm-contrib-graph-runtime-create-e

[TVM Discuss] [Questions] How is __tvm_module_startup invoked?

2020-06-05 Thread JC Li via TVM Discuss
One thing I still don't understand is how __tvm_module_startup is being called at program startup though? @liangfu? Anyone? --- [Visit Topic](https://discuss.tvm.ai/t/how-is-tvm-module-startup-invoked/6891/4) to respond. You are receiving this because you enabled mailing list mode. To u

[TVM Discuss] [Questions] How is __tvm_module_startup invoked?

2020-06-05 Thread JC Li via TVM Discuss
I think I got it. The LLVM host module compilation generates it. src/target/llvm/llvm_module.cc: ``` TVM_REGISTER_GLOBAL("target.build.llvm").set_body_typed([](IRModule mod, std::string target) { auto n = make_object(); n->Init(mod, target); return runtime::Module(n); }); ... void Init(

[TVM Discuss] [Application] Competitive gemm/matmul example?

2020-06-05 Thread Jason Knight via TVM Discuss
Sorry to hear about your difficulties @srush. Indeed we have a number of irons in the fire to address this: * Pip instalability of TVM is something we are still targetting for Q2 (so this month!) * We have an offer out to someone for a full time Developer Advocate role to help build docs (eg t

[TVM Discuss] [Questions] How is __tvm_module_startup invoked?

2020-06-05 Thread JC Li via TVM Discuss
@liangfu, I think this question goes to you best, :slight_smile: The above question originates when I looked at this function in src/runtime/crt/packed_func.h: ``` TVMPackedFunc* g_fexecs = 0; uint32_t g_fexecs_count = 0; // Implement TVMModule::GetFunction // Put implementation in this file s

[TVM Discuss] [Questions] How is __tvm_module_startup invoked?

2020-06-05 Thread JC Li via TVM Discuss
I'm studying apps/bundle_deploy example code which runs on crt runtime. This function was called at the beginning to pre-load all operator functions: ``` // src/runtime/crt/crt_backend_api.c: int TVMBackendRegisterSystemLibSymbol(const char* name, void* ptr) { g_fexecs = vrealloc(g_fexecs, si