Hangchen YU

icon company Meta Platforms icon location Menlo Park, CA icon location A pupil in systems

Results 14 comments of Hangchen YU

Compliation flags

Hi, sorry that I never ran these openmp benchmarks by myself, but I'd say that O3 optimization is faster *in general*. Some algorithms and data structures may fail the higher-degree...

Compiling errors when unsing benchmark-cuda

Hi William, have you downloaded the dataset from http://lava.cs.virginia.edu/Rodinia/download_links.htm?

[BUG] ODrive corrupted my file on Googledrive

Thanks for sharing @nophead . This is really a severe bug and makes me pause using odrive. I'd wait and look forward to the analysis and fix of this issue.

UTF-8

Is there any update for this? ``` mosh-server needs a UTF-8 native locale to run. Unfortunately, the local environment ([no charset variables]) specifies the character set "US-ASCII", The client-supplied environment...

Soft lockup

Hi Weixi, What Java benchmark are you running? It looks like the spinlock is held quite for a while by some other thread. Because this project isn't not in active...

更改配置文件保存路径

请问linux下配置文件保存在哪里？我locate了gui-config.json文件，却没有结果

Support CUDA 11.0

> What's required for this? To support CUDART lib, we need to annotate public CUDA APIs as well as internal APIs such as `__cudaRegisterFatBinary`. CUDA >=10.2 introduced a few new...

Preserve ordering of ava_async APIs between multiple threads

The mistake happens when the multi-threaded program introduces some other assumptions. For example, it can assume a CUDA kernel is "enqueued" when a `cudaLaunchKernel` API call returns. After that, another...

Migration barrier

We probably need some way to organize codes for the experimental purpose. Will take consideration on it.

A segmentation fault may accur in "__cudaPopCallConfiguration" in tf_opt.c

Thanks for reporting this issue. We are refactoring CUDA specs (including tensorflow specs) in these days and I'll take a check on this problem.