我想要使用3090跑一个很大的四面体模型,101318个顶点,401016个四面体。但是会报错,提示如下:
SNode index might be out of int32 boundary but int64 indexing is not supported yet. Struct fors might not work either.
[E 09/01/24 23:57:53.005 32689] [cuda_driver.h:operator()@92] CUDA Error CUDA_ERROR_OUT_OF_MEMORY: out of memory while calling malloc_async_impl (cuMemAllocAsync)
有什么解决方案吗?24g显存应该够的吧,只是顶点索引超出就没办法跑了吗?