一个服务器上,同时安装了Nvidia 2070 super显卡和Nvidia 3090显卡,安装paddle paddle之后,在2070s显卡上可正常使用,如下图所示:
paddle.fluid.install_check.run_check()
Running Verify Paddle Program ...
W1028 18:11:05.836931 11559 device_context.cc:237] Please NOTE: device: 0, CUDA Capability: 75, Driver API Version: 11.1, Runtime API Version: 10.0
W1028 18:11:05.839093 11559 device_context.cc:245] device: 0, cuDNN Version: 7.5.
W1028 18:11:05.839110 11559 device_context.cc:271] WARNING: device: 0. The installed Paddle is compiled with CUDNN 7.6, but CUDNN version in your machine is 7.5, which may cause serious incompatible bug. Please recompile or reinstall Paddle with compatible CUDNN version.
Your Paddle works well on SINGLE GPU or CPU.
I1028 18:11:07.455816 11559 parallel_executor.cc:440] The Program will be executed on CUDA using ParallelExecutor, 1 cards are used, so 1 programs are executed in parallel.
I1028 18:11:07.456137 11559 build_strategy.cc:365] SeqOnlyAllReduceOps:0, num_trainers:1
I1028 18:11:07.456391 11559 parallel_executor.cc:307] Inplace strategy is enabled, when build_strategy.enable_inplace = True
I1028 18:11:07.456578 11559 parallel_executor.cc:375] Garbage collection strategy is enabled, when FLAGS_eager_delete_tensor_gb = 0
Your Paddle works well on MUTIPLE GPU or CPU.
Your Paddle is installed successfully! Let's start deep Learning with Paddle now
但是在3090显卡上验证的时候,却一直卡住,不出结果,如下:
Running Verify Paddle Program ...
W1028 18:04:00.730057 10037 device_context.cc:237] Please NOTE: device: 0, CUDA Capability: 86, Driver API Version: 11.1, Runtime API Version: 10.0
W1028 18:04:00.732218 10037 device_context.cc:245] device: 0, cuDNN Version: 7.5.
W1028 18:04:00.732234 10037 device_context.cc:271] WARNING: device: 0. The installed Paddle is compiled with CUDNN 7.6, but CUDNN version in your machine is 7.5, which may cause serious incompatible bug. Please recompile or reinstall Paddle with compatible CUDNN version.
亏了亏了 或者是自己编译?
后续肯定会支持的 害得再等等
太难了
CPU能跑。。GPU不能跑 。亏大了~
买了张3060ti来搞学习的 。结果跑不过。 郁闷死 。不是京东购入 退不了~
2.0好像支持11了把? 有人试过吗
听tensorflow2.4支持CUDA 11
最近不是一直听说30系显卡的事,现在很多框架都还不支持CUDA11,所以外头很多人收2080Ti来组深度学习的机器。
等等更新吧 2.0是不是可以用? 听说2.0可以用11
等版本更新,目前应该不行吧,应该还不支持那么高的cuda
我也遇到了。试了各个版本的paddle以及不同的cuda版本,都是一样的结果,验证的时候就卡住,算法都执行不了。有没有什么方案呀,为了用paddle才买来3090显卡,结果用不了,1万多白买了啊。
我记得是的
3090只支持CUDA11.1吧,PaddlePaddle还U不支持
这个好像还不支持呢吧? 3090不是听说需要11才支持?
我也一直不出结果,CUDNN是7.6
Please recompile or reinstall Paddle with compatible CUDNN version.