Is the GPU being used?

[FalconLK]

Hello,

I made it running but not sure if the GPU is being used.

Ignoring visible gpu device (device: 0, name: GRID K520, pci bus id: 0000:00:03.0, compute capability: 3.0) with Cuda compute capability 3.0. The minimum required Cuda capability is 3.5.

Could you please confirm this?

[yh1105]

Hello,GPU is used, please check your local config.-------- 原始邮件 --------主题：[yh1105/datasetforTBCCD] Is the GPU being used? (#1)发件人：FalconLK 收件人：yh1105/datasetforTBCCD 抄送：Subscribed Hello, I made it running but not sure if the GPU is being used.

Ignoring visible gpu device (device: 0, name: GRID K520, pci bus id: 0000:00:03.0, compute capability: 3.0) with Cuda compute capability 3.0. The minimum required Cuda capability is 3.5.

Could you please confirm this?

—You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or mute the thread. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/yh1105/datasetforTBCCD/issues/1", "url": "https://github.com/yh1105/datasetforTBCCD/issues/1", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

[yh1105]

another question, when you run bcb...., please make sure you are using python3, I'm not sure if using "python bcb..."can get the token node(AST+), if using "python bcbwithidfinetune.py", it might have the same efficent as "python bcbnoidfinetune.py". so please using "python3 bcb..." for bcb.

[yh1105]

for bcb, it has 1.5 million training pairs, for 1 epoch ,it run about 12 hours, so I suggest using "nohup python3 bcbwithidfinetune.py >recordbcbwithidfinetune.txt 2>&1 &" .

[FalconLK]

Thank you very much for the answer!

However, as far as I know, the following statement is meaning that the GPU is not being used.

Ignoring visible gpu device (device: 0, name: GRID K520, pci bus id: 0000:00:03.0, compute capability: 3.0) with Cuda compute capability 3.0. The minimum required Cuda capability is 3.5.

You can check it out as well in here (https://stackoverflow.com/questions/50995707/ignoring-visible-gpu-device-with-compute-capability-3-0-the-minimum-required-cu)

That's why I'm re-building the tensorflow from the source... it also takes a long long time..

Thank you anyways for the details!

[yh1105]

Oh... If it also takes a long long time for you, I think you can only use 50-70 thousands pairs, only add a break in train.

[FalconLK]

Oh... If it also takes a long long time for you, I think you can only use 50-70 thousands pairs, only add a break in train.

I mean building the tensorflow from the source is also taking a long time :)

I actually want to make sure the GPU is working properly without having issues like the one I mentioned :)

[FalconLK]

I found a solution to make the GPU work.

To be clear, the statement

Ignoring visible gpu device (device: 0, name: GRID K520, pci bus id: 0000:00:03.0, compute capability: 3.0) with Cuda compute capability 3.0. The minimum required Cuda capability is 3.5.

means that the GPU is not being used. So you can check if you saw the same message in your machine.

I don't know why the process (GPU version) is not that faster than the normal (CPU version) process.

There still must be something wrong..

[yh1105]

`2019-05-09 21:57:52.916428: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA 2019-05-09 21:57:53.540173: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties: name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582 pciBusID: 0000:89:00.0 totalMemory: 10.91GiB freeMemory: 9.52GiB 2019-05-09 21:57:53.540237: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:89:00.0, compute capability: 6.1)``

Sorry for I just saw the question. I havn't saw the same message in my machine, the message in my machine is list above.

[FalconLK]

I found the problem and fixed it ;)

It was related to Cuda and CudNN. Although I couldn't be successful with the latest version, I managed to do it at least.

I'll ask you when I have further questions.

Thank you very much for replying anyways! Awesome 👍