コピーされたリンク
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
I'm glad you finally got it to work! I have a new motherboard and Intel Arrow Lake CPU (which has a NPU) that I haven't installed yet, so I was not able to test the NPU functionality of GeekBench AI on my computer. Sorry for the confusion.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Ancak bunun dışında yapay zeka uygulamaları içeren süreçlerde çalışmıyor, bunu nasıl çözebilirim?
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
However, it does not work in processes involving AI applications other than this, how can I solve this?
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Unfortunately, unless applications support it, nothing you can do.
Copilot has a requirement of 40 TOPS. From their thinking when a much more powerful GPU is available, there is no reason for them to develop their code for a NPU. But obviously for laptops and especially phones where battery usage is of concern a NPU is much more power efficient. Additionally, it has its valid uses for certain AI work loads. Its use is much more common on newer phones that support such technologies.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hello iso-uran,
It appears that the NPU utilization is functioning correctly, but it is only activated when OpenVINO is utilized. I would like to extend my gratitude to @PC1997 for providing your knowledge to @iso-uran in this thread. I will investigate this matter internally to conduct a more thorough analysis
Best regards,
Dhanniel M.
Intel Customer Support Technician
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hello iso-uran,
Thank you for your patience. Based on my investigation, I concur with the remarks made by @PC1997. The utilization of the Neural Processing Unit (NPU) can vary depending on how an application is programmed to leverage either the NPU or the GPU. Since you are currently working on a language model, I recommend consulting the community or support resources of the AI framework you are using. They can provide guidance on how to effectively utilize the NPU within your framework.
If you have any further questions, please feel free to create a new thread, as this inquiry will now be closed and no longer monitored.
Best regards,
Dhanniel M.
Intel Customer Support Technician
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Thank you for your support. As far as I understand, this is a situation caused by the settings of the applications. Do I understand correctly? Are these settings on the software side or in a way that I can interfere with, how can I change the settings to use the NPU actively?
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Currently, Microsoft Copilot requires the Arm-based Snapdragon Elite X chip. This is one of the CPU+NPU processors that meets the minimum requirements of 40 TOPS.
Support will be added for other NPU's from Intel and AMD that meet the minimum processing performance in the future.
Here is an excellent resource on how to program applications to use the Neural Processing Unit in Copilot:
https://learn.microsoft.com/en-us/windows/ai/npu-devices/
Thanks to @DhannielM_Intel for the mention.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Finally an actual full solution is available:
https://www.extremetech.com/computing/microsoft-copilot-for-windows-is-transitioning-to-a-native-app
"The app is designed to use the device's Neural Processing Unit (NPU) for task completion and query responses, which could enhance speed and efficiency. Currently, this is limited to Windows Insiders, with a broader release anticipated later this year."
