当前位置:网站首页>Jetson XAVIER NX上ResUnet-TensorRT8.2速度與顯存記錄錶(後續不斷補充)
Jetson XAVIER NX上ResUnet-TensorRT8.2速度與顯存記錄錶(後續不斷補充)
2022-07-02 19:59:00 【少有人走的路_心智旅程】
ResUnet-TensorRT8.2速度與顯存記錄錶
| 精度 | 模式 | 圖像尺寸 | 類別數 | 批次 | 線程數 | 推理時間 | 完整處理時間 | 顯存 |
|---|---|---|---|---|---|---|---|---|
| FP32 | 20W 2CORE | 512*640 | 2 | 1 | 1 | 168ms | 179ms | 2.2G |
| FP32 | 20W 2CORE | 512*640 | 2 | 1 | 2 | 172ms | 184ms | 3.0G |
| FP16 | 20W 2CORE | 512*640 | 2 | 1 | 1 | 58ms | 68ms | 1.6G |
| FP16 | 20W 2CORE | 512*640 | 2 | 1 | 2 | 58ms | 68ms | 1.9G |
| FP32 | 20W 2CORE | 512*612 | 6 | 1 | 1 | 167ms | 209ms | 2.2G |
| FP32 | 20W 2CORE | 512*612 | 6 | 1 | 2 | 170ms | 234ms | 3.0G |
| FP16 | 20W 2CORE | 512*612 | 6 | 1 | 1 | 57ms | 97ms | 1.6G |
| FP16 | 20W 2CORE | 512*612 | 6 | 1 | 2 | 58ms | 106ms | 1.9G |
說明:
1.模式是指Jetson設備的功耗模式,對於本人的Jetson XAVIER NX來說,總共有8種模式,如果想達到最大推理速度的話,選擇20W 2CORE模式。在主界面的右上角有個MODE的選擇,選擇20W 2CORE模式即可。
(本人選擇20W 6CORE測試下來跟20W 2CORE差不多,只快了1ms,所以選擇20W 2CORE即可)

2.推理時間是指平均每張圖進行doInference(即執行cudaMemcpyAsync)所需要的推理時間。
完整處理時間推理時間加上前處理與後處理時間。
3.對於Jetson設備來說,CPU和GPU共用,所以顯存就是內存。對於Jetson XAVIER NX來說內存總共8G。
而查看的方式不能直接使用nvidia-smi的命令行,必須安裝jetson-stats。
具體操作方式可參考以下博客。
Jetson設備上查看顯存(內存)——jetson-stats
4.為什麼本人會有8個模式,而且這個系統下的TensorRT是8.2.1.8版本,不是7版本,猜測原因是在最初燒錄系統的時候使用的鏡像是比較新的。
而且相比TensorRT7版本,速度快了近20ms,具體可以看本人之前的博客。
Jetson XAVIER NX上ResUnet-TensorRT7速度與顯存記錄錶(後續不斷補充)
边栏推荐
- Istio部署:快速上手微服务,
- What is the Bluetooth chip ble, how to select it, and what is the path of subsequent technology development
- 疫情封控65天,我的居家办公心得分享 | 社区征文
- API documentation tool knife4j usage details
- Sometimes only one line of statements are queried, and the execution is slow
- 职场四象限法则:时间管理四象限与职场沟通四象限「建议收藏」
- 功能、作用、效能、功用、效用、功效
- 外包干了三年,废了...
- ShardingSphere-JDBC5.1.2版本关于SELECT LAST_INSERT_ID()本人发现还是存在路由问题
- Taiwan SSS Xinchuang sss1700 replaces cmmedia cm6533 24bit 96KHz USB audio codec chip
猜你喜欢

One side is volume, the other side is layoff. There are a lot of layoffs in byte commercialization department. What do you think of this wave?

RPD出品:Superpower Squad 保姆级攻略
![[real case] trap of program design - beware of large data](/img/bd/d72cc5ce23756cea873c9ced6b642a.jpg)
[real case] trap of program design - beware of large data

KT148A语音芯片ic的软件参考代码C语言,一线串口

Spark source code compilation, cluster deployment and SBT development environment integration in idea

Complete example of pytorch model saving +does pytorch model saving only save trainable parameters? Yes (+ solution)

KT148A语音芯片使用说明、硬件、以及协议、以及常见问题,和参考代码
In depth understanding of modern web browsers (I)

Yes, that's it!

Set up sentinel mode. Reids and redis leave the sentinel cluster from the node
随机推荐
Automatic reading of simple books
Taiwan SSS Xinchuang sss1700 replaces cmmedia cm6533 24bit 96KHz USB audio codec chip
Implementation of online shopping mall system based on SSM
Spark source code compilation, cluster deployment and SBT development environment integration in idea
Shardingsphere jdbc5.1.2 about select last_ INSERT_ ID () I found that there was still a routing problem
Conscience summary! Jupyter notebook from Xiaobai to master, the nanny tutorial is coming!
Cs5268 perfectly replaces ag9321mcq typec multi in one docking station solution
Solution to blue screen after installing TIA botu V17 in notebook
Development skills of rxjs observable custom operator
【每日一题】241. 为运算表达式设计优先级
rxjs Observable 自定义 Operator 的开发技巧
笔记本安装TIA博途V17后出现蓝屏的解决办法
Postman download and installation
面试经验总结,为你的offer保驾护航,满满的知识点
KT148A语音芯片ic的硬件设计注意事项
简书自动阅读
[Chongqing Guangdong education] reference materials for labor education of college students in Nanjing University
For (Auto A: b) and for (Auto & A: b) usage
Introduction to mongodb chapter 03 basic concepts of mongodb
[internship] solve the problem of too long request parameters