2022-04-03
Wedding Postponement Announcement @ Cui Ping Shan Guesthouse, Shijiazhuang

Update: Due to the pandemic, the wedding originally scheduled for May 1, 2022, and postponed to July 17, 2022, is postponed again to May 1, 2023. Please move to the latest version of the wedding invitation, thank you!

Read More

2022-03-31
A Few Things About Buying a House in Haidian

On February 8, 2021, at 08:53, I, my girlfriend, and my parents were on our way to work. I finally made up my mind, said this is the one, and transferred a deposit of 100,000 yuan to the property developer. At this point, it had only been a month since we started looking at houses intensively.

Why the rush?

There are three contexts: First, Beijing requires a household registration or work residence permit to buy a house. Although I could have solved the household registration when I started my job after graduating with a doctorate in June 2019, the process was slow. I got the acceptance letter from the Beijing Municipal Human Resources and Social Security Bureau in December 2019, the household registration transfer certificate in September 2020, and my new ID card in January 2021.

Second, there are not many new properties in Haidian District, and demand exceeds supply. Although Haidian District has quite a lot of vacant land, the land released each year is very limited. Plus, Haidian District is home to the wealthy, such as major internet companies. The house prices in the core area are as high as over 100,000 yuan, and those outside the Fifth Ring Road are also 80,000 yuan. Some remote or shared property rights houses can be 50~60,000 yuan. Even such houses are sold out in seconds. For example, there are two properties near where I live (Haidian District, outside the Fifth Ring Road, Shucun). One is the luxury Yuanming Tiansong, which opened in November 2021, with 238 houses, the lowest total price is also 23 million yuan, sold out in 15 minutes, reportedly all paid in full; the other is the demand-improvement type Xuefu No.1 Courtyard, with more than 900 houses, the lowest total price is 11 million yuan, also sold out in one day. Obviously, I don’t have the financial strength to buy such a house.

Third, the down payment ratio in Beijing is very high, non-ordinary residences (houses in the city are generally non-ordinary residences) first set 40%, second set 80%, I was determined to be the second set because I had a loan record in my hometown, so the down payment ratio is as high as 80%, so it is impossible to buy too expensive a house. For me, the problem of taking on a loan is not big, according to the bank’s algorithm, I can take on about 12 million yuan in loans, but the down payment is the main bottleneck.

City Choice

Faced with such a high down payment, we mainly considered several options:

  1. Buy a smaller house in a prime location in Haidian, and change houses later if there is a need to improve life. At this time, we need to consider mainly include the convenience of commuting, the comfort of living, and the investment property of real estate. Although many people are pessimistic about the property market against the backdrop of a low birth rate and economic downturn, I believe that there is still a strong rigid demand in prime locations in first-tier cities, and it is still in short supply.
  2. Buy a slightly larger house in an ordinary location in Haidian or Changping District, basically one step in place in terms of area, and the house price is not expensive, generally 50~60,000 yuan per square meter. But Changping’s school district is not as good as Haidian’s, and if the household registration is moved to Changping, whether future policies will allow it to be moved back to Haidian is an unknown (currently it is possible to move freely within the city).
  3. My classmates in Hangzhou basically all bought houses of more than 120 square meters, because the house prices in Hangzhou are only 3040,000 yuan per square meter (recently the increase is also quite fierce, it is already 4050,000 yuan), and Hangzhou’s down payment is the first set 30%, the second set 60%, much friendlier than Beijing. Hangzhou’s Binjiang is developing well, Huawei Hangzhou Research Institute and North Research Institute (Environmental Protection Park) also have a history of more than ten years, Hangzhou Research Institute is near Binjiang Internet Town, there are several good large shopping malls, Environmental Protection Park is mostly wasteland (hope it can be developed in ten years). I have been on a business trip to Hangzhou Research Institute for more than half a year, to be honest, my impression of Hangzhou is good, our engagement ring and wedding ring were bought in Hangzhou.
  4. Because of work reasons, I went on business trips to Chengdu and Xi’an, both cities gave me a good impression, especially Chengdu’s life is slower. The house prices are a notch lower than Hangzhou. At the entrance of Yaotang Mansion in Chengdu, I often see various sports cars and luxury cars, the entrance is an international school, life is very convenient, it looks more like a big city than Shangdi and Environmental Protection Park in Beijing where I am. These colleagues earn about the same salary as me, but they live very comfortably in Chengdu and Xi’an, where the house price may be the most important factor. The first time I went to Chengdu Research Institute was at the end of winter and the beginning of spring, when the trees in Beijing were still withered, but the flowers were red and the trees were green in Chengdu Research Institute, with weeping cherry blossoms blooming in patches, reflecting in the lake. When we went out to play, we also went to Suzhou and other cities, the water towns of Suzhou in the south of the Yangtze River are also unforgettable.
  5. My friends in the United States basically all bought houses, for example, in Seattle, a house of 180 square meters can be bought for 1 million US dollars, which is even cheaper than many houses in Hangzhou. Even considering the factor of rising house prices in Seattle (up more than 50% in a year), it is still only the price of a better house in Hangzhou, and most houses in Beijing, even in Changping, are more expensive. In addition, the area of a house cannot be directly compared with a house in China, because there is a lot of shared area in high-rise residences in China, for example, the house I finally bought has a shared area of up to 25%. Europe and the United States do not have such exaggerated down payment ratios. The comfort of living in a house is generally higher than that of a high-rise building. But choosing to stay in China is choosing a more competitive life. If considering going to Europe or the United States in the future, according to the income level of the same level of internet companies, the pressure to buy a house will be much less.
Read More

2022-03-06
Release of 2017-2018 Positioning Trajectory

The positioning trajectory is important data for studying life routines and summarizing life travel history with minute-level accuracy. The positioning trajectory information from January 21, 2017, to July 7, 2018, is now released as follows:

Click here to see the 2017-2018 positioning trajectory

Please note that sometimes there will be no movement in the positioning for a whole day because I did not carry the locator or the locator ran out of power. From this batch of trajectory data, it can be seen that during my PhD, I went to the company mostly at noon or even in the afternoon, and did not return home from the company until midnight. Sometimes I would go for a run as soon as I got up or at midnight.

Read More

2022-01-08
Our Wedding Photos

October 6, 2021, Lijiang

Short video (HD, 467 MB, 6:26)

Click here to view the wedding photo online album (70 refined photos)

Click here to view the wedding photo online album (some original photos)

Read More

2021-08-24
1Pipe: Scalable Total Order Communication in Data Center Networks

1Pipe is a causal and total order communication primitive to scatter groups of messages via data center network. With in-network computation using Barefoot or Arista switches, 1Pipe achieves scalability and high performance with low CPU and network overheads. Published in SIGCOMM’21.

Read More

2021-07-06
AKG: Automatic Kernel Generation for Neural Processing Units using Polyhedral Transformation

AKG (Auto Kernel Generator) is a tensor compiler for NPUs. AKG leverages polyhedral schedulers to perform a much wider class of transformations, and extends the semantics of the polyhedral representation to combine complex tiling techniques and hierarchical fusion strategies. Published in MICRO’20 and PLDI’21.

Read More

2021-05-01
Engagement Party of Li Bojie & Meng Jiaying

12:00, May 1, 2021, Hebei Yunzhen Century Hotel

Click here to see the photo book I made

Proposal video: (HD, 111 MB, 4:23)

Engagement ceremony video: (SD, 37 MB, 10:41)

Read More

2021-02-26
Looking for a topic? Breaking through bottlenecks? Teacher-student relationship? MSRA Joint Training Star Friends Guide You Through

This article is reprinted from the Microsoft Research Asia (MSRA) official account, thanks to MSRA for the invitation!

Pursuing a doctorate is a lonely and challenging journey. If there are senior brothers and sisters to guide you, they can definitely help you out of confusion and bravely make choices. On February 8, five alumni who graduated from the Microsoft Research Asia Joint Doctoral Training Program shared their personal insights during their doctoral studies and subsequent work online to answer questions and encourage everyone to stick to their dreams. This event was organized and hosted by Sun Lijun, Senior Academic Manager of Microsoft Research Asia, and Dou Anqi, the person in charge of the internship program.

We hope that the sharing of these senior brothers and sisters can encourage and inspire every one of you on the road to a doctorate, and let you enjoy this beautiful and challenging time~

Guest speakers:

  • Fu Xiaoming, 2016 USTC-Microsoft Joint Training Doctor, currently Associate Researcher at University of Science and Technology of China
  • Zhang Chi, 2017 Sun Yat-sen University-Microsoft Joint Training Doctor, currently DeepMotion Co-Founder, R&D Director
  • Huang Danqing, 2019 Sun Yat-sen University-Microsoft Joint Training Doctor, currently a researcher at Microsoft Research Asia
  • Li Bojie, 2019 USTC-Microsoft Joint Training Doctor, currently a senior research engineer at Huawei 2012 Lab
  • Li Xiao, 2019 USTC-Microsoft Joint Training Doctor, currently a researcher at Microsoft Research Asia
Read More

2019-12-08
Doctoral Thesis: High Performance Data Center Systems Based on Programmable Network Cards

Doctoral Thesis from University of Science and Technology of China, Author: Bojie Li

Chinese Version: High Performance Data Center Systems Based on Programmable Network Cards (PDF, 8 MB)

AI Translated Unofficial English Version: High Performance Data Center Systems with Programmable Network Interface Cards (PDF, 8 MB)

Publication Date: 2019-05-26.

Read More

2019-08-19
SocksDirect: Datacenter Sockets can be Fast and Compatible

Communication intensive applications in hosts with multi-core CPU and high speed networking hardware often put considerable stress on the native socket system in an OS. Existing socket replacements often leave significant performance on the table, as well have limitations on compatibility and isolation.

In this paper, we describe SocksDirect, a user-space high performance socket system. SocksDirect is fully compatible with Linux socket and can be used as a drop-in replacement with no modification to existing applications. To achieve high performance, SocksDirect leverages RDMA and shared memory (SHM) for inter-host and intra-host communication, respectively. To bridge the semantics gap between socket and RDMA/SHM, we optimize for the common cases while maintaining compatibility in general. SocksDirect achieves isolation by employing a trusted monitor daemon to handle control plane operations such as connection establishment and access control. The data plane is peer-to-peer between processes, in which we remove multi-thread synchronization, buffer management, large payload copy and process wakeup overheads in common cases. Experiments show that SocksDirect achieves 7 to 20x better message throughput and 17 to 35x better latency compared with Linux socket, and reduces Nginx HTTP latency by 5.5 times.

Read More
RSS