1 Aurora InnovationContinuous Regression Testing for Safer and Faster Refactoring Pejman Ghorbanzade Aurora Innovation3 Aurora Innovation Engineers spend 17 hours per week maintaining software. *Stripe Aurora Innovation A roller coaster story "Touca gives us the con�dence to develop new features faster and with fewer problems." 10+ paying customers 100+ work�ows continuously tested 1000+ unexpected

Faster App @kevinzhow About Kevin • Develop iOS App Since 2011 • Ruby on Rails Since 2010 • Design & Linux Since 2006 • Author of PRODUCTER PNChart • Founder of Catch Inc Fact of Tech Fact CoreData • Build in Support • Package Size • GUI Support Realm • Great Threading Support • Faster • But Fat Threading CoreData Notification Realm Sync When Commit use Batch update Network

大规模生产环境下的 Faster-CPython 主讲人： 王文洋 老板思维 已知：公司有xx个计算集群 每个集群有xxxxx个core Python进程占比xx% 如果：提升 10% 那么：可以节省 xx * xxxxx * xx% * 10%个core 降本 xx * xxxxx * xx% * 10% * n >> 我的工资 结论：。。。 Why Why Faster？ Why CPython？ 项目的存量代码 老板每天看到的PR Why CPython？ Why CPython？ 居然要写代码 居然要动环境 Why CPython？ Faster CPython https://github.com/faster-cpython https://github.com/faster-cpython Faster CPython CPython How To Faster In Older Versions git cherry-pick Faster CPython In 3.6 100% 137% 145% 178% 3.6.15 3.10.1 3.6 FASTER 3.11 How To Faster 3.11 Faster Runtime • Cheaper, lazy Python frames

with mixed precision Tensor Cores on NVIDIA Volta and NVIDIA Turing™, so you can get results much faster than training without Tensor Cores. PyTorch Release 23.07 PyTorch RN-08516-001_v23.07 | 10 implementation paper that leverages mixed-precision arithmetic and Tensor Cores on V100 GPUs for faster training times while maintaining target accuracy. This model script is available on GitHub and NGC implementation, which leverages mixed precision arithmetic by using Tensor Cores on NVIDIA V100 GPUs for 1.3x faster training time while maintaining target accuracy. This model script is available on GitHub and NGC

Smaller = Faster Log scale Just buy new hardware and your code runs faster Year YearPerformance improvement in the ‘80sPerformance improvement in the ‘80s just chill and wait for faster hardware!gle Typical performance evaluationA′ A Which is faster?A′ A Which is faster? 90sA′ A 87.5s Which is faster? 90sA′ A Which is faster? 2.8% faster! A′ 87.5s 90s0.00 0.25 0.50 0.75 1.00 85 Version A A' Is faster than ? A′ A 2.8% faster what about variance?0 5 10 15 85.0 87.5 90.0 92.5 95.0 Time (s) Number of runs Version A A' Which is faster? A′ A ×30 ×300

“Falcon looks great so far. I hacked together a quick test for a tiny server of mine and was ~40% faster with only 20 minutes of work.” “I’m loving #falconframework! Super clean and simple, I finally have as UTF-8 in the response. If data is already a byte string, use the data attribute instead (it’s faster). data bytes – Byte string representing response content. Use this attribute in lieu of body when although such strings are also limited to US-ASCII. Note Falcon can process a list of tuples slightly faster than a dict. Raises: ValueError – headers was not a dict or list of tuple. set_stream(stream, stream_len)[source]

Edit Find in Files 1.8x faster 1 2 0 5 10 15 20 25 30 35 40 45 50 47 27 Series1 Open C++ File 2.4x faster 1 2 0 0.5 1 1.5 2 2.5 2.02 0.82 Series1 IntelliSense Ready* 1.9x faster *) Colorization + Member Member List 1 2 0 1 2 3 4 5 6 5.42 2.84 Series1 Open Member List 29% faster Unreal Engine 4.27.1 IntelliSense Ready 18x faster Learn more at http://aka.ms/18xfaster Performance of Visual Studio Files 1.8x faster Open C++ File 2.4x faster IntelliSense Ready* 1.9x faster *) Colorization + Member List Open Member List 29% faster Unreal Engine 4.27.1 IntelliSense Ready 18x faster Learn more

internet was unprecedented. Consider now that AI user and usage trending is ramping materially faster…and the machines can outpace us. The pace and scope of change related to the artificial intelligence helped, directly or via your work, and are driving technology forward.• Seem Like Change Happening Faster Than Ever? Yes, It Is • AI User + Usage + CapEx Growth = Unprecedented • AI Model Compute Costs OutlineWeekly Active Users, MM 4 Charts Paint Thousands of Words… Seem Like Change Happening Faster Than Ever? Yes, It Is AI User + Usage + CapEx Growth = Unprecedented Developers in Leading Chipmaker’s

(three times faster) and code base size (easily half of our original [framework] code).” “Falcon looks great so far. I hacked together a quick test for a tiny server of mine and was ~40% faster with only wherever you need it. Fast. Same hardware, more requests. Falcon turns around requests several times faster than most other Python frameworks. For an extra speed boost, Falcon compiles itself with Cython when you can install Uvicorn along with more optimized alternatives such as uvloop (a faster event loop), httptools (a faster HTTP protocol implementation) etc: $ pip install uvicorn[standard] See also a longer

(three times faster) and code base size (easily half of our original [framework] code).” “Falcon looks great so far. I hacked together a quick test for a tiny server of mine and was ~40% faster with only wherever you need it. Fast. Same hardware, more requests. Falcon turns around requests several times faster than most other Python frameworks. For an extra speed boost, Falcon compiles itself with Cython when you can install Uvicorn along with more optimized alternatives such as uvloop (a faster event loop), httptools (a faster HTTP protocol implementation) etc: $ pip install uvicorn[standard] See also a longer