For the record, at Uber, we're moving many of our microservices to what @copyconstruct calls macroservices (wells-sized services).
Exactly b/c testing and maintaining thousands of microservices is not only hard - it can cause more trouble long-term than it solves the short-term.
i don't claim to be an expert entrepreneur or anything, but DSSCAW's entire marketing budget last year was about $500. it went entirely to mailing properly-fitted shirts to up-and-coming lady computer scientists active on twitter, and the impact/$ ratio was unbelievable.
https://t.co/dsB6KzSVGc here at DSSCAW we live to make effective use of large pages. so many programs are effectively just churning through their TLBs (not to mention unexpected IPI shootdowns). contract us and get your caches into summer beach shape. you better work, bitch!
first cold atlanta day of this year, and that means cheaper, easier FLOPs, but also more static electricity as we get ready to start a major FPGA project here at DSSCAW-WHQ
while we're here -- been too busy to tweet hot cs news, sorry -- DSSCAW has operated for 10 months. i have no idea whether i'm in tax compliance, nor how to determine as much, and assume some Gödelian principle means i can't actually know. our tax system is a charade, kthxbye.
Here it is -- this is the article accompaniment to our AMD Ryzen TDP deep-dive from last week. We'd prefer you watch the video (as, frankly, that's what pays the bills), but the article is up for those who prefer it and for quicker reference. Find it here: https://t.co/msWbe06rIo
good lord between EPYC 7742 and 3950X the ladies and gents over at @AMD really ate their Wheaties this year. feelin' froggy and jumped up on it for sure https://t.co/pgNhEOzsEh
Exciting to see XNNPack open-sourced: https://t.co/fHKBsQ42Hf - it's a library of optimized floating point arithmetic operations to speed up deep learning calculations on Arm and x86.
We're already using it in @TensorFlow Lite, I'd love to see it more widely adopted!
i love the smell of deliverables in the morning. today marks successful completion of a major contract with @PureStorage; thanks for the hacks, had a great time. enjoy Pure//Accelerate next week!
https://t.co/IFngLqJNXd… how delightful. i remember first learning Jia-Wei/Kung's result in @richvuduc's life-changing @GTCSE 6230, and thinking it one of the real unsung gems of CS. sometimes i run through it aloud in the shower, melodically.
@SignalHound this looks like a rather lovely bit of equipment you've got coming out. you need to figure out some CUDA/OpenCL we can write you in exchange for a hookup. got FFTs? https://t.co/ESfMeGqkzm
https://t.co/CN6exgdw3s that was unexpected. i know what we'll be up to here at DSSCAW WHQ for the next eighty hours straight or so. if these chips are clockin', don't come a-knockin'