byronyi
Re:Graphcore Looks Like a Complete Failure in Machine Learning Trainin...
on Friday July 02, 2021 @01:54AM (#27708211)
The analysis is bullshit.
Re:Cerebras’ new monster AI chip adds 1.4T transistors
on Friday April 23, 2021 @05:03AM (#26912609)
I'd bet on SPIR-V.
Re:Cerebras’ new monster AI chip adds 1.4T transistors
on Friday April 23, 2021 @03:48AM (#26912172)
+1 please put some info in your profile.
Re:DLRM: An advanced, open-source deep learning recommendation model
on Wednesday July 03, 2019 @12:40PM (#20345983)
I don't get it. How is it different from "Wide & Deep Learning for Recommender Systems (2016)" by Google?
The whole project, up til now, contains less than 2,500 lines of Python, and 271 lines of Markdown as doc, 159 lines of Shell script, and a Kaggle benchmark on Criteo CTR which dated back to 2014. The project seems to be published by FB Research, while I do not see much "research" value in this project or the companion blog.
Re:256 TFlops on a Single 7nm Chip: Huawei Ascend 910 Unveiled
on Wednesday October 10, 2018 @04:44AM (#18183096)
Unlike TPU that is only made available through Google Cloud Platform, Ascend 910 will be available on market in 2019 Q2, in form of PCIe line cards and OEM servers.
Show HN: GPU Direct RDMA Out-Of-Band Tensor Transport for TensorFlow
Re:Fixing Python Performance with Rust
on Wednesday October 19, 2016 @09:51PM (#12749318)
Indeed. A more interesting comparison would be between C/C++ based shared library and Rust based one.
Re:Fixing Python Performance with Rust
on Wednesday October 19, 2016 @09:49PM (#12749314)
Why does it have to do with scaling in the first place?
Performance has nothing to do with scalability whatsoever, by the way.