Talk:Nvidia Tesla

Duplicated information
The entire table of models in this article is basically a copy of the table in List of Nvidia graphics processing units. Duplicated information is bad practice and hard to maintain. Should we remove the table from Nvidia Tesla and link to the other page instead? Mbevand (talk) 13:16, 8 December 2015 (UTC)

Half Precision FP16 is not in Data Table
Half Precision FP16 is not in Data Table — Preceding unsigned comment added by 95.90.228.8 (talk) 19:53, 6 February 2018 (UTC)

Tesla T4 Turing Card Data available
Only 75 Watt for PCIe only See

https://www.techpowerup.com/gpu-specs/tesla-t4.c3316 — Preceding unsigned comment added by 79.214.184.205 (talk) 21:22, 22 December 2018 (UTC)

The Ampere data seems to be inconsistent between models
Under some definition the number of CUDA cores per SM is now 128, because with Ampere there are 64 FP32 Cores and 64 mixed FP32/INT cores per SM. The TFLOPs numbers probably are too high for some models as a FMA operation is counted twice under some definitions, but the column in this article explicitly states FMA. Sebastian --2A00:20:C059:E92B:556:813F:DEA6:74B2 (talk) 19:41, 5 November 2021 (UTC)

Content Body needs rewriting
The main content body of this article needs some serious updating. It's pretty outdated and could definitely use some more information on use cases, features, etc. Also, as another user mentioned, the majority of the article is literally just a copy and paste from List of Nvidia graphics processing units and doesn't really contribute any information. Justanotherintrovert (talk) 13:48, 29 April 2022 (UTC)