Friday, 2017-11-17

*** crushil has quit IRC00:16
*** goldenfri has quit IRC08:44
*** zhipeng has joined #openstack-cyborg13:46
*** jkilpatr has quit IRC14:03
*** jkilpatr has joined #openstack-cyborg14:16
zhipengjkilpatr ttk2[m] i've written a coarse draft for my idea mentioned last team meeting14:17
*** zhipeng has quit IRC14:17
ttk2[m]zhipeng: so lets say we do this and people can schedule based on types, is their application now responsible fo rhandling supporting all the different types of accerlators they may get with the same request?14:19
ttk2[m]I guss you can narrow it down14:19
ttk2[m]deep learning + cuda X support = get different types of gpu's but the code alway sworks14:20
zhipengh[m]That is another important part if the equation needs to be solved14:24
zhipengh[m]But for HPC and NFV, it should work without too much trouble14:25
zhipengh[m]For preloaded FPGA programs14:26
ttk2[m]well yes, provided we handle compatibility checking behind the scenes some.14:29
*** crushil has joined #openstack-cyborg14:59
zhipengh[m]ttk2: the main point here is about scalability. Name and type won't get us far for accelerator, that is the fundamental problem for us15:35
ttk2[m]how so? too specific?15:36
zhipengh[m]Yep, especially for FPGA15:37
*** goldenfri has joined #openstack-cyborg17:40
*** crushil has quit IRC18:26
*** crushil has joined #openstack-cyborg18:49

Generated by 2.15.3 by Marius Gedminas - find it at!