On Tue, Nov 21, 2017 at 5:50 PM, Tim Harvey <thar...@gateworks.com> wrote: > On Fri, Nov 10, 2017 at 3:57 AM, Outback Dingo <outbackdi...@gmail.com> wrote: >> Yeah i can do tthat, as a side note, do you know anything about >> cpu_scaling or thermal management on this board maybe ? it runs >> terribly hot >> > > As you guys worked out the GW5520 does not use the i.MX6 FEC GbE and > instead has 2x Intel i210 mac/phy devices requiring the igb driver. > This was done for performance reasons as the i.MX6 FEC is bandwidth > limited due to it's internal bus routing to the SoC. > > Regarding the board running hot I can say the following: > - there's about 1.2W of power on the PCIe bridge. The i.MX6 only has a > single PCIe host controller so we need a bridge for the 2x GbE's and > the 2x miniPCI sockets. We have not found any way to lower the power > draw of the PLX switch > - to reduce power draw on the i.MX6 (where the rest of the heat is > coming from) you can use CPU frequency scaling as suggested > - adjust your thermal solution to your enclosure / mouting. We have > many customer that use thermal gap filling pads on both the i.MX6 as > well as the PLX PCIe switch to get heat off the board > > Please feel free to send questions to supp...@gateworks.com - I > happened across this post rather randomly.
so we ran make kernel_menuconfig and set the performance to ondemand from the performance setting in the kernel and recompiled / reflashed however thinking the board would run cooler it appears with a temperature gun that the device is actually running hotter..... any thoughts on this and is it possible to think the changes I made to the cpu throttling somehow prevent the cpu getting into idle/sleep mode? and is the scaling_min_freq iadjustable? Thanks > > Regards, > > Tim > > Tim Harvey - Principal Software Engineer > Gateworks Corporation - http://www.gateworks.com/ > 3026 S. Higuera St. San Luis Obispo CA 93401 > 805-781-2000 _______________________________________________ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
