[Libre-soc-dev] daily kan-ban update 03jul2022
tpearson at raptorengineering.com
Tue Jul 5 18:08:32 BST 2022
----- Original Message -----
> From: "Libre-Soc General Development" <libre-soc-dev at lists.libre-soc.org>
> To: "Libre-Soc General Development" <libre-soc-dev at lists.libre-soc.org>
> Cc: "cestrauss" <cestrauss at gmail.com>
> Sent: Tuesday, July 5, 2022 11:54:02 AM
> Subject: Re: [Libre-soc-dev] daily kan-ban update 03jul2022
> Em 04/07/2022 19:02, Cesar Strauss escreveu:
>> Em 03/07/2022 15:49, Tobias Platen via Libre-soc-dev escreveu:
>>> orangecrab WERROR: Failed to find a route for arc 2 of net
>> Maybe some attribute is missing, on the ls2 clock generator, or the
>> PLL. I'll compare with LiteX / LiteDRAM, and see if I can spot the
> The LiteX target file for the OrangeCrab has an "ECLKBRIDGECS":
> The LiteX target file for the Versa doesn't have it.
> Neither does our ecp5_crg.py, on ls2 (for any target).
> From the "ECP5 and ECP5-5G High-Speed I/O Interface" manual:
> "Edge Clocks (ECLK) are high-speed, low-skew I/O dedicated clocks. They
> are arranged in groups of two per I/O bank on the left and right sides
> of the device. Each of these Edge Clocks can be used to implement a
> high-speed interface. There is an Edge Clock Bridge (ECLKBRIDGECS) that
> allows you to build large interfaces by bridging the Edge Clocks from
> one bank to the other on the same side or from one side to the other side."
> I guess the Versa has all DDR3 pins on just one bank, while the
> OrangeCrab has them on different banks, thus needing an ECLKBRIDGECS.
Acrtic Tern uses the same overall design, since it talks to two different DRAM devices per FPGA (wide data bus):
More information about the Libre-soc-dev