As far as I know, all Rocket Chip configs should work in Verilator. We
actually use Verilator for CI builds of Rocket Chip on Travis, since we can't
run propritary tools there:
https://travis-ci.org/ucb-bar/rocket-chip
Unless this has changed, all Rocket Chip commits need to pass all the tests on
Travis (using Verilator as a simulator) before they can be merged to master.
You can see the exact set of configs here:
https://github.com/ucb-bar/rocket-chip/blob/master/regression/Makefile#L43
DefaultConfig and TinyConfig are both on that list, so they're both regularly
tested on Verilator.
On Fri, 07 Oct 2016 09:50:35 PDT (-0700),
nhon....@abeesemi.net wrote:
> Palmer,
>
> Did you mean you have better luck using Verilator for TinyConfig or for the
> default config?
>
> Thanks,
>
> Nhon
>
> On Saturday, October 1, 2016 at 10:30:51 AM UTC-7, palmer.dabbelt wrote:
>>
>> I tried earlier this week, but it looks like iverilog doesn't have support
>> for SystemVerilog DPI, which the Rocket Chip test harness is written in.
>> Here's a vague bug report
>>
>>
https://github.com/steveicarus/iverilog/issues/100
>>
>> which convinced me to stop trying to make it work. Rocket Chip supports
>> Verilator out of the box, use the emulator directory. Here's the Makefile
>>
>>
https://github.com/ucb-bar/rocket-chip/blob/master/emulator/Makefile
>>
>> I've had good luck with Verilator for RTL simulation.
>>
>> On Sat, Oct 1, 2016 at 9:54 AM, Nhon <
nhon....@abeesemi.net <javascript:>>
>> wrote:
>>
>>> Hi, all,
>>>
>>> Has anyone tried using iverilog instead of VCS for vsim? Simulation
>>> quality wise, is iverilog good enough for actual ASIC chip? Any idea how
>>> fast or slow iverilog is vs. VCS?
>>>
>>> I vaguely remember some project (lowrisc?) has used this, but not sure.
>>> If so, can someone share the Makefile? I am interested in learning how the
>>> various simulation options are replaced by iverilog's.
>>>
>>> If there is a better open source verilog, please let me know as well.
>>>
>>> Thanks!
>>>
>>> Nhon
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "RISC-V HW Dev" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an
>>> email to
hw-dev+un...@groups.riscv.org <javascript:>.
>>> <javascript:>.
>>> <
https://groups.google.com/a/groups.riscv.org/d/msgid/hw-dev/51af9ab7-2590-42e7-9fa6-c064b516fe04%40groups.riscv.org?utm_medium=email&utm_source=footer>
>>> .
>>>
>>
>>