[lowrisc-dev] Re: Question on lowrisc project

Pawan Reddy Sibbala ps849 at cornell.edu
Mon Feb 22 05:23:17 GMT 2016


Hi Wei,

I wanted to find out about the IP to see if we can modify the RTL of the
memory controller.

On a different note, I was looking at the source code of the rocket system
to see which coherence protocol was implemented in it, but I see that many
coherence protocols are defined ( MSI , MESI , MOESI, directory based ) in
the source files. I couldn't figure out which protocol is being used. Could
you tell me what protocol is used for coherence?

Thanks and regards,
Pawan

On Sun, Feb 21, 2016 at 12:24 PM, Wei Song <ws327 at cam.ac.uk> wrote:

> Hello Pawan,
>
> It is the one you get from the Vivado IP library (mig_7series ver 2.4).
> I believe the memory controller is a soft one but the source is encrypted.
> To me, hard or soft does not make much difference.
>
> Best regards,
> Wei
>
>
> On 21/02/2016 15:43, Pawan Reddy Sibbala wrote:
>
> Hello Wei,
>
> In your last mail, you mentioned that you're using Xilinx IPs for memory
> controller. Is it Xilinx Hard IP or are you using soft IP?
>
> Thanks and regards,
> Pawan
>
> On Mon, Feb 15, 2016 at 11:26 AM, Pawan Reddy Sibbala <
> <ps849 at cornell.edu>ps849 at cornell.edu> wrote:
>
>> Thank you for the quick response, Wei.
>>
>> On Mon, Feb 15, 2016 at 11:02 AM, Wei Song < <ws327 at cam.ac.uk>
>> ws327 at cam.ac.uk> wrote:
>>
>>> Hello Pawan,
>>>
>>> On 15/02/2016 15:54, Pawan Reddy Sibbala wrote:
>>> > Hello Wei,
>>> >
>>> > I have two questions that I'd like to ask you.
>>> >
>>> > 1. The rocket core on the zedboard FPGA  currently runs at 25Hz. Could
>>> > you tell me the maximum frequency at which it can run ? I was
>>> > wondering if you already experimented with the parameters in
>>> > "clocking.vh" to find the maximum frequency.
>>>
>>> No, I did not. The maximal frequency is in between of 25 MHz and 50 MHz.
>>> For the same configuration and the same FPGA, running at 50 MHz has some
>>> timing violations.
>>> The critical paths are in FPUs. So suppose if the FPU is disabled, it
>>> may run at 50 MHz.
>>> As running on FPGA is not the target and the design is not optimized for
>>> FPGA, there should be space for speed optimization.
>>> Even without any optimization, I believe it can easily run to 30 or 35
>>> MHz, which I had not tried.
>>>
>>> >
>>> > 2. Does the new release of the untethered version of the rocketchip
>>> > use a opensource IP for implementing the Memory Controller or does it
>>> > use the Xilinx IP?
>>>
>>> Those are Xilinx IPs.
>>>
>>> >
>>> > Thanks and regards,
>>> > Pawan
>>> >
>>>
>>> Best regards,
>>> Wei
>>>
>>
>>
>>
>> --
>> Pawan Reddy Sibbala
>> M.Eng ECE
>> Cornell University
>>
>
>
>
> --
> Pawan Reddy Sibbala
> M.Eng ECE
> Cornell University
>
>
>


-- 
Pawan Reddy Sibbala
M.Eng ECE
Cornell University


More information about the lowrisc-dev mailing list