Re: [FE] Frame size poll
Hugh-
Be careful!
It looks suspiciously like you are sending messages to the world saying
that you agree with me.
Geoff
At 09:25 AM 9/3/2004 -0700, Hugh Barrass wrote:
>Kevin,
>
>As a hardware designer, my first pick was 2) however I have designed a
>number of switch fabrics in various organizations and I have also looked
>closely at many others and the 2048 number may pose a problem for some
>equipment or architectures. Many shared memory architectures use tags
>that are stored along with the frames therefore a frame size of 2048
>would mean that the storage quanta would be just over 2k - a "bad"
>number. Given that this may be an "architectural" feature rather than an
>implementation, it would act as an impediment to rapid development of
>compliant solutions using those architectures.
>
>As a result, I would support 1) or 3) with any number up to 2000 (an
>arbitrary value, less than 2k). Note, however, that I haven't consulted
>any colleagues on this so I cannot say that this response would be
>favorable with others in my team.
>
>Hugh.
>
>Kevin Daines wrote:
>
>>All,
>>
>>
>>Now to the poll. Would you prefer:
>>
>>1) 1875 (which is based on the repeater calculation)
>>2) 2048 (a power of 2)
>>3) some other number (please specify)
>>
>>