On Tue, Nov 22, 2011 at 1:16 PM, Simon Quellen Field <sfield@scitoys.com> wrote:
> Before you talk about gigabit Ethernet on a cost reduced platform, ask
> yourselves
> if the processor can actually deliver even 100 megabits per second. Most
It probably can do megabit just fine, and since the processor is
32-bit it can probably do gigE just fine (32 bits of data
transferred/manipulated per instruction), DMA data transfer might get
into the mix to lighten the processor load too.
> disk
> drives on the target machine can only save data at perhaps 30 megabytes per
> second.
In this day and age 30 MB/s write speed is quite slow, SSDs can get
pretty dang fast, into the 100s of MB/s (this depends on file size,
because transferring a bunch of small files usually doesn't saturate
the bus. You have to start using Jumbo Frames to really use gigE,
which again just increases the size of each packet of info so the
overhead decreases for the total data.
> That is only three times better than the 100 MB/s Ethernet can deliver.
> How much extra cost is it worth to get a three-fold improvement in transfer
> times?
> How much data are you planning to transmit?
You've got a lot of data coming through gigE, I've only used it in a
professional setting where I was streaming high-res RAW camera data at
up to around the maximum throughput of gigE. This was all being saved
to a SATA (SATA II maybe) spinning hard disk through writing
individual images. We thought that writing a single file to save on
open close time, but I left the project just after that.
> How important is it to be able to transmit a file in 0.4 milliseconds
> instead of 4
> milliseconds?
I've calculated with this openSpectrometer sensor the max throughput I
would ever want would be about 4Mbps. If it was a fast processor and I
was doing video, I think compression (even lossless) could be
implemented.
--
Nathan McCorkle
Rochester Institute of Technology
College of Science, Biotechnology/Bioinformatics
--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diybio@googlegroups.com.
To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/diybio?hl=en.
0 comments:
Post a Comment