On 3/16/2010 4:23 PM, Roland Rambau wrote:
Eric,
careful:
Am 16.03.2010 23:45, schrieb Erik Trimble:
Up until 5 years ago (or so), GigaByte meant a power of 2 to EVERYONE,
not just us techies. I would hardly call 40+ years of using the various
giga/mega/kilo prefixes as a power of 2 in computer science as
non-authoritative.
How long does it take to transmit 1 TiB over a 1 GB/sec tranmission
link, assuming no overhead ?
See ?
hth
-- Roland
I guess folks have gotten lazy all over.
Actually, for networking, it's all "GigaBIT", but I get your meaning.
Which is why it's all properly labeled "1Gb" Ethernet, not "1GB" ethernet.
That said, I'm still under the impression that Giga = 1024^3 for
networking, just like Mega = 1024^2. After all, it's 100Mbit Ethernet,
which doesn't mean it runs at 100Mhz.
That is, on Fast Ethernet, I should be sending a max 100 x 1024^2 BITS
per second.
Data amounts are (so far as I know universally) employing powers-of-2,
while frequencies are done in powers-of-10. Thus, baud (for modems) is
in powers-of-10, as are CPU/memory speeds. Memory (*RAM of all sorts),
bus THROUGHPUT (i.e. PCI-E is in powers-of-2), networking throughput,
and even graphics throughput is in powers-of-2.
If they want to use powers-of-10, then use the actual "normal" names,
like graphics performance ratings have done (i.e. 10 billion texels, not
"10 Gigatexels". Take a look at Nvidia's product literature:
http://www.nvidia.com/object/IO_11761.html
It's just the storage vendors using the broken measurements. Bastards!
--
Erik Trimble
Java System Support
Mailstop: usca22-123
Phone: x17195
Santa Clara, CA
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss