What’s the (remaining) value of cryptographic hashing of small values these days?

A friend forwarded this to me today – one more in a long narrative of the incredibly reduced value of hashing
to make it hard for anyone seeing the hash to determine the original data being
hashed (for small-sized inputs).  Hashing
a password/passphrase, hashing a “unique identifier” – these
approaches to obscuring (for lack of a better word) the password/unique ID seem
effectively moot to me.  I honestly don’t
know that there’s any real value in performing the hash and then storing or
exchanging it – frankly, the difference in your risk between
“sending/storing the password” and “sending/storing the hash of
the password” seems pretty small.
Even small inputs + small salts seem doomed, given these
massive advanced in FP calculation arrays. 
Further, it seems like “stored, single-value salts” are just
as pointless, given the amount of research that attackers generally put into
discovering these stored/fixed salt values – so storing a hugely long salt
value just feels wrong to me for many threat scenarios.
Do any of the SHA-3 algorithms take massively more time
on these hash-calculating clusters?  If
not, what other options can our products use for protecting small values like
passwords and unique identifiers?
I get that there are still some attack vectors against
which a hashed password is a useful mitigation vs. the raw password
itself.  It just seems like we’re getting
further and further away from the knee-jerk “hash it and you’re much better
off” that I was taught at the feet of my cryptographic elders.
Note: I believe there’s still value in using
modern/advanced hash functions to predict the integrity of a known piece of
information (digital signatures, message authentication):
  • e.g. hash a large document and sign it to later assert
    with some degree of confidence that the document hasn’t been tampered with
  • e.g. compare the previously-stored password hash to
    determine if the supplicant has possession of that password

“A presentation at the Passwords^12 Conference in
Oslo, Norway, has moved the goalposts on password cracking yet again.
Speaking on Monday, researcher Jeremi Gosney (a.k.a epixoip) demonstrated a rig
that leveraged the Open Computing Language (OpenCL) framework and a technology
known as Virtual Open Cluster (VCL) to run the HashCat password cracking
program across a cluster of five, 4U servers equipped with 25 AMD Radeon GPUs
communicating at 10 Gbps and 20 Gbps over Infiniband switched fabric. Gosney’s
system elevates password cracking to the next level, and effectively renders
even the strongest passwords protected with weaker encryption algorithms, like
Microsoft’s LM and NTLM, obsolete. In a test, the researcher’s system was able
to generate 348 billion NTLM password hash checks per second. That renders even
the most secure password vulnerable to compute-intensive brute force and
wordlist (or dictionary) attacks. A 14 character Windows XP password hashed
using LM for example, would fall in just six minutes, said Per Thorsheim,
organizer of the Passwords^12 Conference. For some context: In June,
Poul-Henning Kamp, creator of the md5crypt() function used by FreeBSD and
other, Linux-based operating systems, was forced to acknowledge that the
hashing function is no longer suitable for production use — a victim of GPU-powered
systems that could perform ‘close to 1 million checks per second on COTS
(commercial off the shelf) GPU hardware,’ he wrote. Gosney’s cluster cranks out
more than 77 million brute force attempts per second against MD5crypt.”
 URL – http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes