this is what I get on my 660Ti under Linux.
buchner@test02:~/CudaMiner-master$ ./cudaminer --algo=keccak -L 64 -l K1024x32 --benchmark
*** CudaMiner for nVidia GPUs by Christian Buchner ***
This is version 2014-02-04 (beta)
based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
Cuda additions Copyright 2013,2014 Christian Buchner
LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm
BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM
YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4
[2014-02-06 15:13:53] 1 miner threads started, using 'keccak' algorithm.
[2014-02-06 15:14:09] GPU #0: GeForce GTX 660 Ti with compute capability 3.0
[2014-02-06 15:14:09] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
[2014-02-06 15:14:09] GPU #0: 32 hashes 0.1 MB per warp.
[2014-02-06 15:14:09] GPU #0: using launch configuration K1024x32
[2014-02-06 15:14:13] GPU #0: GeForce GTX 660 Ti, 98313 khash/s
[2014-02-06 15:14:13] Total: 98313 khash/s
[2014-02-06 15:14:18] GPU #0: GeForce GTX 660 Ti, 98940 khash/s
[2014-02-06 15:14:18] Total: 98940 khash/s