cudpp_hash_testrig crashes on Win7 x64, CUDA 4.0, GTX 480

[GoogleCodeExporter]

What steps will reproduce the problem?
run cudpp_hash_testrig -all

Output:

Using device 0:
GeForce GTX 480; global mem: 1543176192B; compute v2.0; clock: 1401000 kHz
Basic hash table: Iteration 0
        Space usage: 1.050000
Table size: 1050000 slots
Maximum iterations: 2816 (141.283414 lg (N))
Number of functions: 4
Hash function constants:
        3499211612,  581869302
        3890346734, 3586334585
         545404204, 4161255391
        3922919429,  949333985
Num items hashing into a slot:
        (0, 23446)
        (1, 88624)
        (2, 168999)
        (3, 214332)
        (4, 203289)
        (5, 156025)
        (6, 98636)
        (7, 54106)
        (8, 25433)
        (9, 10856)
        (10, 4230)
        (11, 1442)
        (12, 410)
        (13, 130)
        (14, 30)
        (15, 9)
        (16, 3)
Slots assigned to each key: (1, 0) (2, 0) (3, 0) (4, 1000000)
Cuda error: Error occurred during hash table clear.
 in file 'c:/src/cudpp/trunk/src/cudpp_hash/hash_table.cu' in line 179 : invalid
 device pointer.
Press any key to continue . . .

Original issue reported on code.google.com by harr...@gmail.com on 28 Jul 2011 at 2:51

[GoogleCodeExporter]

Verified fixed.

Original comment by harr...@gmail.com on 5 Aug 2011 at 4:56

• Changed state: Verified

[GoogleCodeExporter]

hey could u post how the error was fixed ?

Original comment by joydsouz...@gmail.com on 20 Mar 2012 at 6:08

[GoogleCodeExporter]

I don't recall exactly, but you can go look through the checkins into the hash 
library between Jul 27 and Aug 4 last year. It was almost certainly a fix we 
made in the hash library.

Original comment by jow...@gmail.com on 20 Mar 2012 at 6:10