How to tell mfakto to use IGP?

[Aramis Wyler]

mfaktc v0.20 (32bit built)
Compiletime options
THREADS_PER_BLOCK 256
SIEVE_SIZE_LIMIT 32kiB
SIEVE_SIZE 193154bits
SIEVE_SPLIT 250
MORE_CLASSES enabled
Runtime options
SievePrimes 25000
SievePrimesAdjust 1
SievePrimesMin 5000
SievePrimesMax 100000
NumStreams 3
CPUStreams 3
GridSize 0
WorkFile worktodo.txt
Checkpoints enabled
CheckpointDelay 900s
Stages enabled
StopAfterFactor class
PrintMode compact
V5UserID (none)
ComputerID (none)
AllowSleep no
TimeStampInResults no
CUDA version info
binary compiled for CUDA 4.20
CUDA runtime version 4.20
CUDA driver version 5.0
CUDA device info
name nForce 980a/780a SLI
compute capability 1.1
maximum threads per block 512
number of multiprocessors 1 (8 shader cores)
clock rate 1200MHz
Automatic parameters
threads per grid 131072
running a simple selftest...
Selftest statistics
number of tests 55
successfull tests 55
selftest PASSED!
got assignment: exp=62704217 bit_min=71 bit_max=73 (22.88 GHz-days)
Starting trial factoring M62704217 from 2^71 to 2^72 (7.63 GHz-days)
k_min = 18827946143160
k_max = 37655892289266
Using GPU kernel "barrett76_mul32"
Date Time | class Pct | time ETA | GHz-d/day Sieve Wait
Mar 22 22:18 | 0 0.1% | 408.03 4d12h | 1.68 25000 96.55%

That's with 2 cores free. It completely capped out one of them and still claims a 96.55% wait. Ah well. I expect that even with some tuning it probably wouldn't perform better than the cpu core it is consuming.