[Discussion] [Oisf-devel] Suricata 1.0.0 released

[Martin Holste]

Understood.  I just bought a new gaming rig at home, so now I have a
spare box with a GeForce 9800 GTX with 128 stream processors just
laying around.  I'll see if I can get it hooked up for testing and
I'll let you know how it goes.

On Fri, Jul 2, 2010 at 10:25 AM, Will Metcalf <william.metcalf at gmail.com> wrote:
>> Congrats!
>
> Thanks Martin!
>
>> Can you briefly describe the current state of the CUDA support?  Is it
>> refined enough that performance is better when using CUDA?  I believe
>> that in the early stages of development it was stated that non-CUDA
>> was still faster than using CUDA.  Also, has anyone tested it using
>> SLI?  This is really exciting stuff as the count of on-board stream
>> processors continues to grow by leaps and bounds on these video cards
>> every month!
>
> To be honest we haven't done much more performance testing, but the
> packet batching code is now there i.e. sending multiple packets at
> once to the graphics card.  If you have the cycles take it for a spin
> and let us know ;-)..  We will be doing more testing/refining of this
> code but as of now it is still experimental.  Just to rehash to build
> with cuda support you simply...
>
> ./configure --enable-cuda
>
> edit the suricata.yaml file and set
>
> mpm-algo: b2g
>
> to
>
> mpm-algo: b2g_cuda
>
> and
>
> #max-pending-packets: 50
>
> to something like
>
> max-pending-packets: 7500
>
> This will cost you some ram btw ;-)
>
> Regards,
>
> Will
>