What about int types? I hear they're very slow on GPUs.
I'm asking because I have many parts which look like this:
mask = tensor1 <= tensor2
tensor3 *= mask
I never use masks for extracting subtensors directly. I just use them to
zero out elements as in the example above.
Can you think of a faster way to do that?
On Monday, February 6, 2017 at 6:02:36 PM UTC+1, nouiz wrote:
>
> The new back-end support float32 and float64 (as uint* int* too).
>
> But gamers GPUs have very slow float64. It will work, but you won't have
> any speed up from the GPU unless you use float32.
>
> Fred
>
> On Mon, Feb 6, 2017 at 7:20 AM, Kiuhnm Mnhuik <[email protected]
> <javascript:>> wrote:
>
>> I'm using Theano 0.9.0b1 with the new back-end.
>> Should I use float32 for everything (even for bool masks) for maximum
>> speed on GPU (GTX 970)?
>>
>> --
>>
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "theano-users" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected] <javascript:>.
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>
--
---
You received this message because you are subscribed to the Google Groups
"theano-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
