opus - Feb 2017 - [PATCH] Add silk/tests/test_unit_optimization_LPC_inv_pred_gain

Hi,

Attached is a patch
with silk/tests/test_unit_optimization_LPC_inv_pred_gain which does the
unit test of silk_LPC_inverse_pred_gain() optimizations. Please review.

The testing loop number is set to 10,000, since all branches in this
function get hit after 9,085 loops of random inputs.

Thanks,
Linfeng
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/opus/attachments/20170214/df8d2209/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Add-silk-tests-test_unit_optimization_LPC_inv_pred_g.patch
Type: text/x-patch
Size: 7885 bytes
Desc: not available
URL:
<http://lists.xiph.org/pipermail/opus/attachments/20170214/df8d2209/attachment.bin>

Hi Linfeng,

Thanks for the patch. Turns out I thought of a way to test not just the
optimized version, but also the C version at the same time. I just
checked in a modified version of your patch. It basically computes the
impulse response of the filter and checks if it explodes. If so, then it
verifies that LPC_inv_pred_gain() indeed detects the unstable filter.
Also, I removed the explicit comparison between the C version and the
optimized version, since --enable-check-asm will automatically do that.
Can you update and resend the LPC_inv_pred_gain() optimization patch?

Cheers,

	Jean-Marc

On 14/02/17 03:20 PM, Linfeng Zhang wrote:
> Hi,
> 
> Attached is a patch
> with silk/tests/test_unit_optimization_LPC_inv_pred_gain which does the
> unit test of silk_LPC_inverse_pred_gain() optimizations. Please review. 
> 
> The testing loop number is set to 10,000, since all branches in this
> function get hit after 9,085 loops of random inputs.
> 
> Thanks,
> Linfeng
> 
> 
> _______________________________________________
> opus mailing list
> opus at xiph.org
> http://lists.xiph.org/mailman/listinfo/opus
>