get out of this range, from the mean vector

[buaapengbo]

supra/src/SupraLib/Beamformer/RxBeamformerMVpcg.cu

Line 225 in 70bc6d9

if (subtractMeans)

I think this line need change to:
if ((subtractMeans) && ((offset + colIdx) < subArraySizeLocal))

if not, index offset + colIdx will out of subArraySizeLocal, and will get the next line's mean value

[buaapengbo]

supra/src/SupraLib/Beamformer/RxBeamformerMVpcg.cu

Lines 225 to 229 in 70bc6d9

	if (subtractMeans)
	{
	xCol -= means[offset + colIdx + linearSampleIdx * subArraySize];
	xRow -= means[offset + rowIdx + linearSampleIdx * subArraySize];
	}

can this change to, not sure

if (subtractMeans)
{
    if ((offset + colIdx) < subArraySizeLocal)
        xCol -= means[offset + colIdx + linearSampleIdx * subArraySize];
    if ((offset + rowIdx) < subArraySizeLocal)
        xRow -= means[offset + rowIdx + linearSampleIdx * subArraySize];
}

[goeblr]

Thanks for the feedback!

With how the means are computed

supra/src/SupraLib/Beamformer/RxBeamformerMVpcg.cu

Line 181 in 70bc6d9

means[vectIdx + linearSampleIdx * subArraySize] = mean / static_cast<float>(num);

for each sample it computes the mean signal in the sliding subaperture. So it is of length subArraySize at max (per sample). (specifically of subArraySizeLocal, but the stride is subArraySize).

So the lines you mention are wrong indeed.
With above, it should be

    if (subtractMeans) 
    { 
        xCol -= means[colIdx + linearSampleIdx * subArraySize]; 
        xRow -= means[rowIdx + linearSampleIdx * subArraySize]; 
    }