What about using unique and a brute force approach:
m = randi(10, 1, 100) % Sample data.
allValues = unique(m); % Find unique numbers.
for k = 1 : length(allValues)
thisValue = allValues(k);
indexes{k} = find(m == thisValue);
fprintf('The number %f is found at indexes : ', thisValue);
fprintf('%d ', indexes{k});
fprintf('\n');
end
% celldisp(indexes)
