[bug] Fix SparseMatrix's dtype; check for dtype in SparseSolver.

[houkensjtu]

Issue: #8045

Brief Summary

SparseMatrix's dtype should be determined by the dtype of the SparseMatrixBuilder from which it's built. However, it's fixed to f32 in the current code base. This PR fix this by passing self.dtype instead of dtype to SparseMatrix() in the builder:

def build(self, dtype=f32, _format="CSR"):                                                                                                                                                                                          
        """Create a sparse matrix using the triplets"""                                                                                                                                                                                 
        taichi_arch = get_runtime().prog.config().arch                                                                                                                                                                                  
        if taichi_arch in [_ti_core.Arch.x64, _ti_core.Arch.arm64]:                                                                                                                                                                     
            sm = self.ptr.build()                                                                                                                                                                                                       
            return SparseMatrix(sm=sm, dtype=self.dtype)   # Previously it was dtype, which is f32 in the current context.

Also, SparseSolver should raise an exception to better notify the user if the dtype of the SparseMatrix is not consistent with the solver's dtype. This is implemented in the sparse_solver.py.

Additional comments

• This PR should resolve the user's question posted on the forum: https://forum.taichi-lang.cn/t/topic/4316
• This PR might fix part of the CI test failures mentioned in issue [Test] SparseSolver test failure on CUDA backend #8045 , but [Test] SparseSolver test failure on CUDA backend #8045 still needs more thorough investigation even with this PR merged to ensure that there is no future issues.

[netlify]

✅ Deploy Preview for docsite-preview ready!

Name	Link
🔨 Latest commit	20c6ee9
🔍 Latest deploy log	https://app.netlify.com/sites/docsite-preview/deploys/646f11140a4b690008b9e494
😎 Deploy Preview	https://deploy-preview-8071--docsite-preview.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site settings.

[houkensjtu]

Please review this PR when you have the time @FantasyVR . Thanks :)

[FantasyVR]

Thanks, LGTM.

[neozhaoliang]

LGTM!