Changeset c036ddb in sasmodels for sasmodels/kernelcl.py

Timestamp:

Aug 7, 2018 10:45:45 PM (6 years ago)

Author:

Paul Kienzle <pkienzle@…>

Branches:

master, core_shell_microgels, magnetic_model, ticket-1257-vesicle-product, ticket_1156, ticket_1265_superball, ticket_822_more_unit_tests

Children:

7e923c2

Parents:

7b0abf8

Message:

refactor so Iq is not needed if Fq is defined

File:

• sasmodels/kernelcl.py (modified) (6 diffs)

sasmodels/kernelcl.py

@@ -420 +420 @@
         if self.program is None:
             compile_program = environment().compile_program
-            with open('model.c','w') as fid:
-                print(self.source['opencl'], file=fid)
             timestamp = generate.ocl_timestamp(self.info)
             self.program = compile_program(
@@ -540 +538 @@
         self.dtype = dtype
         self.dim = '2d' if q_input.is_2d else '1d'
-        # plus three for the normalization values
-        self.result = np.empty(2*q_input.nq+2,dtype)
+        # leave room for f1/f2 results in case we need to compute beta for 1d models
+        num_returns = 1 if self.dim == '2d' else 2  #
+        # plus 1 for the normalization value
+        self.result = np.empty((q_input.nq+1)*num_returns, dtype)
 
         # Inputs and outputs for each kernel call
@@ -549 +549 @@
 
         self.result_b = cl.Buffer(self.queue.context, mf.READ_WRITE,
-                                  q_input.global_size[0] * dtype.itemsize)
+                                  q_input.global_size[0] * num_returns * dtype.itemsize)
         self.q_input = q_input # allocated by GpuInput above
 
@@ -557 +557 @@
                      else np.float16 if dtype == generate.F16
                      else np.float32)  # will never get here, so use np.float32
-    __call__= Iq
 
     def Iq(self, call_details, values, cutoff, magnetic):
+        # type: (CallDetails, np.ndarray, np.ndarray, float, bool) -> np.ndarray
+        self._call_kernel(call_details, values, cutoff, magnetic)
+        #print("returned",self.q_input.q, self.result)
+        pd_norm = self.result[self.q_input.nq]
+        scale = values[0]/(pd_norm if pd_norm != 0.0 else 1.0)
+        background = values[1]
+        #print("scale",scale,background)
+        return scale*self.result[:self.q_input.nq] + background
+    __call__ = Iq
+
+    def beta(self, call_details, values, cutoff, magnetic):
+        # type: (CallDetails, np.ndarray, np.ndarray, float, bool) -> np.ndarray
+        if self.dim == '2d':
+            raise NotImplementedError("beta not yet supported for 2D")
+        self._call_kernel(call_details, values, cutoff, magnetic)
+        w_norm = self.result[2*self.q_input.nq + 1]
+        pd_norm = self.result[self.q_input.nq]
+        if w_norm == 0.:
+            w_norm = 1.
+        F2 = self.result[:self.q_input.nq]/w_norm
+        F1 = self.result[self.q_input.nq+1:2*self.q_input.nq+1]/w_norm
+        volume_avg = pd_norm/w_norm
+        return F1, F2, volume_avg
+
+    def _call_kernel(self, call_details, values, cutoff, magnetic):
         # type: (CallDetails, np.ndarray, np.ndarray, float, bool) -> np.ndarray
         context = self.queue.context
@@ -575 +599 @@
         ]
         #print("Calling OpenCL")
-        call_details.show(values)
+        #call_details.show(values)
         #Call kernel and retrieve results
         wait_for = None
@@ -601 +625 @@
                 v.release()
 
-        pd_norm = self.result[self.q_input.nq]
-        scale = values[0]/(pd_norm if pd_norm != 0.0 else 1.0)
-        background = values[1]
-        #print("scale",scale,values[0],self.result[self.q_input.nq],background)
-        return scale*self.result[:self.q_input.nq] + background
-        # return self.result[:self.q_input.nq]
-     #NEEDS TO BE FINISHED FOR OPENCL
-     def beta():
-         return 0
-
     def release(self):
         # type: () -> None