Updated TH to ATen porting guide (markdown)

TH-to-ATen-porting-guide.md

@@ -42,32 +42,6 @@ In code in TH, you will see references to scalar_t. In fact, all code in TH is i
 
 This is obviously horrible, and in ATen we do things more normally using templates. However, this doesn't mean you should go out and template everything. Instead, there is usually a critical loop inside the kernel which actually needs to be templated, and then everything else can be compiled once generically. At the point this occurs, you should use `AT_DISPATCH_FLOATING_TYPES_AND_HALF` to perform this dispatch from code which doesn't at compile time what scalar_t is, to code that knows what scalar_t. For usage guidance, search for some examples.
 
-### Dispatch and OpenMP
-
-Many functions in THNN are parallelized in the following way:
-
-```
-    int64_t batch;
-#pragma omp parallel for private(batch)
-    for (batch = 0; batch < numBatch; ++batch) {
-      THNN_(SpatialFractionalMaxPooling_updateOutput_frame)(
-        input->data<scalar_t>() + batch * numPlanes * inputH * inputW,
-        output->data<scalar_t>() + batch * numPlanes * outputH * outputW,
-        THIndexTensor_(data)(indices) + batch * numPlanes * outputH * outputW,
-        randomSamples->data<scalar_t>() + batch * numPlanes * 2,
-        numPlanes, inputW, inputH, outputW, outputH, poolSizeW, poolSizeH);
-    }
-  }
-```
-
-This snippet of code is particularly delicate to port, for a few reasons:
-
-1. You need to do a dispatch before you can retrieve the appropriately typed scalar_t pointers, implying use of the AT_DISPATCH_ macro. However...
-2. #pragma omp declarations don't work inside lambdas...
-3. You do NOT want to do a dynamic dispatch every time around the batch loop; you should do it once to get into the tight loop.
-
-The recommendation is to make a new helper function to contain the OpenMP pragma, and call it from inside the AT_DISPATCH_ macro. Do the extraction of the pointers once, outside of the loop (since they're same every time around). Also, separate code paths for batch==1 and batch>1, because parallelism across 1 batch in the former case is redundant and affects performance.
-
 ## C++ torch API guidance
 
 For the most part, the at::Tensor API matches the Python Torch tensor API directly; so you can check the official docs to find out what is callable: https://pytorch.org/docs/stable/tensors.html