Update README.md - added pfm instructions

README.md

@@ -151,7 +151,7 @@ For the final deadline, submit your repository to the “Path (Final)” assignm
 To set command line arguments in QT, go to the Projects tab on the left, under build and run, under your currently configured kit, click run, and you should have a screen where you can enter command line arguments to run the executable with.
 - There are a lot of different probability calculations that go into computing the contribution of a ray to the final image. Make sure you really understand all of this math before you start trying to implement anything. You don’t want to get into the situation where your code is producing images that don’t quite look right, and all you can do is resort to aimlessly tweaking parts of the code (e.g. fiddling with constants, flipping minus signs) to try and make them look right.
 - Don’t try to implement all the required features at once. Implement them one by one, thoroughly debugging as you go.
-- It's useful to be able to look at your rendered output images _before_ you've implemented a tone mapping function. To do this, you can write your pathtracer's raw per-pixel radiance values to a high-dynamic range image format. The simplest such format is [PFM (short for Portable Float Map)](https://www.pauldebevec.com/Research/HDR/PFM/). You can also try using a library such as [tinyexr](https://github.com/syoyo/tinyexr) to read/write the industry-standard OpenEXR file format. Many different image viewing/editing apps support these formats.
+- It's useful to be able to look at your rendered output images _before_ you've implemented a tone mapping function. To do this, you can write your pathtracer's raw per-pixel radiance values to a high-dynamic range image format. The simplest such format is [PFM (short for Portable Float Map)](https://www.pauldebevec.com/Research/HDR/PFM/). We've provided a helper function called `outputPFM` in CS123Common.h that writes raw radiance values to a PFM file. You can use a PFM viewer such as [PfmPad](https://sourceforge.net/projects/pfmpad/) to see the output files.
 - Path tracers can take a long time to produce an image. The first thing you should do is **make sure to compile your code in "Release" mode** (or with all compiler optimizations enabled, if you’re not using Qt Creator). To speed up your code-test-iterate cycle, you’ll want to render low-resolution images (or, small windows cut out of a larger image that focus on some region of interest). In addition, you can also parallelize your code. Using [OpenMP’s parallel for loop](http://supercomputingblog.com/openmp/tutorial-parallel-for-loops-with-openmp/) is a good option, and there is a commented out line in the starter code (pathtracer.cpp, line 21) that illustrates how to do so. **Strongly recommended!**
 - When you allocate memory, make sure to zero it out. Bugs can creep into your code from uninitialized values.
 - If you’re noticing ‘speckling’ artifacts, i.e. individual isolated pixels that look incorrect: try using an image editor to identify the coordinates of a problematic pixel (many image editors will display this information somewhere as you mouse over the image). Then, set a breakpoint in your code that only fires on that pixel, and use a debugger to step through and see what is going wrong.