You asked this 8 days before NASA published a description of on how they're testing the heating: http://www.nasa.gov/sls/base-heating-tests.html
Apparently the answer is to build a 2% scale model of the entire rocket including engines and boosters and fire it in short duration bursts of 50-150 milliseconds at a time in a shock tunnel (CUBRC is working with NASA, so probably these tunnels) that's producing flight conditions for various stages of the flight.
(Image from NASA in the article linked above)
The short duration of testing is is mandated by what the model can handle without melting and very short operation time of a shock tunnel (Wikipedia quotes 20 milliseconds, though that isn't for a CUBRC tunnel).
Testing of the mini models will provide data on the convective heating environments the base of the rocket will experience upon ascent for both planned and unplanned flight conditions. Data from the test series will be used to verify flight hardware design environments and set specifications for the design of the rocket's base thermal protection system. The system keeps major hardware, wiring and crews on future missions safe from the extreme heat the boosters and engines create while burning during ascent.
Note that this isn't testing the actual shielding, it's just providing the data required to determine how much shielding is required. Given knowledge of the heating it has to withstand the shielding can be tested separately to prove it handles the heat load before stacking all the rockets together. I wouldn't be surprised if the first time they test the shielding with an actual SRB on one side and a block of SSMEs on the other is the first time they launch it.
Experimental data is collected by:
The full-stack configuration has 200 heat flux and pressure sensors within the aft section of the rocket to collect data on the base flow environment.
and
The test program uses new technology that wasn't available during past human spaceflight programs, such as high-speed visible light and infrared cameras, laser diagnostics and new designs of the model propulsion systems that more adequately simulate the SLS full-scale systems.
The first article I linked mentioned that the testing should be done by early summer of 2015. I haven't found any recent references to the testing, but I did find this older NASA post about the scale rockets before they were attached to a model of the rocket for shock tunnel testing. That included a video of a ~65ms long hot-fire test of the core:
No video of firing with the boosters attached but there was this still image (with an old configuration where the core was turned 45 degrees with respect to the boosters):