Will SLI/Crossfire bypass VRAM limitations?

Question

Would running 2 (2GB) AMD R9 270X cards in Crossfire, bypass the DAG limitations on VRAM size?

I made the mistake of thinking my single 2GB R9 270X running alongside my 6GB GTX 1060 would work out. However the DAG has exceed the 2GB limit as of this post and now I can no longer run both.

See output from Claymore's:

ETH: Authorized
Setting DAG epoch #133...
Setting DAG epoch #133 for GPU0
Create GPU buffer for GPU0
Setting DAG epoch #133 for GPU1
GPU0 - not enough GPU memory to place DAG, you cannot mine this coin with this GPU
GPU0 - OpenCL error -61 - cannot allocate big buffer for DAG. Check readme.txt for possible solutions.
Create GPU buffer for GPU1
GPU 0 failed
Setting DAG epoch #133 for GPU0
GPU0, OpenCL error -38 - cannot write buffer for DAG
GPU 0 failed
GPU1 DAG creation time - 8587 ms
Setting DAG epoch #133 for GPU1 done
GPU0 t=35C fan=20%; GPU1 t=58C fan=29%
WATCHDOG: GPU error, you need to restart miner :(
Restarting OK, exit...

I have read Can you mine with 2GB cards or less? Thinking about the DAG issues already and I understand combining seperate cards in one mining terminal would not work, however my inquiry is more so about combining them on the physical level with SLI/Crossfire.

Answer 1

Practically speaking, no.

As I understand it, SLI and Crossfire implementations were designed to mirror memory contents (think textures or 3d models) such that each GPU could access the same contents locally (through its own memory controller). By doing this, computations (shaders, etc.) can be striped between GPUs with relatively few performance impacts.

Even if you could pool the memory together, access to memory that isn't directly wired to a GPU's memory controller would be very slow (relatively speaking), as it would need to go through some other slower bus (the SLI/Crossfire interface) and would also need to synchronize with the other GPU's memory controller.