Hi everyone, I have a simple doubt. While running performance.cpp to get the latency results, why doesnot the relinearization happen for N=1024 and 2048. The variable context.use_keyswitching() is zero whenever N size is smaller. For N sizes greater than 2048, i.e. 4096 and 8192 .. We use relinearization. Can somebody explain this ?
Hi everyone, I have a simple doubt. While running performance.cpp to get the latency results, why doesnot the relinearization happen for N=1024 and 2048. The variable context.use_keyswitching() is zero whenever N size is smaller. For N sizes greater than 2048, i.e. 4096 and 8192 .. We use relinearization. Can somebody explain this ?