Optimize usage of secp256k1

A comment in secp256k1/secp256k1.h says:

/** Opaque data structure that holds context information (precomputed tables etc.).
 *
 *  The purpose of context structures is to cache large precomputed data tables
 *  that are expensive to construct, and also to maintain the randomization data
 *  for blinding.
 *
 *  Do not create a new context object for each operation, as construction is
 *  far slower than all other API calls (~100 times slower than an ECDSA
 *  verification).
 *
 *  A constructed context can safely be used from multiple threads
 *  simultaneously, but API call that take a non-const pointer to a context
 *  need exclusive access to it. In particular this is the case for
 *  secp256k1_context_destroy and secp256k1_context_randomize.
 *
 *  Regarding randomization, either do it once at creation time (in which case
 *  you do not need any locking for the other calls), or use a read-write lock.
 */
typedef struct secp256k1_context_struct secp256k1_context;

In bitshares_fmt_plug.c, electrum_fmt_plug.c, opencl_electrum_modern_fmt_plug.c, we currently recreate the context for each use. We should optimize these formats to use either one shared context structure, or per-thread ones, but not recreated/destroyed each time.

Read the context lifecycle in bitshares_fmt_plug.c, electrum_fmt_plug.c, and opencl_electrum_modern_fmt_plug.c, along with the context guidance in secp256k1/secp256k1.h. Trace where contexts are created, used, and destroyed, then verify that reuse is implemented without violating the documented thread-safety requirements.