perf: increase cache sizes from 256 to 1024 for ~1000 followed pubkeys

With 256 slots and ~1000 followed users, the direct-mapped caches had
~75% collision rate (multiple pubkeys mapping to the same slot, causing
constant eviction and recomputation). Increasing to 1024 slots covers
most follow lists with minimal collisions.

Memory impact:
  pubkeyCache: 1024 × ~96 bytes = ~96KB (was ~24KB)
  pTableCache: 1024 × ~1KB = ~1MB (was ~256KB)
Total ~1.1MB — acceptable for mobile.

https://claude.ai/code/session_017UbWduFi1sLUsgVUMUH2nx

quartz/src/commonMain/kotlin/com/vitorpamplona/quartz/utils/secp256k1/ECPoint.kt

@@ -105,7 +105,9 @@ internal object ECPoint {
     //
     // 256 entries × 16 AffinePoints × 64 bytes = ~256KB total cache.
 
-    private const val P_TABLE_CACHE_SIZE = 256
+    // 1024 entries to cover ~1000 followed pubkeys with minimal collisions.
+    // Memory: 1024 × 16 AffinePoints × 64 bytes = ~1MB. Acceptable for mobile.
+    private const val P_TABLE_CACHE_SIZE = 1024
     private const val P_TABLE_CACHE_MASK = P_TABLE_CACHE_SIZE - 1
 
     private class CachedPTable(

quartz/src/commonMain/kotlin/com/vitorpamplona/quartz/utils/secp256k1/Secp256k1.kt

@@ -85,8 +85,9 @@ object Secp256k1 {
     // saving the sqrt for repeated pubkeys (~13% of verify cost per cache hit).
     //
     // Simple fixed-size direct-mapped cache (no LRU overhead). Size must be power of 2.
-    // 256 entries × (32 + 32 + 32) bytes = ~24KB. Cache collisions just evict silently.
-    private const val PUBKEY_CACHE_SIZE = 256 // power of 2
+    // 1024 entries covers most follow lists (~1000 users) with few collisions.
+    // Memory: 1024 × ~96 bytes = ~96KB.
+    private const val PUBKEY_CACHE_SIZE = 1024 // power of 2
     private const val PUBKEY_CACHE_MASK = PUBKEY_CACHE_SIZE - 1
 
     private class CachedPubkey(