Various improvements.

.gitignore

@@ -64,6 +64,7 @@ Programs/Public-Input/*
 *.d
 local/
 *.map
+*.gch
 
 # Packages #
 ############

BMR/Register.cpp

@@ -574,11 +574,11 @@ void EvalRegister::inputb(GC::Processor<GC::Secret<EvalRegister> >& processor,
  }
  for (auto& access : accesses)
  access.prepare_masks(my_os);
- party.P->Broadcast_Receive(oss, true);
+ party.P->unchecked_broadcast(oss);
  my_os.reset_write_head();
  for (auto& access : accesses)
  access.received_masks(oss);
- party.P->Broadcast_Receive(oss, true);
+ party.P->unchecked_broadcast(oss);
  for (auto& access : accesses)
  access.received_labels(oss);
 }
@@ -622,7 +622,7 @@ void EvalRegister::other_input(EvalInputter& inputter, int from)
 
 void EvalInputter::exchange()
 {
- party.P->Broadcast_Receive(oss, true);
+ party.P->unchecked_broadcast(oss);
  for (auto& tuple : tuples)
  {
  if (tuple.from != party.P->my_num())
@@ -646,7 +646,7 @@ void EvalInputter::exchange()
 #endif
  }
 
- party.P->Broadcast_Receive(oss, true);
+ party.P->unchecked_broadcast(oss);
 }
 
 void EvalRegister::finalize_input(EvalInputter& inputter, int, int)

BMR/Register.hpp

@@ -236,7 +236,7 @@ void EvalRegister::load(vector<GC::ReadAccess<T> >& accesses,
  }
  }
 
- party.P->Broadcast_Receive(keys, true);
+ party.P->unchecked_broadcast(keys);
 
  int base = 0;
  for (auto access : accesses)

CHANGELOG.md

@@ -1,5 +1,11 @@
 The changelog explains changes pulled through from the private development repository. Bug fixes and small enhancements are committed between releases and not documented here.
 
+## 0.2.1 (Dec 11, 2020)
+
+- Virtual machines automatically use the modulus used during compilation
+- Non-linear computation modulo a prime without large gap in bit length
+- Fewer communication rounds in several protocols
+
 ## 0.2.0 (Oct 28, 2020)
 
 - Rep4: honest-majority four-party computation with malicious security

CONFIG

@@ -40,9 +40,9 @@ endif
 
 # MAX_MOD_SZ (for FHE) must be least and GFP_MOD_SZ (for computation)
 # must be exactly ceil(len(p)/len(word)) for the relevant prime p
-# Default for GFP_MOD_SZ is 2, which is good for 128-bit p
+# GFP_MOD_SZ only needs to be set for primes of bit length more that 256.
 # Default for MAX_MOD_SZ is 10, which suffices for all Overdrive protocols
-# MOD = -DMAX_MOD_SZ=10 -DGFP_MOD_SZ=2
+# MOD = -DMAX_MOD_SZ=10 -DGFP_MOD_SZ=5
 
 LDLIBS = -lmpirxx -lmpir -lsodium $(MY_LDLIBS)
 LDLIBS += -lboost_system -lssl -lcrypto

Compiler/GC/instructions.py

@@ -521,7 +521,7 @@ class inputb(base.DoNotEliminateInstruction, base.VarArgsInstruction):
 
 class inputbvec(base.DoNotEliminateInstruction, base.VarArgsInstruction,
  base.Mergeable):
- """ Copy private input to secret bit register bit by bit. The input is
+ """ Copy private input to secret bit registers bit by bit. The input is
  read as floating-point number, multiplied by a power of two, rounded to an
  integer, and then decomposed into bits.