add some examples and modify README

fmlab-iis · Mar 5, 2020 · b2ab3ad · b2ab3ad
1 parent 0670cea
commit b2ab3ad
Show file tree

Hide file tree

Showing 2 changed files with 139 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -31,7 +31,7 @@ You can see the `gcc2cryptoline.so`  under the `build` directory and how to use
 ```
 mkdir build
 cd build
-cmake -DCMAKE_BUILD_TYPE=Release ..
+cmake ..
 ```
 the plugin is `gcc2cryptoline.so`
 
@@ -81,6 +81,11 @@ Becasue of the gramma of cryptoline parser, we need to escape some characters su
   - c like format:	`-fdump-tree-optimized`
   - raw format:		`-fdump-tree-optimized-raw`
 
+## Troubleshooting
+Not all GIMPLE statements are supported (maybe throw some errors in runtime).
+
+Please report such errors!
+
 ## Citation
 ### BibTeX
 ```

diff --git a/examples/relic/default/relic_fp_mul_low/fp_mula_low_tuned.cl b/examples/relic/default/relic_fp_mul_low/fp_mula_low_tuned.cl
@@ -0,0 +1,133 @@
+/* Multiplies a digit vector by a digit and adds this result to another digit vector. Computes c = c + a * digit. */
+
+proc main (uint64 c0, uint64 c1, uint64 c2, uint64 c3,
+           uint64 a0, uint64 a1, uint64 a2, uint64 a3,
+           uint64 digit) =
+{
+  true
+  &&
+  true
+}
+
+
+/* Start with undefined rhs */
+mov a18_0@uint64 a0;
+mov a18_8@uint64 a1;
+mov a18_16@uint64 a2;
+mov a18_24@uint64 a3;
+mov c17_0@uint64 c0;
+mov c17_8@uint64 c1;
+mov c17_16@uint64 c2;
+mov c17_24@uint64 c3;
+mov digit20@uint64 digit;
+/* End with undefined rhs */
+
+
+
+/* BB's index is 2 */
+
+/* _27 = *c_17(D); */
+mov v27 c17_0;
+/* _15 = (__int128 unsigned) _27; */
+vpc v15@uint128 v27@uint64;
+/* _14 = *a_18(D); */
+mov v14 a18_0;
+/* _12 = _14 w* digit_20(D); */
+umulj v12 v14 digit20;
+/* _38 = _12 + _15; */
+uadd v38 v12 v15;
+/* _43 = (long long unsigned int) _38; */
+cast v43@uint64 v38@uint128;
+/* *c_17(D) = _43; */
+mov c17_0 v43;
+/* _45 = _38 >> 64; */
+usplit v45 tmp_to_use v38 64;
+vpc tmp_to_use_p@uint64 tmp_to_use;
+assert true && tmp_to_use_p = v43;
+assume tmp_to_use_p = v43 && true;
+/* _57 = MEM[(dig_t *)c_17(D) + 8B]; */
+mov v57 c17_8;
+/* _58 = (__int128 unsigned) _57; */
+vpc v58@uint128 v57@uint64;
+/* _59 = MEM[(const dig_t *)a_18(D) + 8B]; */
+mov v59 a18_8;
+/* _61 = digit_20(D) w* _59; */
+umulj v61 digit20 v59;
+/* _39 = _58 + _61; */
+uadd v39 v58 v61;
+/* r_64 = _39 + _45; */
+uadd r64 v39 v45;
+/* _65 = (long long unsigned int) r_64; */
+cast v65@uint64 r64@uint128;
+/* MEM[(dig_t *)c_17(D) + 8B] = _65; */
+mov c17_8 v65;
+/* _67 = r_64 >> 64; */
+usplit v67 tmp_to_use r64 64;
+vpc tmp_to_use_p@uint64 tmp_to_use;
+assert true && tmp_to_use_p = v65;
+assume tmp_to_use_p = v65 && true;
+/* _79 = MEM[(dig_t *)c_17(D) + 16B]; */
+mov v79 c17_16;
+/* _80 = (__int128 unsigned) _79; */
+vpc v80@uint128 v79@uint64;
+/* _81 = MEM[(const dig_t *)a_18(D) + 16B]; */
+mov v81 a18_16;
+/* _83 = digit_20(D) w* _81; */
+umulj v83 digit20 v81;
+/* _40 = _80 + _83; */
+uadd v40 v80 v83;
+/* r_86 = _40 + _67; */
+uadd r86 v40 v67;
+/* _87 = (long long unsigned int) r_86; */
+cast v87@uint64 r86@uint128;
+/* MEM[(dig_t *)c_17(D) + 16B] = _87; */
+mov c17_16 v87;
+/* _89 = r_86 >> 64; */
+usplit v89 tmp_to_use r86 64;
+vpc tmp_to_use_p@uint64 tmp_to_use;
+assert true && tmp_to_use_p = v87;
+assume tmp_to_use_p = v87 && true;
+/* _1 = MEM[(dig_t *)c_17(D) + 24B]; */
+mov v1 c17_24;
+/* _2 = (__int128 unsigned) _1; */
+vpc v2@uint128 v1@uint64;
+/* _3 = MEM[(const dig_t *)a_18(D) + 24B]; */
+mov v3 a18_24;
+/* _6 = _3 w* digit_20(D); */
+umulj v6 v3 digit20;
+/* _41 = _2 + _6; */
+uadd v41 v2 v6;
+/* r_21 = _41 + _89; */
+uadd r21 v41 v89;
+/* _9 = (long long unsigned int) r_21; */
+cast v9@uint64 r21@uint128;
+/* MEM[(dig_t *)c_17(D) + 24B] = _9; */
+mov c17_24 v9;
+/* _10 = r_21 >> 64; */
+usplit v10 tmp_to_use r21 64;
+vpc tmp_to_use_p@uint64 tmp_to_use;
+assert true && tmp_to_use_p = v9;
+assume tmp_to_use_p = v9 && true;
+/* carry_23 = (dig_t) _10; */
+vpc carry23@uint64 v10@uint128;
+/* return carry_23; */
+
+
+/* Start with unused lhs */
+mov r0 c17_0@uint64;
+mov r1 c17_8@uint64;
+mov r2 c17_16@uint64;
+mov r3 c17_24@uint64;
+mov carry_out carry23@uint64;
+/* End with unsed lhs */
+
+
+{
+  (limbs 64 [r0, r1, r2, r3, carry_out])
+  = (
+      (limbs 64 [c0, c1, c2, c3])
+      + (limbs 64 [a0, a1, a2, a3]) * digit
+  )
+  &&
+  true
+}