Upstream changes from the latest Astral developments.

.gitignore

@@ -5,3 +5,7 @@
 /**/*~
 !boot_code/crt0.S
 /simplified-runtime.xml
+*.log
+*.wlf
+modelsim.ini
+transcript

.gitmodules

@@ -0,0 +1,6 @@
+[submodule "hwpe/neureka/pulp-nnx"]
+	path = hwpe/neureka/pulp-nnx
+	url = https://github.com/pulp-platform/pulp-nnx
+[submodule "hwpe/redmule/redmule-golden-model"]
+	path = hwpe/redmule/redmule-golden-model
+	url = https://github.com/yvantor/redmule-golden-model.git

astral.yaml

@@ -0,0 +1,19 @@
+astral:
+  parMatrixMul8:
+    path: ./astral/parMatrixMul8
+    command: make clean all run
+  parMatrixMul16:
+    path: ./astral/parMatrixMul16
+    command: make clean all run
+  parMatrixMul32:
+    path: ./astral/parMatrixMul32
+    command: make clean all run
+  dmr_matmul:
+    path: ./astral/dmr_matmul
+    command: make clean all run
+  redmule:
+    path: ./astral/redmule
+    command: make clean all run
+  neureka:
+    path: ./astral/neureka
+    command: make clean all run

astral/dmr_matmul

@@ -0,0 +1 @@
+../reliability_tests/dmr_matmul

astral/ecc_test

@@ -0,0 +1 @@
+../reliability_tests/ecc_test

astral/hello/Makefile

@@ -0,0 +1,5 @@
+PULP_APP = test
+PULP_APP_SRCS = hello.c
+PULP_CFLAGS = -O3
+
+include $(PULP_SDK_HOME)/install/rules/pulp.mk

astral/hello/hello.c

@@ -0,0 +1,24 @@
+/*
+ * Copyright (C) 2018 ETH Zurich and University of Bologna
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <stdio.h>
+
+int main()
+{
+  printf("Hello !\n");
+
+  return 0;
+}

astral/icache_fi_conv16

@@ -0,0 +1 @@
+../reliability_tests/icache_fi_conv16

astral/neureka

@@ -0,0 +1 @@
+../hwpe/neureka/

astral/parMatrixMul16/Makefile

@@ -0,0 +1,8 @@
+PULP_APP = test
+PULP_APP_SRCS = matrixMul.c
+
+PULP_CFLAGS = -O3
+
+include $(PULP_SDK_HOME)/install/rules/pulp.mk
+
+#pulp-bench-reg --name=parMatrixMul16.cycles --module=pulp_rtl_testset --pipeline=$(PIPELINE) --artefact=pulp_rtl_testset --cmd="make run -f Makefile.sdk" --probe-regexp='matrixMul -> success, nr. of errors: 0, execution time: (\d+)' --params="platform($(platformName)),compiler($(OR1K_TOOLCHAIN_TYPE)),nbCores(4),elemSize(16)" --probe-regexp='matrixMulTranspose -> success, nr. of errors: 0, execution time: (\d+)' --params="platform($(platformName)),compiler($(OR1K_TOOLCHAIN_TYPE)),nbCores(4),elemSize(16),transposed"

astral/parMatrixMul16/gen_stimuli.py

@@ -0,0 +1,54 @@
+#!/usr/bin/env python
+
+import sys
+import random
+
+
+def write_arr(f, name, arr):
+    f.write('const short %s[] = {\n' % name)
+    for v in arr:
+        f.write('%d,\n' % (v))
+    f.write('};\n\n')
+    return
+
+################################################################################
+f = open('parMatrixMul16_stimuli.h', 'w')
+
+
+SIZE = 24 
+RANGE = int(2**7/SIZE)
+
+m_a   = []
+m_b   = []
+m_exp = []
+
+for i in range(0,SIZE):
+    for j in range(0,SIZE):
+        a = random.randint(-RANGE, RANGE-1)
+        b = random.randint(-RANGE, RANGE-1)
+
+        m_a.append(a)
+        m_b.append(b)
+
+for i in range(0,SIZE):
+    for j in range(0,SIZE):
+        r = 0
+
+        for k in range (0,SIZE):
+            r = r + m_a[i * SIZE + k] * m_b[k * SIZE + j]
+
+        m_exp.append(r)
+
+
+write_arr(f, 'm_a',   m_a)
+write_arr(f, 'm_b',   m_b)
+write_arr(f, 'm_exp', m_exp)
+
+f.write('#define SIZE %d\n' % SIZE)
+
+
+f.write('__attribute__ ((section(".heapsram"))) short g_mA[SIZE][SIZE];\n')
+f.write('__attribute__ ((section(".heapsram"))) short g_mB[SIZE][SIZE];\n')
+f.write('__attribute__ ((section(".heapsram"))) short g_mC[SIZE][SIZE];\n')
+f.write('__attribute__ ((section(".heapsram"))) short g_mB_tmp[SIZE][SIZE];\n')
+

astral/parMatrixMul16/matrixMul.c

@@ -0,0 +1,177 @@
+/*
+ * Copyright (C) 2018 ETH Zurich and University of Bologna
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/* 
+ * Mantainer: Luca Valente, [email protected]
+ */
+
+#include "pulp.h"
+
+#include "parMatrixMul16_stimuli.h"
+
+void check_matrix_mul(testresult_t *result, void (*start)(), void (*stop)());
+void check_matrix_mul_transpose(testresult_t *result, void (*start)(), void (*stop)());
+
+testcase_t testcases[] = {
+  { .name = "matrixMul",          .test = check_matrix_mul           },
+  { .name = "matrixMulTranspose", .test = check_matrix_mul_transpose },
+  {0, 0}
+};
+
+unsigned int num_cores;
+
+int main()
+{
+  if (rt_cluster_id() != 0)
+    return bench_cluster_forward(0);
+
+  num_cores = get_core_num();
+
+  if(rt_core_id() < num_cores) {
+    run_suite(testcases);
+  }
+
+  synch_barrier();
+
+  return 0;
+}
+
+void matrix_init();
+unsigned int matrix_check();
+
+void check_matrix_mul(testresult_t *result, void (*start)(), void (*stop)()) {
+  int core_id;
+  unsigned int i, j, k;
+  unsigned int chunk;
+  unsigned int lb, ub;
+
+  core_id = get_core_id();
+
+  // number of rows each core has to multiply
+  chunk = SIZE / num_cores;
+  // lower bound
+  lb = core_id * chunk;
+  // upper bound
+  ub = lb + chunk;
+
+  if(core_id == 0) {
+    matrix_init();
+  }
+
+  if(num_cores != 1) synch_barrier();
+
+  // start benchmark
+  start();
+
+  for(i = lb; i < ub; i++) {
+    for(j = 0; j < SIZE; j++) {
+      g_mC[i][j] = 0;
+
+      for(k = 0; k < SIZE; k++) {
+        g_mC[i][j] += g_mA[i][k] * g_mB[k][j];
+      }
+    }
+  }
+
+  if(num_cores != 1) synch_barrier();
+
+  stop();
+
+  if(core_id == 0) {
+    result->errors = matrix_check();
+  }
+}
+
+void check_matrix_mul_transpose(testresult_t *result, void (*start)(), void (*stop)()) {
+  int core_id;
+  unsigned int i, j, k;
+  unsigned int chunk;
+  unsigned int lb, ub;
+
+  core_id = get_core_id();
+
+  // number of rows each core has to multiply
+  chunk = SIZE / num_cores;
+  // lower bound
+  lb = core_id * chunk;
+  // upper bound
+  ub = lb + chunk;
+
+  if(core_id == 0) {
+    matrix_init();
+  }
+
+  if(num_cores != 1) synch_barrier();
+
+  // start benchmark
+  start();
+
+  // transpose array before using it
+  for(i = lb; i < ub; i++) {
+    for(j = 0; j < SIZE; j++) {
+      g_mB_tmp[i][j] = g_mB[j][i];
+    }
+  }
+
+  if(num_cores != 1) synch_barrier();
+
+  for(i = lb; i < ub; i++) {
+    for(j = 0; j < SIZE; j++) {
+      g_mC[i][j] = 0;
+
+      for(k = 0; k < SIZE; k++) {
+        g_mC[i][j] += g_mA[i][k] * g_mB_tmp[j][k];
+      }
+    }
+  }
+
+  if(num_cores != 1) synch_barrier();
+
+  stop();
+
+  if(core_id == 0) {
+    result->errors = matrix_check();
+  }
+}
+
+void matrix_init() {
+  unsigned int i, j;
+
+  // init, copy to TCDM
+  for(i = 0; i < SIZE; i++) {
+    for(j = 0; j < SIZE; j++) {
+      g_mA[i][j] = m_a[i * SIZE + j];
+      g_mB[i][j] = m_b[i * SIZE + j];
+      g_mC[i][j] = 0;
+    }
+  }
+}
+
+unsigned int matrix_check() {
+  unsigned int errors = 0;
+  unsigned int i, j;
+  // check
+  for(i = 0; i < SIZE; i++) {
+    for(j = 0; j < SIZE; j++) {
+      if(g_mC[i][j] != m_exp[i * SIZE + j]) {
+        printf("At index %d, %d\n", i, j, 0, 0);
+        errors++;
+      }
+    }
+  }
+
+  return errors;
+}