Add example_data_procssing_from_flash (#568)

sw/applications/example_data_processing_from_flash/gen_stimuly.py

@@ -0,0 +1,65 @@
+#!/usr/bin/env python
+
+## Copyright 2024 EPFL
+## Solderpad Hardware License, Version 2.1, see LICENSE.md for details.
+## SPDX-License-Identifier: Apache-2.0 WITH SHL-2.1
+
+# type " python gen_stimuly.py " in the terminal to generate the matrices.h file
+
+import sys
+import random
+import numpy as np
+
+def write_arr(f, name, arr, ctype, size):
+    f.write("const " + ctype + " " + name + "[] = {\n")
+
+    for row in arr:
+        for elem in row[:-1]:
+            f.write('%d,' % (elem))
+        f.write('%d,\n' % (row[-1]))
+
+    f.write('};\n\n')
+    return
+
+def write_arr_flash_only(f, name, arr, ctype, size):
+    f.write( ctype + " __attribute__((section(\".xheep_data_flash_only\"))) __attribute__ ((aligned (16)))" + name + "[] = {\n")
+
+    for row in arr:
+        for elem in row[:-1]:
+            f.write('%d,' % (elem))
+        f.write('%d,\n' % (row[-1]))
+
+    f.write('};\n\n')
+    return
+
+
+################################################################################
+f = open('matrices.h', 'w')
+f.write('#ifndef MATRICES_H_\n')
+f.write('#define MATRICES_H_\n')
+f.write('// This file is automatically generated\n')
+
+
+SIZE  = 64
+RANGE = 10
+
+m_a   = []
+m_b   = []
+m_exp = []
+
+# Generate random 8 bit integers from -RANGE to RANGE for A and B
+A = np.random.randint(0, RANGE, size=(SIZE, SIZE), dtype=np.int32)
+B = np.random.randint(0, RANGE, size=(SIZE, SIZE), dtype=np.int32)
+C = np.zeros((SIZE, SIZE), dtype=np.int32)
+
+# Test the function with A and B
+C = np.matmul(A,B)
+
+write_arr_flash_only(f, 'A',   A, 'int32_t', SIZE)
+write_arr(f, 'B',   B, 'int32_t', SIZE)
+write_arr(f, 'C',   C, 'int32_t', SIZE)
+
+f.write('#define MATRIX_SIZE %d\n' % SIZE)
+
+
+f.write('#endif')

sw/applications/example_data_processing_from_flash/main.c

@@ -0,0 +1,87 @@
+// Copyright 2024 EPFL
+// Solderpad Hardware License, Version 2.1, see LICENSE.md for details.
+// SPDX-License-Identifier: Apache-2.0 WITH SHL-2.1
+//
+// File: sw/applications/example_data_processing_from_flash/main.c
+// Author:  Francesco Poluzzi
+// Date: 29/07/2024
+
+/**
+ * @file main.c
+ * @brief Example of data processing (matrix multiplication) reading data from flash memory
+ *
+ * Simple example that read a matrix from flash memory in many step and performs
+ * matrix multiplication. This is useful for applications where the
+ * data size does not fit in the available SRAM memory, so some data needs to be
+ * stored as "flash_only" and read trough the spi interface. This usually requires
+ * filling a buffer and tiling the data processing.
+*/
+
+#include <stdio.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+#include "x-heep.h"
+#include "w25q128jw.h"
+#include "main.h"
+#include "dma_sdk.h"
+
+#define TILING_ROWS 2
+
+ /* By default, printfs are activated for FPGA and disabled for simulation. */
+#define PRINTF_IN_FPGA  1
+#define PRINTF_IN_SIM   0
+#if TARGET_SIM && PRINTF_IN_SIM
+        #define PRINTF(fmt, ...)    printf(fmt, ## __VA_ARGS__)
+#elif PRINTF_IN_FPGA && !TARGET_SIM
+    #define PRINTF(fmt, ...)    printf(fmt, ## __VA_ARGS__)
+#else
+    #define PRINTF(...)
+#endif
+
+int32_t buffer_data[MATRIX_SIZE*TILING_ROWS] = {0};
+int32_t output_matrix[MATRIX_SIZE*MATRIX_SIZE] = {0};
+
+int main(int argc, char *argv[]) {
+
+    soc_ctrl_t soc_ctrl;
+    soc_ctrl.base_addr = mmio_region_from_addr((uintptr_t)SOC_CTRL_START_ADDRESS);
+
+    #ifdef TARGET_SIM
+        PRINTF("This application is meant to run on FPGA only\n");
+        return EXIT_SUCCESS;
+    #endif
+
+    if ( get_spi_flash_mode(&soc_ctrl) == SOC_CTRL_SPI_FLASH_MODE_SPIMEMIO ) {
+        PRINTF("This application cannot work with the memory mapped SPI FLASH"
+            "module - do not use the FLASH_EXEC linker script for this application\n");
+        return EXIT_SUCCESS;
+    }
+
+    // Pick the correct spi device based on simulation type
+    spi_host_t* spi = spi_flash;
+
+    // Init SPI host and SPI<->Flash bridge parameters
+    if (w25q128jw_init(spi) != FLASH_OK){
+        PRINTF("Error initializing SPI flash\n");
+        return EXIT_FAILURE;
+    } 
+
+    for (int i = 0; i < MATRIX_SIZE; i+=TILING_ROWS) {
+        // read first half matrix A from flash and perform matmul
+        if(fill_buffer(&A[i*MATRIX_SIZE], buffer_data, MATRIX_SIZE*TILING_ROWS)!=FLASH_OK){
+            PRINTF("Error reading from flash\n");
+            return EXIT_FAILURE;
+        }
+        matmul(buffer_data, B, &output_matrix[i*MATRIX_SIZE], TILING_ROWS, MATRIX_SIZE, MATRIX_SIZE);
+    }
+
+    for(int i = 0; i < MATRIX_SIZE*MATRIX_SIZE; i++){
+        if (output_matrix[i] != C[i]){
+             PRINTF("Result[%d][%d]:golden model   %d : %d\n", (i/MATRIX_SIZE), (i % MATRIX_SIZE), output_matrix[i], C[i]);
+            return EXIT_FAILURE;
+       }
+    }
+    PRINTF("All tests passed!\n");
+    return EXIT_SUCCESS;
+}

sw/applications/example_data_processing_from_flash/main.h

@@ -0,0 +1,42 @@
+// Copyright 2024 EPFL
+// Solderpad Hardware License, Version 2.1, see LICENSE.md for details.
+// SPDX-License-Identifier: Apache-2.0 WITH SHL-2.1
+//
+// File: sw/applications/example_data_processing_from_flash/main.h
+// Author:  Francesco Poluzzi
+// Date: 29/07/2024
+
+#ifndef MAIN_H_
+#define MAIN_H_
+
+#include <stdio.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+#include "x-heep.h"
+#include "w25q128jw.h"
+#include "matrices.h"
+
+w25q_error_codes_t fill_buffer(uint32_t *source, uint32_t *buffer, uint32_t len);
+void matmul(int32_t *A, int32_t *B, int32_t *res, int rowsA, int colsA, int colsB);
+
+w25q_error_codes_t fill_buffer(uint32_t *source, uint32_t *buffer, uint32_t len){
+    uint32_t *source_flash = heep_get_flash_address_offset(source);
+    w25q_error_codes_t status = w25q128jw_read_standard(source_flash, buffer, len*4);
+    return status;
+}
+
+void matmul(int32_t *A, int32_t *B, int32_t *res, int rowsA, int colsA, int colsB) {
+    for (int i = 0; i < rowsA; i++) {
+        for (int j = 0; j < colsB; j++) {
+            int32_t sum = 0;
+            for (int k = 0; k < colsA; k++) {
+                sum += A[i*colsA + k] * B[k*colsB + j];
+            }
+            res[i*colsB + j] = sum;
+        }
+    }
+}
+
+
+#endif // DATA_H_