Remove all debug and printf

rtl/ip/matmul/matmul_axi_slave.v

@@ -341,8 +341,6 @@ module matmul_axi_slave (
                 if (c0_part == 2'd0)      s_axi_rdata = target_c0[31:0];
                 else if (c0_part == 2'd1) s_axi_rdata = target_c0[63:32];
                 else                      s_axi_rdata = {30'b0, target_c0[65:64]};
-                $display("RCH B0_C off=%h elem=%d part=%d data=%h raw=%h",
-                         current_r_offset, c0_elem, c0_part, s_axi_rdata, target_c0);
             end
             else if (current_r_offset >= ADDR_B1_A && current_r_offset < ADDR_B1_B)
                 s_axi_rdata = b1_a[(current_r_offset - ADDR_B1_A) >> 2];
@@ -352,8 +350,6 @@ module matmul_axi_slave (
                 if (c1_part == 2'd0)      s_axi_rdata = target_c1[31:0];
                 else if (c1_part == 2'd1) s_axi_rdata = target_c1[63:32];
                 else                      s_axi_rdata = {30'b0, target_c1[65:64]};
-                $display("RCH B1_C off=%h elem=%d part=%d data=%h raw=%h",
-                         current_r_offset, c1_elem, c1_part, s_axi_rdata, target_c1);
             end
         end
     end
@@ -449,10 +445,6 @@ module matmul_axi_slave (
                     else
                         b1_c[calc_idx] <= ({2'b0, mult_p_r[0]} + {2'b0, mult_p_r[1]}) +
                                           ({2'b0, mult_p_r[2]} + {2'b0, mult_p_r[3]});
-                    $display("ACCUM calc_idx=%d expr=%h",
-                             calc_idx,
-                             ({2'b0, mult_p_r[0]} + {2'b0, mult_p_r[1]}) +
-                             ({2'b0, mult_p_r[2]} + {2'b0, mult_p_r[3]}));
                     
                     if (calc_idx == 4'd15) begin
                         calc_done_pulse <= 1'b1;

sdk/software/examples/asm/user-sample.c

@@ -195,17 +195,6 @@ int main(void)
     dma_start_transfer(1, phys_bank_C[(MATMUL_GROUP_NUM - 1) % 2], EXTRAM_PHYS_C(MATMUL_GROUP_NUM - 1), 192, 20);
     dma_wait_polling(1); // 必须等待最后一组写回完成
 
-    /* ── DEBUG: read back specific C elements from ExtRAM ─────── */
-    #define C_WORD(g, elem, part) extram[DST_BASE_WORDS + (g)*48 + (elem)*3 + (part)]
-    printf("DBG g=0 C00=[%08lx %08lx %08lx] C11=[%08lx %08lx %08lx] C23=[%08lx %08lx %08lx]\n",
-        C_WORD(0,0,0),C_WORD(0,0,1),C_WORD(0,0,2),
-        C_WORD(0,5,0),C_WORD(0,5,1),C_WORD(0,5,2),
-        C_WORD(0,11,0),C_WORD(0,11,1),C_WORD(0,11,2));
-    printf("DBG g=1 C00=[%08lx %08lx %08lx] C11=[%08lx %08lx %08lx] C23=[%08lx %08lx %08lx]\n",
-        C_WORD(1,0,0),C_WORD(1,0,1),C_WORD(1,0,2),
-        C_WORD(1,5,0),C_WORD(1,5,1),C_WORD(1,5,2),
-        C_WORD(1,11,0),C_WORD(1,11,1),C_WORD(1,11,2));
-
     /* ── CRC32 over result region ─────────────────────────────── */
     crc = compute_crc32(extram);
 

sdk/software/examples/fft/fft.c

@@ -0,0 +1,14 @@
+#include "fft.h"
+#include "common_func.h"
+
+void fft_start() {
+    RegWrite(FFT_CSR_REG, FFT_CTRL_START);
+}
+
+void fft_wait() {
+    while ((RegRead(FFT_CSR_REG) & FFT_STAT_DONE) == 0) {}
+}
+
+unsigned int fft_get_csr() {
+    return RegRead(FFT_CSR_REG);
+}

sdk/software/examples/fft/fft.h

@@ -0,0 +1,16 @@
+#pragma once
+
+#define FFT_BASE          0xbf400000
+#define FFT_IN_RE_BASE    (FFT_BASE + 0x1000)
+#define FFT_IN_IM_BASE    (FFT_BASE + 0x2000)
+#define FFT_OUT_RE_BASE   (FFT_BASE + 0x3000)
+#define FFT_OUT_IM_BASE   (FFT_BASE + 0x4000)
+#define FFT_CSR_REG       (FFT_BASE + 0xF000)
+#define FFT_CTRL_START    (1 << 4)
+#define FFT_STAT_DONE     (1 << 1)
+#define FFT_STAT_BUSY     (1 << 0)
+#define FFT_POINT_NUM     1024
+
+void fft_start();
+void fft_wait();
+unsigned int fft_get_csr();

sdk/software/examples/fft_dma/dma.c

@@ -0,0 +1,37 @@
+#include "dma.h"
+
+void dma_start_transfer(int ch_id, uint32_t src, uint32_t dst, uint32_t bytes, uint32_t tag) {
+    volatile dma_ch_regs_t* ch = DMA_CH(ch_id);
+
+    // 检查通道是否空闲 防止覆盖正在运行的任务
+    if (ch->STATUS & STATUS_BUSY_BIT) {
+        // printf("channel %d busy\n", ch_id);
+        return; // 通道正忙，处理报错或重试
+    }
+
+    // 写入基本地址和长度
+    ch->SRC_ADDR = src;
+    ch->DST_ADDR = dst;
+    ch->LENGTH   = bytes;
+    ch->TAG      = tag; // 可选，填入任务ID
+
+    // 置 Burst 属性并触发传输
+    // AXI SIZE: 2 (代表 2^2 = 4 Bytes，与 32-bit 数据线匹配)
+    // AXI LEN: 15 (代表 16 beats 突发传输，最高效)
+    uint32_t ctrl_val = CTRL_BURST_SIZE(2) | CTRL_BURST_LEN(15);
+    
+    // 写入 CTRL 并拉高 Bit 0 (Start)
+    ch->CTRL = ctrl_val | CTRL_START_BIT;
+}
+
+void dma_wait_polling(int ch_id) {
+    volatile dma_ch_regs_t* ch = DMA_CH(ch_id);
+    
+    // 死等 Done bit 置 1
+    while (!(ch->STATUS & STATUS_DONE_BIT)) {
+    }
+
+    // 清除 Done 标志位 (Write 1 to Clear Bit 1)
+    // 注意：写 1 清 0 的设计，所以我们对 bit 1 写入 1
+    ch->STATUS = STATUS_DONE_BIT; 
+}

sdk/software/examples/fft_dma/dma.h

@@ -0,0 +1,30 @@
+#pragma once
+
+#include <stdint.h>
+
+
+typedef struct {
+    volatile uint32_t SRC_ADDR;  // 0x00: 源地址
+    volatile uint32_t DST_ADDR;  // 0x04: 目的地址
+    volatile uint32_t LENGTH;    // 0x08: 传输总字节数
+    volatile uint32_t TAG;       // 0x0C: 软件 Tag (硬件不处理，留给软件标记任务用)
+    volatile uint32_t CTRL;      // 0x10: 控制寄存器 (配置 Burst 属性 + 启动)
+    volatile uint32_t STATUS;    // 0x14: 状态寄存器 (Busy, Done)
+    volatile uint32_t RESV[10];  // 0x18 ~ 0x3C: 保留空间，凑齐 0x40 字节
+} dma_ch_regs_t;
+
+// DMA 控制器基地址
+#define CDMA_MUX_BASE 0xbf300000
+#define DMA_CH(i)     ((volatile dma_ch_regs_t*)(CDMA_MUX_BASE + (i) * 0x40))
+
+// CTRL 寄存器位定义
+#define CTRL_START_BIT       (1 << 0)
+#define CTRL_BURST_SIZE(x)   (((x) & 0x7) << 3)  // Bits [5:3]: AXI AxSIZE (0=1B, 1=2B, 2=4B, 3=8B)
+#define CTRL_BURST_LEN(x)    (((x) & 0xFF) << 6) // Bits [13:6]: AXI AxLEN (0=1 beat, 15=16 beats)
+
+// STATUS 寄存器位定义
+#define STATUS_BUSY_BIT      (1 << 0)
+#define STATUS_DONE_BIT      (1 << 1)
+
+void dma_start_transfer(int ch_id, uint32_t src, uint32_t dst, uint32_t bytes, uint32_t tag);
+void dma_wait_polling(int ch_id);

sdk/software/examples/fft_dma/fft.c

@@ -0,0 +1,14 @@
+#include "fft.h"
+#include "common_func.h"
+
+void fft_start() {
+    RegWrite(FFT_CSR_REG, FFT_CTRL_START);
+}
+
+void fft_wait() {
+    while ((RegRead(FFT_CSR_REG) & FFT_STAT_DONE) == 0) {}
+}
+
+unsigned int fft_get_csr() {
+    return RegRead(FFT_CSR_REG);
+}

sdk/software/examples/fft_dma/fft.h

@@ -0,0 +1,16 @@
+#pragma once
+
+#define FFT_BASE          0xbf400000
+#define FFT_IN_RE_BASE    (FFT_BASE + 0x1000)
+#define FFT_IN_IM_BASE    (FFT_BASE + 0x2000)
+#define FFT_OUT_RE_BASE   (FFT_BASE + 0x3000)
+#define FFT_OUT_IM_BASE   (FFT_BASE + 0x4000)
+#define FFT_CSR_REG       (FFT_BASE + 0xF000)
+#define FFT_CTRL_START    (1 << 4)
+#define FFT_STAT_DONE     (1 << 1)
+#define FFT_STAT_BUSY     (1 << 0)
+#define FFT_POINT_NUM     1024
+
+void fft_start();
+void fft_wait();
+unsigned int fft_get_csr();