diff --git a/.gitignore b/.gitignore
index c8d76e0..510df94 100644
--- a/.gitignore
+++ b/.gitignore
@@ -5,4 +5,5 @@ tools/obj_pb2.py
 work-obj93.cf
 *.ghw
 *.gen.vhdl
+*.bin
 .*.swp
diff --git a/dsp/bootloader.c b/dsp/bootloader.c
new file mode 100644
index 0000000..8dc1a02
--- /dev/null
+++ b/dsp/bootloader.c
@@ -0,0 +1,40 @@
+#include "sys.h"
+
+int main() {
+  uint16_t addr;
+  uint8_t len;
+  uint16_t val;
+
+  led0->output = 1;
+
+  while(1) {
+    uint8_t c = uart_read(uart0);
+    if (c == 'c') {
+      c = uart_read(uart0);
+      addr = c << 8;
+      c = uart_read(uart0);
+      addr |= c;
+
+      len = uart_read(uart0);
+      for (int i = 0; i < len; i += 2) {
+        c = uart_read(uart0);
+        val = c << 8;
+        c = uart_read(uart0);
+        val |= c;
+
+        *((uint16_t*)(addr + i)) = val;
+      }
+
+      uart_write(uart0, 'a');
+      uart_write(uart0, len);
+      led0->output = 0xf0;
+    } else if (c == 'j') {
+      c = uart_read(uart0);
+      addr = c << 8;
+      c = uart_read(uart0);
+      addr |= c;
+
+      ((void (*)())addr)();
+    }
+  }
+}
diff --git a/dsp/hello.c b/dsp/hello.c
new file mode 100644
index 0000000..b382731
--- /dev/null
+++ b/dsp/hello.c
@@ -0,0 +1,35 @@
+#include "sys.h"
+
+#define CYCLES_PER_MS 6666  // ish
+
+void busy_sleep_1ms() {
+  for (int i = 0; i < CYCLES_PER_MS; ++i) {
+    // nothing
+  }
+}
+
+/** waits a general amount of time */
+void busy_sleep(int ms) {
+  for (int i = 0; i < ms; ++i) {
+    busy_sleep_1ms();
+  }
+}
+
+int strlen(const char* str) {
+  int i = 0;
+  while(str[i] != '\0') {
+    i++;
+  }
+  return i;
+}
+
+void log(const char* stuff) {
+  uart_writen(uart0, stuff, strlen(stuff));
+}
+
+int main() {
+  while(1) {
+    log("Hello, world!\r\n");
+    busy_sleep(1000);
+  }
+}
diff --git a/dsp/makefile b/dsp/makefile
index 2261c0c..c326be7 100644
--- a/dsp/makefile
+++ b/dsp/makefile
@@ -8,7 +8,13 @@ LD = ../bin/ld
 
 CFLAGS = -I../wave -I../uart
 
-boot_rom.gen.vhdl: main.o ../uart/uart.o
+offset = $(shell printf "%d" 0x1100)
+
+boot_rom.gen.vhdl: bootloader.o ../uart/uart.o
+
+hello.bin: hello.o ../uart/uart.o
+
+synth.bin: main.o ../uart/uart.o
 
 sim_sources = dsp_test.vhdl
 sources = boot_rom.gen.vhdl dsp.vhdl ram.vhdl     \
@@ -24,6 +30,9 @@ sources = boot_rom.gen.vhdl dsp.vhdl ram.vhdl     \
 	../lab/dsp/au_base_project.runs/synth_1/runme.sh
 
 
+%.bin:
+	$(LD) -o $@ --offset $(offset) $^
+
 %.gen.vhdl:
 	$(LD) -o $@ --vhdl $*.vhdl.in $^
 
@@ -38,7 +47,7 @@ work-obj93.cf: $(sim_sources) $(sources)
 PHONY: bin sim rom clean
 
 clean:
-	rm -rf *.o *.gen.vhdl *.ghw work-obj93.cf
+	rm -rf *.o *.gen.vhdl *.ghw work-obj93.cf *.bin
 
 .DELETE_ON_ERROR:
 
diff --git a/dsp/prog.py b/dsp/prog.py
new file mode 100644
index 0000000..02be094
--- /dev/null
+++ b/dsp/prog.py
@@ -0,0 +1,46 @@
+import serial
+import struct
+import sys
+
+offset = 0x1100
+tty = 'tty'
+baud = 1000000
+
+
+def write(s, offset, dat):
+    for i in range(0, len(dat), 128):
+        chunk = dat[i: i + 128]
+        cmd = struct.pack('>cHB', b'c', offset + i, len(chunk))
+
+        print(f'Sending {len(chunk)} bytes @0x{offset + i:04x}')
+
+        s.write(cmd)
+        s.write(chunk)
+
+        ack = s.read(2)
+        if len(ack) < 2:
+            raise RuntimeError(f'timeout waiting for full ack. got {ack}')
+        if ack[0] != b'a'[0]:
+            raise RuntimeError(f'expected ack, got this instead: {ack}')
+        print(f'Ack! len={ack[1]}')
+
+
+def jump(s, offset):
+        addr = struct.pack('>H', offset)
+
+        s.write(b'j')
+        s.write(addr)
+
+
+def main():
+    binfile = sys.argv[1]
+    with open(binfile, 'rb') as f:
+        dat = f.read()
+
+    s = serial.Serial(tty, baud, timeout=1)
+    write(s, offset, dat)
+    jump(s, offset)
+
+
+if __name__ == "__main__":
+    main()