Cypress DCT-1D Podręcznik Użytkownika Strona 1

Laboratory manual for TSEA44Olle Seger, Per Karlström, Andreas EhliarComputer EngineeringDepartment of Electrical EngineeringLinköping University, S-5

10CHAPTER 1. THE SYST EMDeviceSystemGatesCLB(1 CLB = 4 slices = Max 128 bits)MultiplierBlocksSelectRAM BlocksDCMsMax I/OPads(1)ArrayRow x Col. SlicesM

1.3. OPEN RISC11Global Clock MuxDCM DCM IOBCLBProgrammable I/OsBlock SelectRAM MultiplierConfigurable LogicFigure 1.3: Virtex-II architectural overvie

12CHAPTER 1. THE SYST EMLUTFXGinputsFXINA MUXFXFXINBDFF/LATQREVDCECLKSRBYBXCECLKSRYDYYQF5MUXF5XLUTFinputsDFF/LATQREVDCECLKSRDXXQa) b)Figure 1.4: a) Vi

1.3. OPEN RISC13WishboneOR1200 CPUMem Ctrl4kBRAM4kBROMUARTParportMaster SlaveEther Ctrl23721001Debug3JTAGPS24VGA5SRAMSRAM1

14CHAPTER 1. THE SYST EMThe OR1200 CPU, [8], consists of several blocks:• High Performance 32-Bit CPU/DSP– 32-bit architecture implementing ORBIS32 in

1.3. OPEN RISC15slave. Furthermore tristate is not used, instead there are two databuses, one in eachdirection. The address bus and the data busses ar

16CHAPTER 1. THE SYST EMThe Ethernet IP Core is capable of operating at 10 or 100 Mbps for Ethernet andFast Ethernet applications. An external PHY is

1.4. SOFTWARE17Command Explanationd <addr> display memory contentm <addr> <data> modify memory contentg <addr> go (execute)l l

18CHAPTER 1. THE SYST EMListing 1.2: Makefile for simpleprog (Listing 1.1)# The na me o f t h e p rog ram we w ant t o c o m pi l ePROGRAM = s i m p l

1.4. SOFTWARE19Listing 1.3: Link script for simpleprog (Listing 1.1)MEMORY{vec tors : ORIGIN = 0 x00000000 , LENGTH = 0 x000020 00sdram : ORIGIN = 0 x

2

20CHAPTER 1. THE SYST EMThe simulator can also be started in an interactive mode byor32-uclinux-sim -f sim.cfg -i prog .In Figure 1.8 we show as an ex

1.4. SOFTWARE21The command help will list the built-in shell commands.An important file is /etc/rc, the start-up file, which is shown in Listing 1.4. If

22CHAPTER 1. THE SYST EMListing 1.5: Program showing contents of a special purpose register.# i n c l u d e < s y s / t y p e s . h># i n c l u

Chapter 2Lab task 0 - Build a UART inVerilog2.1 IntroductionIn this introductory lab exercise you will learn the HDL Verilog. We r equire that youare

24CHAPTER 2. LAB TASK 0 - BUILD A UART IN VERILOG2.2.2 The hardwareThe system clock is running at 40 MHz. You will need a reset-signal and a send-sign

2.2. A SIMPLE UART25Listing 2.1: Test bench for the UART.‘ t i m e s c a l e 1 n s / 10 psmodule l a b 0 _ t b ( ) ;re g c l k _ i ;re g r s t _ i ;re

26CHAPTER 2. LAB TASK 0 - BUILD A UART IN VERILOG2.3 ExercisesPreparation task 1Draw a HW diagram of the UART. Use simple components like counters, re

2.4. GTKTERM USAGE27/ / b l u e DIP s w i t c hNET "switch_i <7>" LOC = "AL3" ; / / SWITCH 1NET "switch_i <6>&quo

28CHAPTER 2. LAB TASK 0 - BUILD A UART IN VERILOG

Chapter 3Lab task 1 - Interfacing to theWishbone bus3.1 IntroductionIn this lab exercise you will get acquainted with the OR 1200 RISC processor andpa

Contents1 The system 71.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.2 Hardware . . . . . . . . . . . . . . . . . .

30CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUS4. simulate the computer running the benchmark program.5. design a module containing hardware

3.2. SOME BASIC FACTS ON THE WISHBONE BUS313. The Master deasserts the wb.stb, wb.cyc and wb.we-signals.4. The slave deasserts the wb.ack-signal.For t

32CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUSIn the return path the addressed slave’s s-bus is connected to all the masters. Thisis handle

3.3. A SIMPLE COMPUTER33ControlUnitControlUnit>=1 txSRsendRSregrx_full F/Fwrtxtx_empty F/FshiftShiftRegrdin&&wb.stbwb.stbwb.wewb.sel[3]wb.d

34CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUS7 0 071523319000_0000123459000_00004sel[0]sel[1]sel[2]sel[3]a) b)tx_empty rx_fullrx_fulltx_em

3.3. A SIMPLE COMPUTER35Check mon2.c to see what the monitor does at startup so that you can verify thatthe hardware does the correct thing.3.3.4 Test

36CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUS3.4 A Benchmark Program3.4.1 JPEG CompressionWe will use the first part, DCT, of the JPEG comp

3.4. A BENCHMARK PROGRAM37Sofar we have presented three ways of computing the 2-D DCT. We compare thecomputation complexity of the algorithms:Algorith

38CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUSPreparation task 5Why do we go through all the trouble inserting the module in Figure 3.8? Wh

3.5. DESIGN A PERFORMANCE COUNTER MODULE393. contain four 32 bit counters that can be read and written on the addresses 0x9900_0000to 0x9900_000c.4. T

4CONTENTS3.3 A Simple Computer . . . . . . . . . . . . . . . . . . . . . . . . . . . 323.3.1 General . . . . . . . . . . . . . . . . . . . . . . . . .

40CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUS3.6 Useful CommandsWe have prepared a makefile based build system that is responsible for both

3.7. HOW TO GET STARTED WRITING/EXECUTING C PROGRAMS41• synthdir/foo.syr: Synthesis report• synthdir/foo_map.mrp: Map report• synthdir/foo.par: Place

42CHAPTER 3. LAB TASK 1 - INTERFACING TO THE WISHBONE BUSmacros shown in Listing 3.5 to access memory mapped I/O. These macros are definedin both the m

Chapter 4Lab task 2 - Design a JPEGaccelerator4.1 The lab systemIn this lab task you will learn how to build a hardware accelerator for the JPEG image

44CHAPTER 4. LAB TASK 2 - DESIGN A JPEG ACCELERATOR...DCT2 Control UnitcsrDCT648x16=12832NC32 NCRAMin8x12=96t_wrt_rd8x12=96WBCtrlTransposeMemoryBlock1

4.2. PROPOSED ARCHITECTURE45to use a block RAM is, in our opinion, to instantiate a library primitive. The code inListing 4.1 instantiates a block RAM

46CHAPTER 4. LAB TASK 2 - DESIGN A JPEG ACCELERATORwi r e [ 7 : 0 ] d a t a _ i , d a t a _ o ;wi r e [ 3 : 0 ] a d d r_a , a d d r _ b ;/ / 1 c o m b

4.3. INTRODUCTION TOµCLINUX47Laboration task 4Design and implement the DCT accelerator with a WB interface.Laboration task 5Write a testbench for your

48CHAPTER 4. LAB TASK 2 - DESIGN A JPEG ACCELERATORservers to be left af ter you log out since this would prohibit other lab groups fromstarting a TFT

4.4. INTRODUCTION TO JPEGFILES49• jpegtest.c contains the test program we will use• testbild.raw is a grayscale image in raw format.• perfctr.c,perfct

CONTENTS56 Lab task 4 - Custom Instructions 616.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616.1.1 Huffman Coding .

50CHAPTER 4. LAB TASK 2 - DESIGN A JPEG ACCELERATOR4.4.2 The jpegtest applicationThis is the main test application we are going to use in the lab seri

4.6. QUANTIZATION510 0.5 1 1.5 2 2.5x 10400.20.40.60.811.21.41.61.8dmadct(N)read(N) Q(N) huff(N)ACCCPUclockcyclesFigure 4.3: Timestamps for JPEG compr

52CHAPTER 4. LAB TASK 2 - DESIGN A JPEG ACCELERATORshifting 17 steps:Y [u, v] = round(A[u, v] − 8192 · δ[u, v]) · R[u, v] · 2−17=−96 −3

4.7. TIPS AND TRICKS534.7 Tips and tricksIn this section we have collected some notes that you might find useful.• If you want to simulate the 2D DCT a

54CHAPTER 4. LAB TASK 2 - DESIGN A JPEG ACCELERATOR

Chapter 5Lab task 35.1 DMA in the DCT AcceleratorIn this lab we will improve the DCT accelerator by using DMA. You can find a speci-fication on how to d

56CHAPTER 5. LAB TASK 3IDLEWAITREADYGETBLOCKWAITREADY_LAST RELEASEBUSFigure 5.1: The proposed state diagram for the DMA accelerator.In Figure 5.1 ther

5.1. DMA IN THE DCT ACCELERATOR57• dma_bram_data:The data we want to write to inmem in jpeg_top.• dma_bram_addr:The address we want to write the data

58CHAPTER 5. LAB TASK 3It is important to note that the DCT accelerator should still work as before if DMAis not in use. But you are (of course) allow

5.2. WHAT TO INCLUDE IN THE LAB REPORT595.2 What to Include in the Lab ReportThe lab report should contain all source code that you have written. (The

6CONTENTS

60CHAPTER 5. LAB TASK 3WBCtrlAddressGenerator...DCT2 Control UnitcsrDCT6432NCNCRAMin8x12=96t_wrt_rd8x12=96TransposeMemoryBlock132dma_bram_addrdma_bram

Chapter 6Lab task 4 - Custom Instructions6.1 IntroductionIn this lab task you will learn how to design and integrate a new instruction into theprocess

62CHAPTER 6. LAB TASK 4 - CUSTOM INSTRUCTIONSsuggests four Huffman tables. Two for the luminance, one for the DC values and onefor the AC values. And

6.3. PROPOSED ARCHITECTURE63tions in software. We will add a new group of special purpose registers to the pro-cessor, therefore we must make some cha

64CHAPTER 6. LAB TASK 4 - CUSTOM INSTRUCTIONS6.3.2 Data PathThe data path unit should defined in theor1200_vlx_dpmodule, a stub can be found infile or12

6.5. SOFTWARE IMPLEMENTATION65purpose register space in group 24 (the gr oup is selected with bit 15 - 11 of a specialpurpose register address), the t

66CHAPTER 6. LAB TASK 4 - CUSTOM INSTRUCTIONS: o u t p u t o p e r a n d s: i n p u t o p e r a n d s: l i s t o f c l o b b e r e d r e g i s t e r s

6.5. SOFTWARE IMPLEMENTATION676.5.2 Integration into jpegfilesWhen you find that the hardware is working and you have written some test programsto verif

68CHAPTER 6. LAB TASK 4 - CUSTOM INSTRUCTIONSCode Name Explanation Stands Alone0xF F C0 SOF0Start of frame for baseline coded pictures. No0xF F E0 APP

6.8. WHAT TO INCLUDE IN THE LAB REPORT69• How to read from a special purpose registers with address 0xC000 in C-code:asm volatile("l.mfspr %0,%1,

Chapter 1The system1.1 IntroductionThis text is intended as a laboratory compendium for the course TSEA44 ComputerHardware - a System On a Chip. We be

70CHAPTER 6. LAB TASK 4 - CUSTOM INSTRUCTIONS• What was bad?• What can we improve for the next year?• Do you have any other ideas for this course?• Di

Bibliography[1] Xilinx Virtex-II Development Kit, www.avnet.com[2] Communications/Memory Module User’s Guide, www.avnet.com[3] Per Karlström,Mikael An

72BIBLIOGRAPHY

Appendix AOpen RISC Reference PlatformThis is the ORP standard memory map. The actual memory map for our system is insection 1.4.1.A.1 Address mapStar

74APPENDIX A. OPEN RISC REFERENCE PLATFORMA.2 InterruptsNumber Peripheral0 Reserved1 Reserved2 UART16550 Controller3 General-Purpose I/O4 Ethernet Con

Appendix BThe Wishbone specificationB.1 IntroductionThe Wishbone specification basically dictates the interfaces and how they should behave. Themethod o

76APPENDIX B. THE WISHBONE SPECIFICATIONTable B.1: Wishbone signals (named from the master side).Name Direction Width Descriptionadr M->S 32 Addres

B.3. WISHBONE CLASSICAL CYCLES77B.2.6 cycThis signal indicates that a valid bus cycle is in progress. This signal should be asserted for theduration o

78APPENDIX B. THE WISHBONE SPECIFICATION−WSS−CLK_IADR_ODAT_ODAT_IWE_OSEL_OSTB_OCYC_OACK_IVALIDVALIDVALIDMaster signalsFigure B.3: Wishbone classical s

B.5. SYSTEM VERILOG INTERFACE79Table B.2: cti and bte signal valuesSignal group Value Descriptioncti 000 Classic cycle001 Constant address burst cycle

8CHAPTER 1. THE SYST EMFigure 1.1: Block diagram of the Avnet main board.1.2 Hardware1.2.1 Virtex-II Development boardIn the course we will use a deve

80APPENDIX B. THE WISHBONE SPECIFICATIONd a t _ t d a t _ i ; / / r e a d d a t a bu sl o g i c s t b ; / / s t r o b el o g i c cy c ; / / c y c l e

Appendix CTips & TrixIn this appendix we have collected a number of tips and trix that you might find useful.• If you encounter some weird problems

82APPENDIX C. TIPS & TRIX• If you cannot find certain commands, make sure that the following commands arepresent in your .bashrc:source /opt/Xilinx

1.2. HARDWARE9AvBus Connectors (x2)1MByteSRAM (x32)CypressCY7C1041V3316MBytesFLASH (x32)MicronMT28F640J3A64MBytesSDRAM (x32)MicronMT48LC16M16A2PCMCIA1