DOCSLIB.ORG

DRAM: Architectures, Interfaces, and Systems a Tutorial

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
DRAM: Architectures, Interfaces, and Systems a Tutorial DRAM TUTORIAL ISCA 2002 Bruce Jacob David Wang University of Maryland DRAM: Architectures, DRAM: why bother? (i mean, Interfaces, and Systems besides the “memory wall” thing? ... is it just a performance issue?) think about embedded systems: think cellphones, think printers, A Tutorial think switches ... nearly every embedded product that used to be expensive is now cheap. why? Bruce Jacob and David Wang for one thing, rapid turnover from high performance to obsolescence guarantees generous supply of CHEAP, HIGH-PERFORMANCE Electrical & Computer Engineering Dept. embedded processors to suit nearly any design need. University of Maryland at College Park what does the “memory wall” mean in this context? perhaps it will take longer for a high- performance design to become http://www.ece.umd.edu/~blj/DRAM/ obsolete? UNIVERSITY OF MARYLAND DRAM TUTORIAL ISCA 2002 Outline Bruce Jacob David Wang • University of Basics Maryland • DRAM Evolution: Structural Path NOTE • Advanced Basics • DRAM Evolution: Interface Path • Future Interface Trends & Research Areas • Performance Modeling: Architectures, Systems, Embedded Break at 10 a.m. — Stop us or starve DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of DRAM ORGANIZATION Maryland first off -- what is DRAM? an DRAM array of storage elements (capacitor-transistor pairs) Storage element Column Decoder “DRAM” is an acronym (explain) why “dynamic”? (capacitor) Word Line Data In/Out Sense Amps - capacitors are not perfect ... need recharging Buffers ... Bit Lines... - very dense parts; very small; Bit Line capactiros have very little charge ... thus, the bit lines are charged up to 1/2 voltage level and the ssense amps detect the Memory minute change on the lines, then recover the full signal d Lines ... Array or w Decoder W .. Switching Ro element DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of BUS TRANSMISSION Maryland so how do you interact with this DRAM thing? let’s look at a traditional organization first ... CPU connects to a memory controller Column Decoder that connects to the DRAM itself. let’s look at a read operation Data In/Out Sense Amps Buffers MEMORY ... Bit Lines... CPU BUS CONTROLLER Memory d Lines ... Array or w Decoder W .. Ro DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of [PRECHARGE and] ROW ACCESS Maryland at this point, all but lines are attt DRAM the 1/2 voltage level. the read discharges the Column Decoder capacitors onto the bit lines ... this pulls the lines just a little bit high or a little bit low; the sense Data In/Out Sense Amps amps detect the change and recover the full signal Buffers ... Bit Lines... the read is destructive -- the MEMORY capacitors have been CPU BUS CONTROLLER discharged ... however, when the sense amps pull the lines to the full logic-level (either high or Memory low), the transistors are kept open and so allow their attached d Lines ... Array or capacitors to become recharged w Decoder W .. (if they hold a ‘1’ value) AKA: OPEN a DRAM Page/Row . Ro or ACT (Activate a DRAM Page/Row) or RAS (Row Address Strobe) DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of COLUMN ACCESS Maryland once the data is valid on ALL of DRAM the bit lines, you can select a subset of the bits and send them to the output buffers ... CAS Column Decoder picks one of the bits big point: cannot do another Data In/Out Sense Amps RAS or precharge of the lines until finished reading the column Buffers data ... can’t change the values ... Bit Lines... on the bit lines or the output of MEMORY the sense amps until it has been CPU BUS CONTROLLER read by the memory controller Memory d Lines ... Array or w Decoder W .. Ro READ Command or CAS: Column Address Strobe DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of DATA TRANSFER Maryland then the data is valid on the data DRAM bus ... depending on what you are using for in/out buffers, you might be able to overlap a litttle Column Decoder or a lot of the data transfer with the next CAS to the same page (this is PAGE MODE) Data In/Out Sense Amps Buffers MEMORY ... Bit Lines... CPU BUS CONTROLLER Memory d Lines ... Array or w Decoder Data Out W .. Ro ... with optional additional CAS: Column Address Strobe note: page mode enables overlap with CAS DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of BUS TRANSMISSION Maryland NOTE DRAM Column Decoder Data In/Out Sense Amps Buffers MEMORY ... Bit Lines... CPU BUS CONTROLLER Memory d Lines ... Array or w Decoder W .. Ro DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang F University of Maryland DRAM CPU Mem Controller E DRAM “latency” isn’t 1 deterministic because of CAS or RAS+CAS, and there may be A significant queuing delays within the CPU and the memory controller B Each transaction has some D E2/E3 overhead. Some types of C overhead cannot be pipelined. This means that in general, longer bursts are more efficient. A: Transaction request may be delayed in Queue B: Transaction request sent to Memory Controller C: Transaction converted to Command Sequences (may be queued) D: Command/s Sent to DRAM E1: Requires only a CAS or E2: Requires RAS + CAS or E3: Requires PRE + RAS + CAS F: Transaction sent back to CPU “DRAM Latency” = A + B + C + D + E + F DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of PHYSICAL ORGANIZATION Maryland NOTE x8 DRAM x2 DRAM x4 DRAM Column Decoder Column Decoder Column Decoder Data Sense Amps Data Sense Amps Data Sense Amps Buffers Buffers Buffers ... Bit Lines... ... Bit Lines... ... Bit Lines... Memory . Memory . Memory . Array . Array . Array w Decoder w Decoder w Decoder Ro Ro Ro x2 DRAM x4 DRAM x8 DRAM This is per bank … Typical DRAMs have 2+ banks DRAM TUTORIAL ISCA 2002 Basics Bruce Jacob David Wang University of Read Timing for Conventional DRAM Maryland let’s look at the interface another way .. the say the data sheets portray it. RAS Row Access [explain] main point: the RAS\ and CAS\ Column Access signals directly control the CAS latches that hold the row and Data Transfer column addresses ... Address Row Column Row Column Address Address Address Address DQ Valid Valid Dataout Dataout DRAM TUTORIAL ISCA 2002 DRAM Evolutionary Tree Bruce Jacob David Wang . University of . Maryland . MOSYS since DRAM’s inception, there have been a stream of changes to the design, from FPM to EDO Structural to Burst EDO to SDRAM. the Modifications changes are largely structural modifications -- nimor -- that Targeting target THROUGHPUT. FCRAM Latency [discuss FPM up to SDRAM Conventional Everything up to and including DRAM $ SDRAM has been relatively (Mostly) Structural Modifications inexpensive, especially when considering the pay-off (FPM Targeting Throughput VCDRAM was essentially free, EDO cost a latch, PBEDO cost a counter, SDRAM cost a slight re-design). however, we’re run out of “free” ideas, and now all changes are considered expensive ... thus there is no consensus on new directions and myriad of choices FPM EDO P/BEDO SDRAM ESDRAM has appeared [ do LATENCY mods starting Interface Modifications with ESDRAM ... and then the INTERFACE mods ] Targeting Throughput Rambus, DDR/2 Future Trends DRAM TUTORIAL ISCA 2002 DRAM Evolution Bruce Jacob David Wang University of Read Timing for Conventional DRAM Maryland Row Access NOTE Column Access Transfer Overlap Data Transfer RAS CAS Address Row Column Row Column Address Address Address Address DQ Valid Valid Dataout Dataout DRAM TUTORIAL ISCA 2002 DRAM Evolution Bruce Jacob David Wang University of Read Timing for Fast Page Mode Maryland Row Access FPM aallows you to keep th esense amps actuve for multiple Column Access CAS commands ... Transfer Overlap much better throughput Data Transfer problem: cannot latch a new RAS value in the column address buffer until the read-out of the data is complete CAS Address Row Column Column Column Address Address Address Address DQ Valid Valid Valid Dataout Dataout Dataout DRAM TUTORIAL ISCA 2002 DRAM Evolution Bruce Jacob David Wang University of Read Timing for Extended Data Out Maryland Row Access solution to that problem -- instead of simple tri-state Column Access buffers, use a latch as well. Transfer Overlap by putting a latch after the column mux, the next column Data Transfer address command can begin sooner RAS CAS Address Row Column Column Column Address Address Address Address DQ Valid Valid Valid Dataout Dataout Dataout DRAM TUTORIAL ISCA 2002 DRAM Evolution Bruce Jacob David Wang University of Read Timing for Burst EDO Maryland Row Access by driving the col-addr latch from an internal counter rather than Column Access an external signal, the minimum cycle time for driving the output Transfer Overlap bus was reduced by roughly 30% Data Transfer RAS CAS Address Row Column Address Address DQ Valid Valid Valid Valid Data Data Data Data DRAM TUTORIAL ISCA 2002 DRAM Evolution Bruce Jacob David Wang University of Read Timing for Pipeline Burst EDO Maryland Row Access “pipeline” refers to the setting up of the read pipeline ... first CAS\ Column Access toggle latches the column address, all following CAS\ Transfer Overlap toggles drive data out onto the bus. therefore data stops coming Data Transfer when the memory controller stops toggling CAS\ RAS CAS Address Row Column Address Address DQ Valid Valid Valid Valid Data Data Data Data DRAM TUTORIAL ISCA 2002 DRAM Evolution Bruce Jacob David Wang University of Read Timing for Synchronous DRAM Maryland Row Access Clock main benefit: frees up the CPU or memory controller from Column Access having to control the DRAM’s internal latches directly ... the RAS Transfer Overlap controller/CPU can go off and do other things during the idle Data Transfer cycles instead of “wait” ..
Load more

Recommended publications
  • Modeling System Signal Integrity Uncertainty Considerations
    WHITE PAPER Intel® FPGA Modeling System Signal Integrity Uncertainty Considerations Authors Abstract Ravindra Gali This white paper describes signal integrity (SI) mechanisms that cause system-level High-Speed I/O Applications timing uncertainty and how these mechanisms are modeled in the Intel® Quartus® Engineering, Intel® Corporation Prime software Timing Analyzer to achieve timing closure for external memory interface designs. Zhi Wong By using the Intel Quartus Prime software to achieve timing closure for external High-Speed I/O Applications memory interfaces, a designer does not need to allocate a separate SI timing Engineering, Intel Corporation budget to account for simultaneous switching output (SSO), simultaneous Navid Azizi switching input (SSI), intersymbol interference (ISI), and board-level crosstalk for Software Engineeringr flip-chip device families such as Stratix® IV and Arria® II FPGAs for typical user Intel Corporation implementation of external memory interfaces following good board design practices. John Oh High-Speed I/O Applications Introduction Engineering, Intel Corporation The widening performance gap between FPGAs, microprocessors, and memory Arun VR devices, along with the growth of memory-intensive applications, are driving the Memory I/O Applications Engineering, need for faster memory technologies. This push to higher bandwidths has been Intel Corporation accompanied by an increase in the signal count and the signaling rates of FPGAs and memory devices. In order to attain faster bandwidths, device makers continue to reduce the supply voltage. Initially, industry-standard DIMMs operated at 5 V. However, due to improvements in DRAM storage density, the operating voltage was decreased to 3.3 V (SDR), then to 2.5V (DDR), 1.8 V (DDR2), 1.5 V (DDR3), and 1.35 V (DDR3) to allow the memory to run faster and consume less power.
    [Show full text]
  • Memory & Devices
    Memory & Devices Memory • Random Access Memory (vs. Serial Access Memory) • Different flavors at different levels – Physical Makeup (CMOS, DRAM) – Low Level Architectures (FPM,EDO,BEDO,SDRAM, DDR) • Cache uses SRAM: Static Random Access Memory – No refresh (6 transistors/bit vs. 1 transistor • Main Memory is DRAM: Dynamic Random Access Memory – Dynamic since needs to be refreshed periodically (1% time) – Addresses divided into 2 halves (Memory as a 2D matrix): • RAS or Row Access Strobe • CAS or Column Access Strobe Random-Access Memory (RAM) Key features – RAM is packaged as a chip. – Basic storage unit is a cell (one bit per cell). – Multiple RAM chips form a memory. Static RAM (SRAM) – Each cell stores bit with a six-transistor circuit. – Retains value indefinitely, as long as it is kept powered. – Relatively insensitive to disturbances such as electrical noise. – Faster and more expensive than DRAM. Dynamic RAM (DRAM) – Each cell stores bit with a capacitor and transistor. – Value must be refreshed every 10-100 ms. – Sensitive to disturbances. – Slower and cheaper than SRAM. Semiconductor Memory Types Static RAM • Bits stored in transistor “latches” à no capacitors! – no charge leak, no refresh needed • Pro: no refresh circuits, faster • Con: more complex construction, larger per bit more expensive transistors “switch” faster than capacitors charge ! • Cache Static RAM Structure 1 “NOT ” 1 0 six transistors per bit 1 0 (“flip flop”) 0 1 0/1 = example 0 Static RAM Operation • Transistor arrangement (flip flop) has 2 stable logic states • Write 1. signal bit line: High à 1 Low à 0 2. address line active à “switch” flip flop to stable state matching bit line • Read no need 1.
    [Show full text]
  • DDR and DDR2 SDRAM Controller Compiler User Guide
    DDR and DDR2 SDRAM Controller Compiler User Guide 101 Innovation Drive Software Version: 9.0 San Jose, CA 95134 Document Date: March 2009 www.altera.com Copyright © 2009 Altera Corporation. All rights reserved. Altera, The Programmable Solutions Company, the stylized Altera logo, specific device designations, and all other words and logos that are identified as trademarks and/or service marks are, unless noted otherwise, the trademarks and service marks of Altera Corporation in the U.S. and other countries. All other product or service names are the property of their respective holders. Altera products are protected under numerous U.S. and foreign patents and pending ap- plications, maskwork rights, and copyrights. Altera warrants performance of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information, product, or service described herein except as expressly agreed to in writing by Altera Corporation. Altera customers are advised to obtain the latest version of device specifications before relying on any published information and before placing orders for products or services. UG-DDRSDRAM-10.0 Contents Chapter 1. About This Compiler Release Information . 1–1 Device Family Support . 1–1 Features . 1–2 General Description . 1–2 Performance and Resource Utilization . 1–4 Installation and Licensing . 1–5 OpenCore Plus Evaluation . 1–6 Chapter 2. Getting Started Design Flow . 2–1 SOPC Builder Design Flow . 2–1 DDR & DDR2 SDRAM Controller Walkthrough .
    [Show full text]
  • AXP Internal 2-Apr-20 1
    2-Apr-20 AXP Internal 1 2-Apr-20 AXP Internal 2 2-Apr-20 AXP Internal 3 2-Apr-20 AXP Internal 4 2-Apr-20 AXP Internal 5 2-Apr-20 AXP Internal 6 Class 6 Subject: Computer Science Title of the Book: IT Planet Petabyte Chapter 2: Computer Memory GENERAL INSTRUCTIONS: • Exercises to be written in the book. • Assignment questions to be done in ruled sheets. • You Tube link is for the explanation of Primary and Secondary Memory. YouTube Link: ➢ https://youtu.be/aOgvgHiazQA INTRODUCTION: ➢ Computer can store a large amount of data safely in their memory for future use. ➢ A computer’s memory is measured either in Bits or Bytes. ➢ The memory of a computer is divided into two categories: Primary Memory, Secondary Memory. ➢ There are two types of Primary Memory: ROM and RAM. ➢ Cache Memory is used to store program and instructions that are frequently used. EXPLANATION: Computer Memory: Memory plays a very important role in a computer. It is the basic unit where data and instructions are stored temporarily. Memory usually consists of one or more chips on the mother board, or you can say it consists of electronic components that store instructions waiting to be executed by the processor, data needed by those instructions, and the results of processing the data. Memory Units: Computer memory is measured in bits and bytes. A bit is the smallest unit of information that a computer can process and store. A group of 4 bits is known as nibble, and a group of 8 bits is called byte.
    [Show full text]
  • Solving High-Speed Memory Interface Challenges with Low-Cost Fpgas
    SOLVING HIGH-SPEED MEMORY INTERFACE CHALLENGES WITH LOW-COST FPGAS A Lattice Semiconductor White Paper May 2005 Lattice Semiconductor 5555 Northeast Moore Ct. Hillsboro, Oregon 97124 USA Telephone: (503) 268-8000 www.latticesemi.com Introduction Memory devices are ubiquitous in today’s communications systems. As system bandwidths continue to increase into the multi-gigabit range, memory technologies have been optimized for higher density and performance. In turn, memory interfaces for these new technologies pose stiff challenges for designers. Traditionally, memory controllers were embedded in processors or as ASIC macrocells in SoCs. With shorter time-to-market requirements, designers are turning to programmable logic devices such as FPGAs to manage memory interfaces. Until recently, only a few FPGAs supported the building blocks to interface reliably to high-speed, next generation devices, and typically these FPGAs were high-end, expensive devices. However, a new generation of low-cost FPGAs has emerged, providing the building blocks, high-speed FPGA fabric, clock management resources and the I/O structures needed to implement next generation DDR2, QDR2 and RLDRAM memory controllers. Memory Applications Memory devices are an integral part of a variety of systems. However, different applications have different memory requirements. For networking infrastructure applications, the memory devices required are typically high-density, high-performance, high-bandwidth memory devices with a high degree of reliability. In wireless applications, low-power memory is important, especially for handset and mobile devices, while high-performance is important for base-station applications. Broadband access applications typically require memory devices in which there is a fine balance between cost and performance.
    [Show full text]
  • Semiconductor Memories
    Semiconductor Memories Prof. MacDonald Types of Memories! l" Volatile Memories –" require power supply to retain information –" dynamic memories l" use charge to store information and require refreshing –" static memories l" use feedback (latch) to store information – no refresh required l" Non-Volatile Memories –" ROM (Mask) –" EEPROM –" FLASH – NAND or NOR –" MRAM Memory Hierarchy! 100pS RF 100’s of bytes L1 1nS SRAM 10’s of Kbytes 10nS L2 100’s of Kbytes SRAM L3 100’s of 100nS DRAM Mbytes 1us Disks / Flash Gbytes Memory Hierarchy! l" Large memories are slow l" Fast memories are small l" Memory hierarchy gives us illusion of large memory space with speed of small memory. –" temporal locality –" spatial locality Register Files ! l" Fastest and most robust memory array l" Largest bit cell size l" Basically an array of large latches l" No sense amps – bits provide full rail data out l" Often multi-ported (i.e. 8 read ports, 2 write ports) l" Often used with ALUs in the CPU as source/destination l" Typically less than 10,000 bits –" 32 32-bit fixed point registers –" 32 60-bit floating point registers SRAM! l" Same process as logic so often combined on one die l" Smaller bit cell than register file – more dense but slower l" Uses sense amp to detect small bit cell output l" Fastest for reads and writes after register file l" Large per bit area costs –" six transistors (single port), eight transistors (dual port) l" L1 and L2 Cache on CPU is always SRAM l" On-chip Buffers – (Ethernet buffer, LCD buffer) l" Typical sizes 16k by 32 Static Memory
    [Show full text]
  • Access Order and Effective Bandwidth for Streams on a Direct Rambus Memory Sung I
    Access Order and Effective Bandwidth for Streams on a Direct Rambus Memory Sung I. Hong, Sally A. McKee†, Maximo H. Salinas, Robert H. Klenke, James H. Aylor, Wm. A. Wulf Dept. of Electrical and Computer Engineering †Dept. of Computer Science University of Virginia University of Utah Charlottesville, VA 22903 Salt Lake City, Utah 84112 Abstract current DRAM page forces a new page to be accessed. The Processor speeds are increasing rapidly, and memory speeds are overhead time required to do this makes servicing such a request not keeping up. Streaming computations (such as multi-media or significantly slower than one that hits the current page. The order of scientific applications) are among those whose performance is requests affects the performance of all such components. Access most limited by the memory bottleneck. Rambus hopes to bridge the order also affects bus utilization and how well the available processor/memory performance gap with a recently introduced parallelism can be exploited in memories with multiple banks. DRAM that can deliver up to 1.6Gbytes/sec. We analyze the These three observations — the inefficiency of traditional, performance of these interesting new memory devices on the inner dynamic caching for streaming computations; the high advertised loops of streaming computations, both for traditional memory bandwidth of Direct Rambus DRAMs; and the order-sensitive controllers that treat all DRAM transactions as random cacheline performance of modern DRAMs — motivated our investigation of accesses, and for controllers augmented with streaming hardware. a hardware streaming mechanism that dynamically reorders For our benchmarks, we find that accessing unit-stride streams in memory accesses in a Rambus-based memory system.
    [Show full text]
  • AN-371, Interfacing RC32438 with DDR SDRAM Memory
    Interfacing the RC32438 with Application Note DDR SDRAM Memory AN-371 By Kasi Chopperla and Harold Gomard Notes Revision History July 3, 2003: Initial publication. October 23, 2003: Added DDR Loading section. October 29, 2003: Added Passive DDR Termination Scheme for Lightly Loaded Systems section. Introduction The RC32438 is a high performance, general-purpose, integrated processor that incorporates a 32-bit CPU core with a number of peripherals including: – A memory controller supporting DDR SDRAM – A separate memory/peripheral controller interfacing to EPROM, flash and I/O peripherals – Two Ethernet MACs – 32-bit v.2.2 compatible PCI interface – Other generic modules, such as two serial ports, I2C, SPI, interrupt controller, GPIO, and DMA. Internally, the RC32438 features a dual bus system. The CPU bus interfaces directly to the memory controller over a 32-bit bus running at CPU pipeline speed. Since it is not possible to run an external bus at the speed of the CPU, the only way to provide data fast enough to keep the CPU from starving is to use either a wider DRAM or a double data rate DRAM. As the memory technology used on PC motherboards transitions from SDRAM to DDR, the steep ramp- up in the volume of DDR memory being shipped is driving pricing to the point where DDR-based memory subsystems are primed to drop below those of traditional SDRAM modules. Anticipating this, IDT opted to incorporate a double data rate DRAM interface on its RC32438 integrated processor. DDR memory requires a VDD/2 reference voltage and combination series and parallel terminations that SDRAM does not have.
    [Show full text]
  • Design and Implementation of High Speed DDR SDRAM Controller on FPGA
    International Journal of Engineering Research & Technology (IJERT) ISSN: 2278-0181 Vol. 4 Issue 07, July-2015 Design and Implementation of High Speed DDR SDRAM Controller on FPGA Veena H K Dr. A H Masthan Ali M.Tech Student, Department of ECE, Associate Professor, Department of ECE, Sambhram Institute of Technology, Bangalore, Sambhram Institute of Technology, Bangalore, Karnataka, India Karnataka, India Abstract — The dedicated memory controller is important is and column. To point to a location in the memory these three the applications in high end applications where it doesn’t are mandatory. The ACTIVE command signal along with contains microprocessors. Command signals for memory registered address bits point to the specific bank and row to be refresh, read and write operation and SDRAM initialisation has accessed. And column is given by READ/WRITE signal been provided by memory controller. Our work will focus on along with the registered address bits that shall point to a FPGA implementation of Double Data Rate (DDR) SDRAM location for burst access. The DDR SDRAM interface makes controller. The DDR SDRAM controller is located in between higher transfer rates possible compared to single data rate the DDR SDRAM and bus master. The operations of DDR (SDR) SDRAM, by more firm control of the timing of the SDRAM controller is to simplify the SDRAM command data and clock signals. Implementations frequently have to use interface to the standard system read/ write interface and also schemes such as phase-locked loops (PLL) and self- optimization of the access time of read/write cycle. The proposed design will offers effective power utilization, reduce the gate calibration to reach the required timing precision [1][2].
    [Show full text]
  • Bioshock Infinite
    SOUTH AFRICA’S LEADING GAMING, COMPUTER & TECHNOLOGY MAGAZINE VOL 15 ISSUE 10 Reviews Call of Duty: Black Ops II ZombiU Hitman: Absolution PC / PLAYSTATION / XBOX / NINTENDO + MORE The best and wors t of 2012 We give awards to things – not in a traditional way… BioShock Infi nite Loo k! Up in the sky! Editor Michael “RedTide“ James [email protected] Contents Features Assistant editor 24 THE BEST AND WORST OF 2012 Geoff “GeometriX“ Burrows Regulars We like to think we’re totally non-conformist, 8 Ed’s Note maaaaan. Screw the corporations. Maaaaan, etc. So Staff writer 10 Inbox when we do a “Best of [Year X]” list, we like to do it Dane “Barkskin “ Remendes our way. Here are the best, the worst, the weirdest 14 Bytes and, most importantly, the most memorable of all our Contributing editor 41 home_coded gaming experiences in 2012. Here’s to 2013 being an Lauren “Guardi3n “ Das Neves 62 Everything else equally memorable year in gaming! Technical writer Neo “ShockG“ Sibeko Opinion 34 BIOSHOCK INFINITE International correspondent How do you take one of the most infl uential, most Miktar “Miktar” Dracon 14 I, Gamer evocative experiences of this generation and make 16 The Game Stalker it even more so? You take to the skies, of course. Contributors 18 The Indie Investigator Miktar’s played a few hours of Irrational’s BioShock Rodain “Nandrew” Joubert 20 Miktar’s Meanderings Infi nite, and it’s left him breathless – but fi lled with Walt “Shryke” Pretorius 67 Hardwired beautiful, descriptive words. Go read them. Miklós “Mikit0707 “ Szecsei 82 Game Over Pippa
    [Show full text]
  • 10. Implementing Double Data Rate I/O Signaling in Cyclone Devices
    10. Implementing Double Data Rate I/O Signaling in Cyclone Devices C51010-1.2 Introduction Double data rate (DDR) transmission is used in many applications where fast data transmission is needed, such as memory access and first-in first-out (FIFO) memory structures. DDR uses both edges of a clock to transmit data, which facilitates data transmission at twice the rate of a single data rate (SDR) architecture using the same clock speed. This method also reduces the number of I/O pins required to transmit data. This chapter shows implementations of a double data rate I/O interface using Cyclone® devices. Cyclone devices support DDR input, DDR output, and bidirectional DDR signaling. For more information on using Cyclone devices in applications with DDR SDRAM and FCRAM memory devices, refer to “DDR Memory Support” on page 10–4. Double Data The DDR input implementation shown in Figure 10–1 uses four internal logic element (LE) registers located in the logic array block (LAB) adjacent Rate Input to the DDR input pin. The DDR data is fed to the first two of four registers. One register captures the DDR data present during the rising edge of the clock. The second register captures the DDR data present during the falling edge of the clock. Figure 10–1. Double Data Rate Input Implementation DFF DFF PRN PRN ddr DQ DQ ddr_out_h p_edge_reg ddr_h_sync_reg DFF DFF PRN PRN DQ DQ ddr_out_l NOT n_edge_reg ddr_l_sync_reg clk Altera Corporation 10–1 May 2008 Preliminary Cyclone Device Handbook, Volume 1 The third and fourth registers synchronize the two data streams to the rising edge of the clock.
    [Show full text]
  • DDR400/333/266, Dual DDR, RDRAM 16 Bit and 32 Bit, SDRAM
    Ace’s Hardware Granite Bay: Memory Technology Shootout Granite Bay: Memory Technology Shootout By Johan De Gelas – December 2002 Dual-Channel DDR SDRAM Arrives for the Pentium 4 DDR400/333/266, Dual DDR, RDRAM 16 bit and 32 bit, SDRAM... almost every memory technology on the market is available for the Pentium 4 platform. One of our previous technical articles discussed the advantages and disadvantages of the different architectures of Rambus and SDRAM based memory technology such as DDR and DDR-II. In this article, we will investigate how the different memory technologies and their supporting chipsets compare on the test bench. The following motherboards were tested: • The ASUS P4T533 features the i850E chipset and 32 bit RDRAM • The ASUS P4T533-C comes with the same chipset but uses two channels of 16 bit RIMMs • The MSI 648 Max comes with SIS 648 chipset which unofficially supports DDR400 • The MSI i845PE comes with Intel's newest i845 chipset, which officially support DDR333 • The Tyan Trinity 7205 and MSI GNB Max feature the Dual DDR266 Granite Bay chipset We are well aware that there have already many tests with Pentium 4 chipsets, Granite Bay included. So why bother to publish another on Ace’s Hardware? The focus of this article is on the memory technology supported by these chipsets. This article will offer you a insight in how the different memory technologies compare in a wide variety of applications. We'll investigate in depth what the advantages and disadvantages are of each memory technology and try to find out what are the reasons behind this.
    [Show full text]