2016 COMPUTING GUIDE Best of the Breed Table of Contents

Total Page:16

File Type:pdf, Size:1020Kb

2016 COMPUTING GUIDE Best of the Breed Table of Contents 2016 COMPUTING GUIDE Best of the Breed Table of Contents Computing Business Unit Storage Monitors Computing Solutions ................1 Kingston Flash ................43 LG Monitors ...............78 Kingston SD Cards ................47 Keyboards and Mice Kingston SSD ................50 Graphics Cards Logitech Mice ...............81 HyperX ................53 EVGA Nvidia Graphics Cards ................2 Logitech Keyboards ...............84 WD Hard Drives ................55 Genius Mice ...............91 Leadtek Nvidia Graphics Cards ................6 WD External Drives ................57 Sapphire Graphics Cards ................11 Optical Drives ................58 Genius Keyboards ...............93 AMD Graphics Cards ................16 Kingston Mouse pads ...............95 Motherboards Software Audio and Webcams ASRock Intel Motherboards ................19 Office ................59 Kingston Headsets ...............96 ASRock AMD Motherboards ................30 Windows ................60 Logitech Audio ...............98 Genius Audio ...............103 Processors Chassis and PSU Notebooks and Tablets Intel Processors ................31 SAMA Chassis ................61 AMD Processors ................35 BitFenix Chassis ................62 Lenovo Tablets ...............106 BitFenix Cooling ................66 Memory Lenovo Notebooks ...............110 BitFenix Accessories ................69 Kingston ValueRAM ................38 HyperX ................41 EVGA Power Supplies ................74 | Computing Solutions PERIPHERALS SYSTEMS COMPONENTS Keyboards & Audio & Notebooks & Graphics Memory Monitors Mice Webcams Chassis PSU Tablets Storage CPU Motherboards Cards EVGA® GeForce® GTX 1080 N1080-8GBFTW-6286 The EVGA GeForce GTX 1080 featuring EVGA ACX 3.0 cooling has arrived. This new graphics card features NVIDIA’s new Graphics Engine: NVIDIA GeForce GTX 1080 “Pascal” graphics processor which is the most advanced gaming Video Memory: 8192 MB, 256 bit GDDR5X GPU ever created. This breakthrough GPU delivers industry-lead- ing performance, innovative new gaming technologies, and Engine Clock: GPU Boost Clock : 1860 MHz immersive, next-gen VR. GPU Base Clock : 1721 MHz These cards also feature EVGA ACX 3.0 cooling technology. EVGA ACX 3.0 once again brings new and exciting features to the award winning EVGA ACX cooling technology. SHP 3.0 gives Resolution: Digital Max Resolution: 7680x4320 increased heatpipes and copper contact area for cooler operation, and optimized fan curve for even quieter gaming. Of course, ACX Interface: PCI-E 3.0 16x 3.0 coolers also feature optimized swept fan blades, double ball DVI-D, DisplayPort, DisplayPort, DisplayPort, HDMI bearings and an extreme low power motor, delivering more air flow with less power, unlocking additional power for the GPU. NVIDIA Graphics Cards Model EVGA® NVIDIA GeForce® GT 710 EVGA® NVIDIA GeForce® GT 710 EVGA® NVIDIA GeForce® GT 730 EVGA® NVIDIA GeForce® GTX 750 Ti EVGA® NVIDIA GeForce® GTX 950 Product Code N710-1GB-2710 N710-2GB-2712 N730-2GB-2732 N750TI-2GD5-3751 N950-2GBSC-2951 1.0GB DDR3, 64-Bit Memory Bus, 2.0GB GDDR3, 64-Bit Memory Bus, 2.0GB GDDR3, 128-Bit Memory Bus, 2.0GB GDDR5, 128-Bit Memory Bus, 2.0GB GDDR5, 128-Bit Memory Bus, Product Description PCI-Express 2.0 PCI-Express 2.0 PCI-Express x16 3.0 PCI-Express x16 3.0 PCI-Express x16 3.0 1x D-SUB Port, 1x DVI-I Port, 1x HDMI 1x D-SUB Port, 1x DVI-I Port, 1x 1x D-SUB Port, 1x DVI-I Port, 1x 1x DVI-I Port, 1x HDMI 1.4a Port, 1x DVI-I Port, 1x HDMI 1.4a Port, Display Configuration Port HDMI Port HDMI Port 1x DisplayPort 1.2 3x DisplayPort 1.2 Clock Speeds 954MHz 954MHz 700MHz 1020MHz (Base) & 1085MHz (Boost) 1152MHz (Base) & 1342MHz (Boost) Memory 1800MHz 1800MHz 1400MHz 5400MHz 6610MHz Nvidia Graphics Cards Model EVGA® NVIDIA GeForce® GTX 1060 EVGA® NVIDIA GeForce® GTX 1060 EVGA® NVIDIA GeForce® GTX 1060 EVGA® NVIDIA GeForce® GTX 1060 EVGA® NVIDIA GeForce® GTX 1070 Product Code N1060-3GB-6160 N1060-3GBSC-6162 N1060-6GB-6161 N1060-6GBSC-6163 N1070-8GBSC-6173 3.0GB GDDR5, 192-Bit Memory Bus, 3.0GB GDDR5, 192-Bit Memory Bus, 6.0GB GDDR5, 192-Bit Memory Bus, 6.0GB GDDR5, 192-Bit Memory Bus, 8.0GB GDDR5, 256-Bit Memory Bus, Product Description PCI-Express x16 3.0 PCI-Express x16 3.0 PCI-Express x16 3.0 PCI-Express x16 3.0 PCI-Express x16 3.0 1x DVI-D Port, 1x HDMI 2.0a Port, 3x 1x DVI-D Port, 1x HDMI 2.0a Port, 3x HDMI 2.0b, DisplayPort 1.4 and HDMI 2.0b, DisplayPort 1.4 and 1x DVI-D Port, 1x HDMI 2.0a Port, 3x Display Configuration DisplayPort 1.4 DisplayPort 1.4 Dual-Link DVI Dual-Link DVI DisplayPort 1.4 Clock Speeds 1506MHz (Base) & 1708MHz (Boost) 1607MHz (Base) & 1835MHz (Boost) 1506MHz (Base) & 1708MHz (Boost) 1607MHz (Base) & 1835MHz (Boost) 1594MHz (Base) & 1784MHz (Boost) Memory 8008MHz 8008MHz 8008MHz 8008MHz 8008MHz 3 NVIDIA Graphics Cards Model EVGA® NVIDIA GeForce® GTX 1070 EVGA® NVIDIA GeForce® GTX 1080 EVGA® NVIDIA GeForce® GTX 1080 EVGA® NVIDIA GeForce® GTX 1080 Product Code N1070-8GBFTW-6276 N1080-8GBACX-6181 N1080-8GBSC-6183 N1080-8GBFTW-6286 8.0GB GDDR5, 256-Bit Memory Bus, 8.0GB GDDR5X, 256-Bit Memory Bus, 8.0GB GDDR5X, 256-Bit Memory Bus, 8.0GB GDDR5X, 256-Bit Memory Bus, Product Description PCI-Express x16 3.0 PCI-Express x16 3.0 PCI-Express x16 3.0 PCI-Express x16 3.0 1x DVI-D Port, 1x HDMI 2.0a Port, 3x 1x DVI-D Port, 1x HDMI 2.0a Port, 3x 1x DVI-D Port, 1x HDMI 2.0a Port, 3x 1x DVI-D Port, 1x HDMI 2.0a Port, 3x Display Configuration DisplayPort 1.4 DisplayPort 1.4 DisplayPort 1.4 DisplayPort 1.4 Clock Speeds 1607MHz (Base) & 1797MHz (Boost) 1607MHz (Base) & 1733MHz (Boost) 1708MHz (Base) & 1847MHz (Boost) 1721MHz (Base) & 1860MHz (Boost) Memory 8008MHz 10000MHz 10000MHz 10000MHz 4 EVGA PRO SLI BRIDGE HB Elegant High Bandwidth Performance EVGA has once again amped up the SLI bridge with the EVGA PRO SLI BRIDGE HB. These new, completely redesigned SLI bridges come with new features and improved performance. Included is an RGBW color switcher that allows you to choose Red, Green, Blue or White color. Also, with the new High Bandwidth technology on NVIDIA GTX 10 series cards, you get double the available transfer bandwidth compared to previous generation cards. 0 Slot Spacing 1 Slot Spacing 2 Slot Spacing 100-2W-0025-LR 100-2W-0026-LR 100-2W-0027-LR Coming Soon Coming Soon Coming Soon ACCELERATE YOUR CREATIVITY NVIDIA® Quadro® INDUSTRY SOLUTIONS Design and Manufacturing Media and Entertainment Science and Medical Imaging Energy Exploration Get real interactive expression with NVIDIA® Quadro®—the world’s most powerful workstation graphics. NVIDIA Quadro M6000 LK-M6000-24GB 24.0GB GDDR5, 384-Bit Memory Bus, PCI-Express x16 3.0 4x DisplayPort 1.2, 1x DVI-DL 4096x2160 (DisplayPort) / 2560x1600 (DVI DL) / 1920x1200 (DVI SL) / 2048x1536 (D-SUB) R 399 R 399 R 399 R 399 R 399 REAL INTERACTIVE EXPRESSION NVIDIA® QUADRO™ M6000 24GB R 399 Implement New Visualise Larger Collaborate Cloud Workflows Data Sets Across Teams DOUBLE CLICK to zoom in and out NVIDIA® NVS™ 810 BRILLIANTLY SIMPLE DIGITAL SIGNAGE The NVIDIA® NVS™ 810 graphics board delivers exceptional display connectivity, cost-effective scalability, and advanced image management capabilities that make it easy to drive any kind of multi-display digital signage setup. It’s the first of its kind to offer eight display outputs, combined with the world’s most advanced GPU architecture—all in a single-slot form factor. Advanced Display Features Simultaneously drive up to eight displays when connected natively or when using DisplayPort 1.2 Multi-Stream Eight DisplayPort 1.2 outputs including Multi-Stream and HBR2 support (capable of supporting resolutions such as 4096x2160 @30Hz when all eight displays are connected) DisplayPort to VGA, DisplayPort to DVI (single-link and dual-link) and DisplayPort to HDMI cables available (resolution support based on dongle specifications) DisplayPort 1.2, HDMI, and DVI support HDCP 12-bit internal display pipeline (hardware support for 12-bit scan out on supported panels, applications and connection) Underscan/overscan compensation and hardware scaling Support for NVIDIA® Mosaic, NVIDIA® nView® multi-display technology,* All pricing advertised excludes VAT. Pricing subject to change without notice. E&OE NVIDIA® Enterprise Management Tools 9 Nvidia NVS Graphics Cards Nvidia Quadro Graphics Cards Model Leadtek® NVIDIA NVS315 Leadtek® NVIDIA NVS510 Leadtek® NVIDIA NVS810 Leadtek® NVIDIA Quadro K420 Leadtek® NVIDIA Quadro K620 Product Code LK-NVS315 LK-NVS510 LK-NVS810 LK-K420 LK-K620 1024 MB GDDR3, 64-bit, 48 CUDA Cores, 2GB GDDR3, 128-bit, 4GB GDDR3, 128-bit, 1024 (512 per 4GB GDDR5, 256-bit, 2GB GDDR3, 128-bit, 384 CUDA Product Description PCI Express x16 192 CUDA Cores GPU) Cores, PCI Express 3.0 x16 384 CUDA Cores Cores, PCI Express 2.0 x16 DisplayPort : 2560x1600 Digital: 3840x2160 4096x2160 @30Hz Resolution & Displays 1x DVI-I, 2x DisplayPort 1.2 DVI-I (1), DP 1.2 (1) DVI-I : 1920x1200 Analog: 1920x1200 4096x2160 @60Hz Nvidia Quadro Graphics Cards Model Leadtek® NVIDIA Quadro K1200 DP Leadtek® NVIDIA Quadro K1200 DVI Leadtek® NVIDIA Quadro M2000 Leadtek® NVIDIA Quadro M4000 Product Code LK-K1200-4GB LK-K1200DVI-4GB LK-M2000-4GB LK-M4000 4GB GDDR5, 128-bit, 512 CUDA Cores, 4GB GDDR5, 128-bit, 512 CUDA 4GB GDDR5, 128-bit, 768 CUDA 8GB GDDR5, 256-bit, 1664 CUDA Product Description PCI Express 2.0 x16 Cores, PCI Express 2.0 x16 Cores, PCI Express 2.0 x16 Cores, PCI Express 3.0 x16 DP 1.2: 3840 x 2160 at 60 Hz DP 1.2: 4096 × 2160 @60 Hz Resolution & Displays DisplayPort, DVI-DL, DVI-SL, D-Sub 4x DVI-D DVI-I DL: 2560 × 1600 at 60 Hz
Recommended publications
  • Small Form Factor 3D Graphics for Your Pc
    VisionTek Part# 900701 PRODUCTIVITY SERIES: SMALL FORM FACTOR 3D GRAPHICS FOR YOUR PC The VisionTek Radeon R7 240SFF graphics card offers a perfect balance of performance, features, and affordability for the gamer seeking a complete solution. It offers support for the DIRECTX® 11.2 graphics standard and 4K Ultra HD for stunning 3D visual effects, realistic lighting, and lifelike imagery. Its Short Form Factor design enables it to fit into the latest Low Profile desktops and workstations, yet the R7 240SFF can be converted to a standard ATX design with the included tall bracket. With 2GB of DDR3 memory and award-winning Graphics Core Next (GCN) architecture, and DVI-D/HDMI outputs, the VisionTek Radeon R7 240SFF is big on features and light on your wallet. RADEON R7 240 SPECS • Graphics Engine: RADEON R7 240 • Video Memory: 2GB DDR3 • Memory Interface: 128bit • DirectX® Support: 11.2 • Bus Standard: PCI Express 3.0 • Core Speed: 780MHz • Memory Speed: 800MHz x2 • VGA Output: VGA* • DVI Output: SL DVI-D • HDMI Output: HDMI (Video/Audio) • UEFI Ready: Support SYSTEM REQUIREMENTS • PCI Express® based PC is required with one X16 lane graphics slot available on the motherboard. • 400W (or greater) power supply GCN Architecture: A new design for AMD’s unified graphics processing and compute cores that allows recommended. 500 Watt for AMD them to achieve higher utilization for improved performance and efficiency. CrossFire™ technology in dual mode. • Minimum 1GB of system memory. 4K Ultra HD Support: Experience what you’ve been missing even at 1080P! With support for 3840 x • Installation software requires CD-ROM 2160 output via the HDMI port, textures and other detail normally compressed for lower resolutions drive.
    [Show full text]
  • Monte Carlo Evaluation of Financial Options Using a GPU a Thesis
    Monte Carlo Evaluation of Financial Options using a GPU Claus Jespersen 20093084 A thesis presented for the degree of Master of Science Computer Science Department Aarhus University Denmark 02-02-2015 Supervisor: Gerth Brodal Abstract The financial sector has in the last decades introduced several new fi- nancial instruments. Among these instruments, are the financial options, which for some cases can be difficult if not impossible to evaluate analyti- cally. In those cases the Monte Carlo method can be used for pricing these instruments. The Monte Carlo method is a computationally expensive al- gorithm for pricing options, but is at the same time an embarrassingly parallel algorithm. Modern Graphical Processing Units (GPU) can be used for general purpose parallel-computing, and the Monte Carlo method is an ideal candidate for GPU acceleration. In this thesis, we will evaluate the classical vanilla European option, an arithmetic Asian option, and an Up-and-out barrier option using the Monte Carlo method accelerated on a GPU. We consider two scenarios; a single option evaluation, and a se- quence of a varying amount of option evaluations. We report performance speedups of up to 290x versus a single threaded CPU implementation and up to 53x versus a multi threaded CPU implementation. 1 Contents I Theoretical aspects of Computational Finance 5 1 Computational Finance 5 1.1 Options . .7 1.1.1 Types of options . .7 1.1.2 Exotic options . .9 1.2 Pricing of options . 11 1.2.1 The Black-Scholes Partial Differential Equation . 11 1.2.2 Solving the PDE and pricing vanilla European options .
    [Show full text]
  • Deep Dive: Asynchronous Compute
    Deep Dive: Asynchronous Compute Stephan Hodes Developer Technology Engineer, AMD Alex Dunn Developer Technology Engineer, NVIDIA Joint Session AMD NVIDIA ● Graphics Core Next (GCN) ● Maxwell, Pascal ● Compute Unit (CU) ● Streaming Multiprocessor (SM) ● Wavefronts ● Warps 2 Terminology Asynchronous: Not independent, async work shares HW Work Pairing: Items of GPU work that execute simultaneously Async. Tax: Overhead cost associated with asynchronous compute 3 Async Compute More Performance 4 Queue Fundamentals 3 Queue Types: 3D ● Copy/DMA Queue ● Compute Queue COMPUTE ● Graphics Queue COPY All run asynchronously! 5 General Advice ● Always profile! 3D ● Can make or break perf ● Maintain non-async paths COMPUTE ● Profile async on/off ● Some HW won’t support async ● ‘Member hyper-threading? COPY ● Similar rules apply ● Avoid throttling shared HW resources 6 Regime Pairing Good Pairing Poor Pairing Graphics Compute Graphics Compute Shadow Render Light culling G-Buffer SSAO (Geometry (ALU heavy) (Bandwidth (Bandwidth limited) limited) limited) (Technique pairing doesn’t have to be 1-to-1) 7 - Red Flags Problem/Solution Format Topics: ● Resource Contention - ● Descriptor heaps - ● Synchronization models ● Avoiding “async-compute tax” 8 Hardware Details - ● 4 SIMD per CU ● Up to 10 Wavefronts scheduled per SIMD ● Accomplish latency hiding ● Graphics and Compute can execute simultanesouly on same CU ● Graphics workloads usually have priority over Compute 9 Resource Contention – Problem: Per SIMD resources are shared between Wavefronts SIMD executes
    [Show full text]
  • Contributions of Hybrid Architectures to Depth Imaging: a CPU, APU and GPU Comparative Study
    Contributions of hybrid architectures to depth imaging : a CPU, APU and GPU comparative study Issam Said To cite this version: Issam Said. Contributions of hybrid architectures to depth imaging : a CPU, APU and GPU com- parative study. Hardware Architecture [cs.AR]. Université Pierre et Marie Curie - Paris VI, 2015. English. NNT : 2015PA066531. tel-01248522v2 HAL Id: tel-01248522 https://tel.archives-ouvertes.fr/tel-01248522v2 Submitted on 20 May 2016 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. THESE` DE DOCTORAT DE l’UNIVERSITE´ PIERRE ET MARIE CURIE sp´ecialit´e Informatique Ecole´ doctorale Informatique, T´el´ecommunications et Electronique´ (Paris) pr´esent´eeet soutenue publiquement par Issam SAID pour obtenir le grade de DOCTEUR en SCIENCES de l’UNIVERSITE´ PIERRE ET MARIE CURIE Apports des architectures hybrides `a l’imagerie profondeur : ´etude comparative entre CPU, APU et GPU Th`esedirig´eepar Jean-Luc Lamotte et Pierre Fortin soutenue le Lundi 21 D´ecembre 2015 apr`es avis des rapporteurs M. Fran¸cois Bodin Professeur, Universit´ede Rennes 1 M. Christophe Calvin Chef de projet, CEA devant le jury compos´ede M. Fran¸cois Bodin Professeur, Universit´ede Rennes 1 M.
    [Show full text]
  • AMD APP SDK V2.8.1 Developer Release Notes
    AMD APP SDK v2.8.1 Developer Release Notes 1 What’s New in AMD APP SDK v2.8.1 1.1 New features in AMD APP SDK v2.8.1 AMD APP SDK v2.8.1 includes the following new features: Bolt: With the launch of Bolt 1.0, several new samples have been added to demonstrate the use of the features of Bolt 1.0. These features showcase the use of valuable Bolt APIs such as scan, sort, reduce and transform. Other new samples highlight the ease of porting from STL and the performance benefits achieved over equivalent STL implementations. Other samples demonstrate the different fallback options in Bolt 1.0 when no GPU is available. These options include a fallback to multicore-CPU if TBB libraries are installed, or falling all the way back to serial-CPU if needed to ensure your code runs correctly on any platform. OpenCV: AMD has been working closely with the OpenCV open source community to add heterogeneous acceleration capability to the world’s most popular computer vision library. These changes are already integrated into OpenCV and are readily available for developers who want to improve the performance and efficiency of their computer vision applications. The new samples illustrate these improvements and highlight how simple it is to include them in your application. For information on the latest OpenCV enhancements, see Harris’ blog. GCN: AMD recently launched a new Graphics Core Next (GCN) Architecture on several AMD products. GCN is based on a scalar architecture versus the VLIW vector architecture of prior generations, so carefully hand-tuned vectorization to optimize hardware utilization is no longer needed.
    [Show full text]
  • AMD Firepro™ W4100 Professional Graphics in a Class of Its Own
    AMD FirePro™ W4100 Professional Graphics In a class of its own Key Features: The AMD FirePro™ W4100 professional • Application optimizations graphics card represents a completely and certifications • AMD Graphics Core Next (GCN) new class of product – one that provides GPU architecture you with great graphics performance • Four Mini DisplayPort outputs • DisplayPort 1.2a support and display versatility while housed in a • AMD Eyefinity technology1 compact, low-power design. • 4K display resolution (up to 4096 x 2160) • 512 stream processors Increase your productivity by working across up to four high-resolution displays with AMD Eyefinity 1 • 645.1 GFLOPS peak single precision technology. Manipulate 3D models and large data sets with ease thanks to 2GB of ultrafast GDDR5 memory. With a stable driver that supports a growing list of optimized and certified applications, the • 2GB GDDR5 memory AMD FirePro W4100 is uniquely suited to provide the performance and quality you expect, and more, • 128-bit memory interface from a professional graphics card. • Up to 72GB/s memory bandwidth • PCIe® 3.0 compliant Peformance Get solid, midrange performance with the AMD FirePro W4100, delivering CAD performance that is up • OpenCL™, DirectX® and OpenGL support to 100% faster than the previous generation2. Equipped with 2GB of ultrafast GDDR5 memory with a • 50W maximum power consumption 128-bit memory interface, the AMD FirePro W4100 delivers up to 72GB/s of memory bandwidth, helping • Discreet active cooling solution improve application responsiveness to your workflows. Accelerate your 3D applications with 512 stream • Low profile single-slot form factor processors and enable more efficient data transfers between the GPU and CPU with PCIe® 3.0 support.
    [Show full text]
  • How to Sell the AMD Radeon™ HD 7790 Graphics Cards Outstanding 1080P Performance and Unbeatable Value for Gamers
    How to Sell the AMD Radeon™ HD 7790 Graphics Cards Outstanding 1080p performance and unbeatable value for gamers. Who’s it for? Gamers who want 1080p gaming and outstanding image quality at a great value. Sell it in 5 seconds. This is where high-quality 1080p gaming begins. Get ready to turn on that graphics eye-candy. With the AMD Radeon™ HD 7790 GPU, you get outstanding 1080p performance in the latest DirectX® 11 games at an unbeatable value. It offers great performance per dollar and allows you to play modern games with all the settings turned up to the max. It’s an all new chip built just for gaming featuring AMD’s latest refinement of AMD PowerTune Technology. Sell it in 60 seconds. > Outstanding 1080p performance in the latest DirectX® 11 games: The AMD Radeon™ HD 7790 Graphics card was engineered to provide superior DirectX® 11.1 performance for gamers with 1080p monitors and, being built on the Graphics Core Next Architecture, is the perfect opportunity to ready your rig for the hottest games of the year. > Unbeatable value for gamers: If you’re looking for great gaming on a budget, it doesn’t get any better than this product. In fact it is up to 21% faster than the competition.1 > Featuring an all-new AMD PowerTune Technology designed to squeeze every bit of performance out of the GPU, the AMD Radeon™ HD 7790 is engineered with intelligent, automatic overclocking to provide the most frame-rates possible. Don’t take our word for it. Here is what others are saying… “…power efficiency, its low noise levels, and the free copy of BioShock Infinite in the box…looks like we have a winning recipe from AMD.” – The Tech Report 2 “…even without BioShock Infinite coming along for the ride, the HD 7790 represents a phenomenal value.” – Hardware Canucks 3 Why it’s great..
    [Show full text]
  • AMD Graphics Core Next | June 2011 SCALABLE MULTI-TASK GRAPHICS ENGINE
    AMD GRAPHIC CORE NEXT Low Power High Performance Graphics & Parallel Compute Michael Mantor Mike Houston AMD Senior Fellow Architect AMD Fellow Architect [email protected] [email protected] At the heart of every AMD APU/GPU is a power aware high performance set of compute units that have been advancing to bring users new levels of programmability, precision and performance. AGENDA AMD Graphic Core Next Architecture .Unified Scalable Graphic Processing Unit (GPU) optimized for Graphics and Compute – Multiple Engine Architecture with Multi-Task Capabilities – Compute Unit Architecture – Multi-Level R/W Cache Architecture .What will not be discussed – Roadmaps/Schedules – New Product Configurations – Feature Rollout 3 | AMD Graphics Core Next | June 2011 SCALABLE MULTI-TASK GRAPHICS ENGINE GFX Command Processor Work Distributor Scalable Graphics Engine MC Primitive Primitive Pipe 0 Pipe n HUB & HOS HOS CS Pixel Pixel R/W MEM Pipe 0 Pipe n L2 Pipe Tessellate Tessellate Scan Scan Geometry Geometry Conversion Conversion RB RB HOS – High Order Surface RB - Render Backend Unified Shader Core CS - Compute Shader GFX - Graphics 4 | AMD Graphics Core Next | June 2011 SCALABLE MULTI-TASK GRAPHICS ENGINE PrimitiveGFX Scaling Multiple Primitive Pipelines Command ProcessorPixel Scaling Multiple Screen Partitions Multi-task graphics engine use of unified shader Work Distributor Scalable Graphics Engine MC Primitive Primitive Pipe 0 Pipe n HUB & HOS HOS CS Pixel Pixel R/W MEM Pipe 0 Pipe n L2 Pipe Tessellate Tessellate Scan Scan Geometry Geometry Conversion Conversion RB RB Unified Shader Core 5 | AMD Graphics Core Next | June 2011 MULTI-ENGINE UNIFIED COMPUTING GPU Asynchronous Compute Engine (ACE) ACE ACE GFX 0 n .
    [Show full text]
  • Real-World Design and Evaluation of Compiler-Managed GPU Redundant Multithreading ∗
    Real-World Design and Evaluation of Compiler-Managed GPU Redundant Multithreading ∗ Jack Wadden Alexander Lyashevsky§ Sudhanva Gurumurthi† Vilas Sridharan‡ Kevin Skadron University of Virginia, Charlottesville, Virginia, USA †AMD Research, Advanced Micro Devices, Inc., Boxborough, MA, USA §AMD Research, Advanced Micro Devices, Inc., Sunnyvale, CA, USA ‡ RAS Architecture, Advanced Micro Devices, Inc., Boxborough, MA, USA {wadden,skadron}@virginia.edu {Alexander.Lyashevsky,Sudhanva.Gurumurthi,Vilas.Sridharan}@amd.com Abstract Structure Size Estimated ECC Overhead Reliability for general purpose processing on the GPU Local data share 64 kB 14 kB (GPGPU) is becoming a weak link in the construction of re- Vector register file 256 kB 56 kB liable supercomputer systems. Because hardware protection Scalar register file 8 kB 1.75 kB is expensive to develop, requires dedicated on-chip resources, R/W L1 cache 16 kB 343.75 B and is not portable across different architectures, the efficiency Table 1: Reported sizes of structures in an AMD Graphics Core of software solutions such as redundant multithreading (RMT) Next compute unit [4] and estimated costs of SEC-DED ECC must be explored. assuming cache-line and register granularity protections. This paper presents a real-world design and evaluation of automatic software RMT on GPU hardware. We first describe These capabilities typically are provided by hardware. Such a compiler pass that automatically converts GPGPU kernels hardware support can manifest on large storage structures as into redundantly threaded versions. We then perform detailed parity or error-correction codes (ECC), or on pipeline logic power and performance evaluations of three RMT algorithms, via radiation hardening [19], residue execution [16], and other each of which provides fault coverage to a set of structures techniques.
    [Show full text]
  • Decoupled Vector-Fetch Architecture with a Scalarizing Compiler By
    Decoupled Vector-Fetch Architecture with a Scalarizing Compiler by Yunsup Lee A dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of Philosophy in Computer Science in the GRADUATE DIVISION of the UNIVERSITY OF CALIFORNIA, BERKELEY Committee in charge: Professor Krste Asanovic,´ Chair Professor David A. Patterson Professor Borivoje Nikolic´ Professor Paul K. Wright Spring 2016 Decoupled Vector-Fetch Architecture with a Scalarizing Compiler Copyright 2016 by Yunsup Lee 1 Abstract Decoupled Vector-Fetch Architecture with a Scalarizing Compiler by Yunsup Lee Doctor of Philosophy in Computer Science University of California, Berkeley Professor Krste Asanovic,´ Chair As we approach the end of conventional technology scaling, computer architects are forced to incorporate specialized and heterogeneous accelerators into general-purpose processors for greater energy efficiency. Among the prominent accelerators that have recently become more popular are data-parallel processing units, such as classic vector units, SIMD units, and graphics processing units (GPUs). Surveying a wide range of data-parallel architectures and their parallel program- ming models and compilers reveals an opportunity to construct a new data-parallel machine that is highly performant and efficient, yet a favorable compiler target that maintains the same level of programmability as the others. In this thesis, I present the Hwacha decoupled vector-fetch architecture as the basis of a new data-parallel machine. I reason through the design decisions while describing its programming model, microarchitecture, and LLVM-based scalarizing compiler that efficiently maps OpenCL kernels to the architecture. The Hwacha vector unit is implemented in Chisel as an accelerator at- tached to a RISC-V Rocket control processor within the open-source Rocket Chip SoC generator.
    [Show full text]
  • AMD Radeon ™ E8860 Embedded GPU
    Product Brief AMD Radeon ™ E8860 Embedded GPU SUPERIOR MULTIDISPLAY VERSATILITY The latest evolution in AMD Radeon™ embedded GPUs leverages advanced The AMD Radeon E8860 GPU provides multi-display flexibility, supporting up to five 3840x2160 @30Hz displays simultaneously in Graphics Core Next architecture, delivering clone mode and extended desktop in static screen. Competitive 5 breakthrough performance and power NVIDIA GPUs can only support up to four independent displays. efficiency gains. The AMD Radeon E8860 GPU supporting AMD Eyefinity technology6 PRODUCT OVERVIEW can expand a high-resolution picture across multiple displays. In addition, one display of 4096x2160 @60Hz over one HDMI™ or DP1.2 The AMD Radeon™ E8860 Embedded discrete GPU – the first interface can be supported, providing a superior viewing experience. embedded GPU developed on the groundbreaking Graphics Core This flexible, one-to-many system-to-display configuration Next (GCN) architecture – pushes AMD Radeon graphics and capability enables ultra-immersive visual experiences via a single parallel processing performance to unprecedented new heights small form factor system. while increasing power efficiency. OPTIMIZED FOR GRAPHICS-INTENSIVE APPLICATIONS Providing 2x higher 3D graphics performance1 and 33% higher The AMD Radeon E8860 GPU was designed to increase multimedia single-precision floating point performance than the AMD Radeon processing performance and power efficiency for a range of 2 E6760 GPU , the AMD Radeon E8860 GPU delivers industry- leading embedded applications, including: 3D video graphics performance, enabling stunning, multi- display visual experiences for a range of embedded applications spanning Digital gaming. Supporting rich 3D and 4K video graphics and digital gaming, digital signage, medical imaging, and avionics.
    [Show full text]
  • PC Powerplay Hardware & Tech Special
    HARDWARE & TECH SPECIAL 2016 THE FUTURFUTURE OF VR COMPUTEX 2016 THE LATEST AND GREATEST IN THE NEW PC HARDWARE STRAIGHT FROM THE SHOW FLOOR 4K GAMING WHAT YOU REALISTICALLY NEED FOR 60 FPS 4K GPU GAMING WAR EVERYONE WINS HOW AMD AND NVIDIA ARE STREAMING PCS DOMINATING OPPOSITE HOW TO BUILD A STREAMING PC FOR ONLY $500! ENDS OF THE MARKET NUC ROUNDUP THEY MAY BE SMALL, BUT THE NEW GENERATION NUCS ARE SURPRISINGLY POWERFUL PRE-BUI T NEW G US! 4K GAMING PC NVIDIA'S NEW RANGE OF 4K GAMING IS AR EALITY, BUT CARDS BENCHMARKEDH D POWER COMES AT A COST AND REVIEWEDE HARDWARE & TECH SPECIAL 2016 Computex 2016 20 Thelatestandgreatestdirectfromtheshowfloor 21 Aorus | 22 Asrock | 24 Corsair | 26 Coolermaster | 28 MSI | 30 In W 34 Asus | 36 Gigabyte | 38 Roccat/Fractal/FSP | 39 OCZ/Crucial | 40 Nvidia 1080 Launch 42 Nvidia dominates the enthusiast market 8 PC PowerPlay TECH SPECIAL 2016 12 52 56 64 Hotware Case Modding Streaming PC 4K Gaming The most drool-worthy Stuart Tonks talks case How to build a streaming PC What you really need for 60+ products on show modding in Australia for around $500 FPS 4K gaming 74 52 56 50 80 VR Peripherals Case Mod gallery Interviews Industry Update VR is now a reality but The greatest and most We talk to MSI’s cooling guru VR and 4K are where it’s at control is still an issue extreme case mods and Supremacy Gaming HARDWARE & TECH BUYER’S GUIDES 66 SSDs 82 Pre-Built PCs 54 Graphics Cards PC PowerPlay 9 Ride the Whirlwind EDITORIAL Holy mother of god, what a month it’s been in the lead-up EDITOR Daniel Wilks to this year’s PCPP Tech Special.
    [Show full text]