Computer

The External GPUs (eGPU) Review 2019 for Razer Core X with Vega 64

This is a review about eGPU with Macbook Pro retina (2012, 2013, 2014, 2015) & Macbook pro (2016, 2018) & razer core X (USB 3.1). I found this topic is a bit challenging than I thought before. But it is quite useful for 3D graphics, video rendering & gaming. This topic tests about Razer Core X (it is a box with voltage transformers only) with Radeon™ RX Vega 64 – AMD.

This

Here is the agenda:

  • Why people need this
  • Cons
  • Who should use it
  • Budget
  • System Requirements
  • How we test the performance
  • Results & Performance
  • How to install (example setup)
  • Challenge & How to Fix

Why?

1.Saving your time and money & keep upgraded devices. eGPU is a way to upgrade your laptop.

As you noticed that most of the corporate technology giants like Apple, Microsoft have their planned obsolescence / built-in obsolescence strategy. Your devices might no longer functional after a certain period of time. Not because technology is growing fast or your devices are broken/ IT giants aren’t able to support your old devices. It’s because they want their consumers to purchase again (That’s why adapter becomes fast growing products). Sometimes you already spend lots of time & money on your equipments (Or you just keep buying stuff all the time). Getting a new eGPU only spends US$730 (Razer Core X + vega 64). But if I buy a new laptop Macbook pro with 32GB, i9 it will cost like ~US$3.8k. eGPU with my 6 year-old mbpr is still more powerful than the new MBP(*score on geekbench). And, I can still keep upgrade the components one by one.

/

2.Boost up laptop performance (Faster Rendering ~+2X), Geekbench4 ~+3X)

You can use eGPU on Video Rendering (davinci resolve, Adobe Premiere Final Cut Pro, ), 3D graphics VFX rendering (Cinema 4D, Maya, Zbrush, After Effect, Blender etc), Graphic Design (PS, AI, ) in Adobe, Sketchapp for Website&app design, Game Design (Unity, Unreal Engine). It saves you half of the time!! Even for gaming (non-professional, I think).

Running Vega 64 is faster 2 times than internal graphics card (mbpr2012:Intel HD Graphics 4000 & Nvidia GeForce GT 650), 3 times higher scores than MacBook pro 2018.

/

Cons

1.You need a cable for it and you cannot move around like before (believe me, it’s not cheap if you need an adapter. If you are using Thunderbolt 2 (macbook retina, #mbpr) or you a Hub-Guy (you need like 6-in-1 hub on your desk, you will get troubles), you need to go through the Challenge & how to fix before you buy a USB type-c eGPU like razer core x when you still using macbook pro before 2016 version. The customer services cannot help you.

/

2.eGPU is not perfect. If you already have a fine desktop with highend graphic card (like us$1k), you might not need it. eGPU market 2019 works it’s because of laptop components limitations & fast growing graphics card industry. Once those technology giants get rid of GPUs manufacturers Nvidia/AMD or laptop’s GPUs is powerful enough, eGPU might not exist. It just like sony walkman replaced by iPod & iPhone. However, Nvidia/AMD are still powerful in this decade. (Nvidia R&D already in artificial intelligence for graphics, it might be a game changer in visual design as well, not just about FLOPS (floating point operations per second ))

3. There are several companies providing cloud gaming services like Nvidia’s GeForce Now, PlayStation Now, LiquidSky, Vortex, and Google Stadia is coming! You might not need eGPU in future. Google just announced March 29, 2019 that google has partnered with AMD to create a specially designed GPUs for the platform that’s more powerful than the PS4 Pro & Xbox One X combined. Gamers can play games through google chrome (even on Youtube) without processing graphics on user’s computer like the old days that means eGPU gaming becomes less important in the future. You don’t need to worry about if the game supports Windows or MacOS because it is cross-platform services. Google will process the graphics on their cloud services for gamers, and google’s processors are more powerful than PS4 & Xbox combined. (It seems Sony is hopeless frighting with Google Cloud service even Microsoft Azure.)

4. eGPU can be experience too. Blackmagic eGPU cost US$699 (US$827 in Hong Kong) , Blackmagic eGPU pro US $1,199 (US1,337 in Hong Kong). My equipment is US$730 (Razer Core X + vega 64). My setup is bit more expensive than lower version of Blackmagic eGPU in U.S. but Vega 64 works faster than the Blackmagic eGPU Pro (with 60% budget only) . And it’s difficult to upgrade the Blackmagic eGPU. Moreover, eGPU is super power consuming. Some graphics cards are loud and noisy.

5. Supporting issue. Most of the graphics works with limited manufacturer. Let say Adobe work with Nvidia and using it’s CUDA graphic standard. There are high changes that it does not support with AMD’s product. And some protocols are overwhelming. Apple Inc developed a graphics standard (technically it’s an API) called Metal since 2014, but there is a standard called OpenGL(Open Graphics Library) as well long supported by the others. If you do not turn on the settings, you might still using your CPU instead of GPU. Like Adobe Encoder if you haven’t turn on your GPU in settings, (e.g. old version Final Cut Pro does not support eGPU exporting video) you pay eGPU & electricity for nothing. (User should install benchmark just in case.) You might realize that GPUs still look like a un developed product in the market in 2010s.

/

Who should use it

  • Mid-level Gamer
  • Graphic Design, 3D Designer, VFX specialist, Interior designer
  • Game designer, Game developer (Unreal engine)
  • Architect (SketchUp)

/

System Requirements

  • Windows® 10 64bit RS1 or above with Thunderbolt™ 3 port
  • Macs running macOS High Sierra 10.13.4 or later.
  • Thunderbolt 2, Thunderbolt 3, Usb 3.1, 3.0 (USB Type C might be Thunderbolt 3, USB3.0 or USB3.1, there are huge differences on speed, the protocol of data transfer)
  • GPUs voltage transformers (watts power)
  • Software supports (e.g. for MacOs user, Razer Core X said it’s only supports serval AMD’s cards officially. You have to unlock by typing code in the terminal if you want to use Nvidia’s product)

/

How we test the performance

/

How to install

Go to: https://egpu.io

search your setup, here is my set up reference:

System

• MacBook Pro (Retina, 15-inch, Mid 2015) (MacBookPro11,5)
• CPU: Intel i7-4870HQ (8) @ 2.50GHz, quad-core
• GPU: AMD Radeon R9 M370X, Intel Iris Pro
• Mac OS 10.14 Mojave

eGPU + Enclosure

• Razer Core X 
• Apple Thunderbolt 3 to Thunderbolt 2 Adapter
• Apple Thunderbolt 2 cable (2m)
• SAPPHIRE Radeon RX Vega 64 DirectX 12 21275-03-20G 8GB 2048-Bit HBM2 PCI Express 3.0 CrossFireX Support Video Card  

Mac Installation (Mac OS 10.14 Mojave)

1) Install Purge-wrangler, go through the enable AMD flow. This will enable Thunderbolt 2 bus to be used for eGPU. It will ask if you want to enable legacy AMD cards. Say no. It will then ask if you want to enable Ti82 enclosures. Say no. 
2) Stick the thunderbolt cable in. It just works.

Windows Installation (Windows 10)

1) Reboot system while the Thunderbolt 2 is plugged in.
2) It just works! No error 12, no finagling needed, nothing. I just booted into my boot camp installation and it worked.

I already had the AMD drivers for my own internal dGPU, so maybe that helped? The AMD taskbar utility just started to show the Vega 64 as the GPU, and that was it.

Results & Performance

To be continued…

#eGPU #externalgraphicscard #GPUs #AMD #NVIDIA #VEGA64 #1080

How to create Laravel project by using Valet

Step 1: Install environment
installation for Php 7.1
brew install homebrew/php/php71
or
installation for OS X install Php 7.1
curl -s https://php-osx.liip.ch/install.sh | bash -s 7.1

reference:
https://php-osx.liip.ch/

installation for Homebrew
/usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
brew install homebrew/php/php71
Update Homebrew
brew update

Composer

php -r "copy('https://getcomposer.org/installer', 'composer-setup.php');"
php -r "if (hash_file('SHA384', 'composer-setup.php') === '669656bab3166a7aff8a7506b8cb2d1c292f042046c5a994c43155c0be6190fa0355160742ab2e1c88d40d5be660b410') { echo 'Installer verified'; } else { echo 'Installer corrupt'; unlink('composer-setup.php'); } echo PHP_EOL;"
php composer-setup.php
php -r "unlink('composer-setup.php');"

install Valet

composer global require laravel/valet

install laravel installer
composer global require "laravel/installer"

Step 2: Add Path to Terminal
echo $PATH
for running valet & composer installation
export PATH=/usr/local/php5/bin:$PATH
export PATH=$PATH:$HOME/.composer/vendor/bin

Step 3: Build up project
valet install
Go to folder you want, we suggest use Code (~/ is user home directory)
create a folder code under “User/”
input cd ~/code in terminal
cd ~/code
tell the plugin that you want to use code as the working path so that app.dev runs the program.
valet park
create a laravel project, naming it as “app”
laravel new app
running “app”
ping app.dev
secure the local domains by using terminal
valet secure

How to Allow Apps from Anywhere in macOS Sierra 10.12 by using Terminal

Tips for install more app for mac os. If you wanna install an app for your macbook pro or mac mini desktop even macbook air from unknown developer, you have to get access for your macbook first. Apple makes it difficult to install unknown sources for security issue. The following Instructions shows you how to get access in your “System Preferences” > “Security & Private”.

Open app download from Anywhere

input this in Terminal
1: Command+space input “Terminal” or “terminal” in Spotlight search
2: input the following code in terminal
sudo spctl --master-disable

Reset to Default Security in macOS Sierra

sudo spctl --master-enable

Show/Hide Hidden Files the Long Way
The long way to show hidden Mac OS X files is as follows:
Show:
defaults write com.apple.finder AppleShowAllFiles YES
Hide:
defaults write com.apple.finder AppleShowAllFiles NO

Bootstrap 4 Beta 正式出爐

由Twitter推特的程式員開發前端工具Bootstrap4終於發佈Beta版本。Bootstrap是一套為網頁開發而訂製的前端框架,最初由 Mark Otto 和 Jacob Thornton 製作命名為Twitter Blueprint。由於前端開發人編寫的風格因人而異,而且每次加載重覆的文檔費神失事。這班前端工程師製作了一套共享文檔,各自利用其開發前端網頁,亦便利了後端工程師的工作。簡單來說,這一套框架讓工作人員共享一套文檔,包括CSS (Cascading Style Sheets), JS(JavaScript),Ajax等文件,用來開發、傳閱、展示和匯報。程式員要需要下載一次和按時更新文件,便可離線工作,大大減少重覆加載的所耗費的資源和時間。而且線上傳閱時,透過CDN(Content delivery network)方法,數行代碼便可執行網頁,無疑方便整個項目由開發到匯報的時間。

自2015年4月宣佈發開,屈指一算已經2個年頭。

以下是這次Beta版本的撮要

  • CSS方面,正式融合Sass,棄守Less
  • 新版Grid system更具彈性。
  • 放棄IE8 & IE9 的支援 (終於放棄治療!!)
  • JS方面,整個JS插件重寫!
  • 重寫簡介文件,網站Document使用了algolia的設計 (多無關痛癢的改動) 。

Progressive Web App

Progressive Web App (PWA) is a term used to denote web apps that uses the latest web technologies. Progressive web apps are technically regular web pages (or websites) but can appear to the user like traditional applications or (native) mobile applications. This new application type attempts to combine features offered by most modern browsers with the benefits of mobile experience.

什麼是演算法(Algorithm)

讀CS朋友告訴我,計算機科學/資訊工程有三本必修的秘笈,一般人是不會碰到的,碰到都不會翻開,大概是易筋經、葵花寶典和九陰真經吧。
那是CLRS,CSAPP和SICP。CLRS即1990年出版的演算法導論Introduction to Algorithms;CSAPP即深入理解計算機系統,在2003年才出版Computer Systems: A Programmer’s Perspective;早於1979年出版的SICP(計算機程序的構造和解釋), Structure and Interpretation of Computer Programs在網上有詳細的免費版本,但CS同鞋都不會選擇,因為那是近四十年前的著作啦(https://mitpress.mit.edu/sicp/)

CLRS代表了四位演算法的泰斗,包括Thomas H. Cormen, Charles E. Leiserson, Ronald Rivest, Clifford Stein. CLRS由四人last name合併得來,1990年初出版叫CLR。

Thomas H. Cormen寫的Introduction to Algorithms,對於門外漢,是相當深的著作,需要一定的數學基礎。
建議看之前,看一下BBC製作的節目,”The secret rules of modern living:Algorithms”。那麼Algorithm究竟是什麼呢?
據Thomas Cormen所說,演算法是一個定義清楚明確的計算過程,它可以輸入一些數值,從而輸出一些結果,而這個計算的步驟,就是演算法。Thomas Cormen提出了三個重要概念,非常重要,第一個是這個計算過程的法則,必定有有限的,它不能無限運作。第2個是精確,“specification must provide a precise description”;第3個是「有效」Efficiency的。這個有機會再用例子說明。

另外要介紹一下麻省的MIT 6.006(Fall 2011)課程,那是2011年的課程,介紹了什麼是sorting,頂點之類的數學概念,非常有趣,我第一次看的時候,花了近一個小時加速去看,我在想,的究竟看了些什麼!!?然後花了非常長的時間才明白一點點,學習斜坡是非常陡峭的。

軟體概念MVC模式與SPA單頁應用淺談

現在的網頁技術發展愈來愈抽象化,大概是還原基本步。由最初能夠做到便足夠,到現在我們還會思考效能和擴展性。
MVC模式一在早在1979年代已提出的概念,用於軟體設計及運作。它就如一種三權分立的模式,大家互不統屬,從而獲得最佳效能。MVC代表了「模型」(Model)、「視圖」(View)和「控制器」(Controller)。這個理論是這樣運作的,網民一旦開啓網頁(或軟體),這時網民其實是發出一個請求,網站伺服器(或軟體)裏的「控制器」(Controller)便會傳達資訊給「模型」(Model),模型會更新資訊給「視圖」(View)。在這零點零零幾秒之間,View便會更新畫面,讓用家瀏覽。整個運算過程,由Controller在Model之間完成,前者被用家選擇和控制,後者則作為協調和規範的角色。

由於最初的網頁設計,例如一頁index.php或JSP當中,裏面包含HTML和一些運算式,這樣不太有利數式分工。HTML是一些基本表現式,但PHP等OOP程式涉及運算邏輯,兩者分離可以減少各自負擔,對於擴展和維修有很大幫助。此外Model還能分擔重覆的部分,讓View重覆使用,能夠減少重覆語句。

SPA是指「單頁應用程式」(Single-page Application)。意思是一個程序之中,以一個頁面處理大部分工作,減少網頁重新載入的部分,換句話說能減少網頁傳輸負擔從而加快瀏覽速度。SPA能夠以javascript完成,現在大多使用API和Ajax和JQuery等技術達到SPA效果。angularjs是由Google維護的一種框架技術,內裏包含了Ajax的技術,可惜速度和第一版的設計被人垢病,現在angularjs已出版了第2版,幾乎重寫了整套框架語法,其陡削的學習曲線亦令不少程序猿卻步。

數據管理101之Data Analysis基礎概念(上)——Entity Relationship Diagram(ERD)與Relational Database Management System(RDBMS)

數據管理的應用需要數據分析員處理和釐清邏輯關係。這裏要介紹一種概念技術和其作者——美籍華裔的陳品山(Peter P.S Chen)。

所謂數據只是一種的紀錄和資料,在未被處理的情況下,它只是一堆文字或檔案。在商業世界裏,數據可以說是一本天書,它能為目標導向的商業組織作出指導,當然前提是數據能被有效收集、統計和分析(這裏是三個完全不同而又缺一不可的步驟)。但在電腦應用範疇裏,我們是以如何分析呢?我們明白電腦應用其實依賴以數學式的邏輯作為語言基礎。那麼數據管理或數據分析,實際上的操作是根據什麼呢?現在的商業機構,會聘用數據分析員(Data analyst)處理數據,那個系統叫數據庫或資料庫,或者稱為「資料庫管理系統(Database management system, DBMS)」。除了電腦科學(Computer Science)外,這些分析員所學習的是「數據模型」(Data Model)理論,例如層次模型(Hierarchical model)、網狀(Network model)、關聯模型(Relational model)、鑽型(Dimensional model)和ER模型,ER模型即ERD。

Entity-Relationship Diagram(ERD)

Entity-Relationship Diagram(ERD)是一種著名的數據模式理論,現在大多軟體開發都會應用這個數據開發模式概念。運用Entity-Relationship Diagram(ERD),能夠視覺化數據個體(Entity)和它們之間的關係(Relationship)及其屬性(Attributes),用來規劃數據結構。換句話說,這個理論有助將現實世界的抽象資訊轉換為數據,讓分析員能夠量化及統計。ERD的視覺化邏輯亦有利數據分析員檢視、統計和管理。在規劃數據管理上,會將不同類型的數據分類,我們叫作邏輯設計(Logical design);在實踐和製作數據庫/資料庫上,稱為物理數據庫設計(Physical database design)。

ERD,我們可譯為「實體關係模組」。在1976年3月,台灣國立台灣大學電機工程學系畢業的陳品山,在ACM Transactions on Database Systems上發表了《The Entity-Relationship Model–Toward a Unified View of Data》一文。由於大眾廣泛使用實體聯繫模型,而這篇文章已成為計算機科學38篇被廣泛引用的論文中之一。他已被譽為全球最具軟體開發技術的16位科學家之一。早在1968年,IBM的推出一種產品IMS(Information Management System, IMS),它使用的是層次模型,一種樹型結構的數據模式。1970年,Edgar Frank Codd提出關聯式資料庫(Relational model),藉助於集合代數(Set)等數學概念來建立數據之間的關係。Edgar F.Codd在IBM工作期間提出的「科德十二定律」(Codd’s 12 rules)來建立數據模型的理論,它是ERD概念的先驅。後來更有EERD(Enhanced Enitity Relatioship Diagram)是ERD的進階加強版,加入其他進階術語。

ERD基本分為三大基本部分,實體(Entity)、屬性(Attribute)、關係(Relationship)。長方形代表實體(Entity)、橢圓形代表屬性(Attribute)、菱形代表關係(Relationship)。例如部門和僱員都是一個實體;它們的關係是部門包括員工。

但是考慮到它們關係並不平等(僱員不會包含一個部門),因此我們需要另外兩組符號標示兩者關係。 (more…)

Build a Blog with Laravel 101

Laravel is an open-source PHP web framework created by Taylor Otwell and intended for the development of web applications following the model–view–controller (MVC) architectural pattern. The very first version was created in 2011. Back in those days, there were already several frameworks among the industry.

Some of the features of Laravel are a modular packaging system with a dedicated dependency manager, different ways for accessing relational databases, utilities that aid in application deployment and maintenance, and its orientation toward syntactic sugar. (more…)

Back
Copy link
Add to
Email
Facebook
Whatsapp
Gmail
More
Save
Like
Cancel