If you've been using macOS for a while, you might remember a time when Apple had both GPU options from ATI (purchased later by AMD) and Nvidia. In fact, the Macintosh was the first platform to sport the GeForce 3 in 2001. Nvidia even made a special chipset that was found in the 2008 MacBooks that helped deliver better GPU performance and skipped the Intel integrated chipsets. </p

Then suddenly, Apple stopped using NVidia chipsets. The last Macs featuring an Nvidia GPU was in 2015.

The video version differs slightly as it includes more personal ancedotes and asides.

The history of Nvidia and Apple

The first Mac to ship with an Nvidia chipset was the Nvidia GeForce 2 MX, with the G4 Digital Audio in 2001, and Apple would also at the same time ship the PowerMacs with an option GeForce 3 GPU.

In 2004, 30-Inch Apple's Cinema Display release was delayed by Nvidia's GeForce 8600 Ultra yields, not producing the cards in a timely enough fashion for Apple's liking. Still, Apple continued to offer plenty of Nvidia options. As important as Apple was during this time frame, it wasn't the goliath it is today.

The year 2008 is when the relationship with NVidia changed during a flurry of events. Apple pulled into a legal battle that was primarily between Nvidia and Intel. To understand this, we have to jump back to 2004.

In 2004, Intel and NVIDIA joined forces for a patent licensing agreement for Intel CPUs with integrated memory controllers, the MCP79 and the MCP89. Then in 2008, Nvidia produced Nehalem-based chipsets that bypassed the Intel Northbridge (Memory controller) and South Bridge (I/0 controller) chipset. Apple was the first PC maker to adopt Nvidia's new chipset. The advantage was that Apple was going to be able to simplify its GPU strategy. It'd allowed Apple to stop using the underwhelming Intel integrated GPUs and unify them to mirror the desktops. At the time, Intel's integrated GPUs were pretty bad and could not support OpenCL, thus limiting the amount of offloading to the GPU that Apple could reliably bank with the OS.

Intel was much more central to Apple as a business partner, and Intel enjoyed Apple in its company roster. Nvidia pulling a fast one on Intel put Apple in the center of its own controversial strategy.

Predictably, Intel then filed suit against Nvidia, throwing Apple's plans into disarray. Neither company was endeared to Apple, as the squabble had many industry people speculating that Apple may look into AMD processors, even though AMD had very few competitive offerings in the laptop space. Nvidia tried to court Apple into its legal saga but ultimately failed, leaving Nvidia feeling spurned. Apple continued to use Nvidia GPUs, but sadly, its lower-end offerings were constrained to Intel's supremely mediocre integrated GPUs. This wasn't the only issue Apple was having with its relationship with Nvidia.

Meanwhile, in 2008 Nvidia was hit with a securities lawsuit around knowingly shipping faulty GPUs and trying to mitigate the problem through firmware, burning $196 million for replacements. HP at the time said it had 24 models of laptops affected, and Dell had 15. Apple had 2, the MacBook Pro using the GeForce 8600M GT.

GPUs were failing at a steady clip (not just for Apple), and Apple had to extend its warranties for consumers in 2009 (ending in 2012) and issued a software update in 2009 trying to mitigate the GPU issues. The problem came down to the soldering that held the a chip it's printed circuit board cracked under thermal stress. This still landed Apple in a class action lawsuit. Nvidia saw Apple as a smaller player and refused to extend support costs beyond an unknown amount of money (it only handed out $10,000,000 to Dell after it threatend to pull from Nvidia), putting another twist in the Apple relationship. This was the dividing moment by most accounts.

By this point, multiple publications reported a frosty air between Nvidia and Apple, although the high-end MacBook Pros would continue to use Nvidia GPUs.

Tried to use an AMD chipset in the MacBook Pros in 2011 and ended up in yet another class action lawsuit over faulty GPUs. Apple would switch back to Nvidia in 2012 MacBook Pros.

2013 marked a substantial shift away from Nvidia. Apple went with long-time Nvidia rival AMD for its partnership to produce custom variations of the Radeon FirePros for the 2013 Mac Pros. The iMac 2014s moved to AMD with the introduction of the 5k iMac.

If there was any hope of Nvidia and Apple reconciling, 2014 was the end of it. Nvidia went litigious against Samsung and Qualcomm over mobile graphics patents, filing a lawsuit over mobile GPUs. They went as far as to try and block shipments of Samsung Galaxy S / Note /Tab lines, with speculation that Nvidia wanted the iOS and Android business. At this time, Apple was still relying on components from Qualcomm and Samsung for its mobile units.

Things seemed quiet. Nvidia had ported CUDA to macOS and created Web Drivers even while Nvidia still was producing GPUs for Apple as their relationship fizzled.

Apple had embraced OpenCL, the popular framework used for GPU accelerated computing tasks. Nvidia had created its own closed alternative, CUDA, and using its marketing power to court various software publishers to use it over OpenCL. NVidia's CUDA did not work on AMD hardware, thus giving Nvidia a competitive advantage if a software maker chose to use CUDA. Adobe embraced CUDA even on macOS and thus earned CUDA a favored position among creative professionals, especially those using the Adobe Suite. Adobe went as far as to build CUDA specific applications for Nvidia GPUs. In the background, Apple was poaching industry talent for it's own GPU ambitions.

Nvidia continued a quiet strategy for macOS by bringing support for its later GPUs on macOS and updating CUDA. This meant classic Mac Pro owners, eGPU users, and Hackintosh users could enjoy the latest Nvidia hardware under macOS, which continued uninterrupted for nearly seven years. Many Mac professionals invested in Nvidia hardware as AMD's offerings generally paled against Nvidia at the higher end, and CUDA offered a lot more performance in Adobe video applications like Premiere Pro and After Effects. Nvidia didn't overtly flaunt its web drivers, and it came as a surprise to many Mac users to learn that they could buy Nvidia GPUs and use them in their Mac Pros. As a personal anecdote, I wrote two popular guides on using a GeForce 700s series and GeForce 1000 series GPU in a Mac Pro.

With the release of macOS 10.14 Mojave, everything changed. Outside of the people on Infinite Loop, no one knew for sure that Apple's grand ambition was to merge macOS and iOS hardware. Most users at the time feared the iOSfication of Apple's software instead of hardware.

For years, Microsoft had a huge leg up in the graphics department by owning its own graphics API in the form of DirectX. OpenGL, Apple's preferred graphics API, had floundered in the late 2000s, whereas DirectX, for all its faults, leaped ahead of OpenGL in graphics capabilities and support.

Rather than wait for the next open-source library, Vulkan, to formalize, Apple developed its own graphics API, Metal, for use with iOS. Microsoft most certainly inspired Metal. Bringing Metal to macOS was all-but given and was ported to macOS in 2019, set to replace both OpenGL and OpenCL and skip Vulkan support.

macOS 10.14 Mojave required metal-compatible GPUs. At some point, during the macOS Mojave beta, Apple pulled Nvidia's ability to sign its code, which ended Nvidia's support for macOS in one spiteful, anti-competitive move. In order for GPUs to be Metal compatible, they needed drivers, and Nvidia wasn't able to release drivers.

Nvidia publicly announced that it had working metal drivers on its forum, but Apple had revoked its developer license leaving the blame squarely at Apple's feet. Nvidia even called out apple on his support page but has now since modified it.

My personal take is that it boiled down to CUDA, Metal, and the M1. CUDA represented a significant problem for Metal adoption. In order to get professional applications on board with Metal, they had to cut out CUDA, and my guess is that NVIDIA was not willing to give up CUDA in its driver. Yet again, this was the impasse between Apple's management and Nvidia.

In order for Apple to launch Apple Silicon very smoothly, they needed everyone to support Apple's current technologies, and CUDA was a roadblock to that success.

Apple also knew the aftermarket install base for NVidia GPUs was quite small and limited to classic Mac Pro users and adventurous people who had eGPUs and the Hackintosh community. The group of people this affected was a group Apple the past decade has seemed vaguely resentful of: users who like modular computing. Axing Nvidia was another blow against modularity and another win for Apple's tight-fisted control of when products are obsolete.

The goalposts have now changed. The question isn't whether Nvidia and Apple will get along. It is now whether Apple will allow external GPUs or dedicated GPUs. At the time of writing this, this MacRumors on its buyer's guide page lists that it thinks that apple will release GPUs that outstrip AMD and Nvidia's current offerings.

Usually, MacRumors is pretty on point. Still, I'm just hyper skeptical the year-over-year gains in the GPU market have been not just consistent but going up also. Nvidia and AMD are two of TSMC's biggest clients. They, too, will have access to the same manufacturing processes as Apple. They've been doing it much longer, and they're very good at it.

I have a very unusual take on this whole thing, and that is that in the future, we're going to see macs that absolutely rock at laptop performances and low wattage.

Also, we'll probably see iMacs in a year or two that can edit 8k natively but also can't ray trace and are pretty crap when it comes to things like TensorFlow.

To quote myself after I received my first Apple Silicon mac in December of 2020: "for the portable class of computing, Apple silicon looks like it'll be unmatched, and expensive brute force versus efficiency will be the story of x86 versus Apple Silicon versus ARM, and I expect there will always be a clear winner. Welcome to the next decade of computing."

Iriun Web Camera on the Mac is a bit of a pain to uninstall. Utilities like AppCleaner will not completely uninstall all the Iriun as it doesn't know about the services that run in the background, which will continue to run the background of your computer regardless if it's uninstalled or not.

Fire up Activity monitor, and search for iriun, and force quit any tasks it has operating (there should be at least two).

Next, you'll need to delete the following (this will be from the library located at the root of your boot volume, not in your /Users/ ):

• /Library/CoreMediaIO/Plug-Ins/DAL/IriunCamera.plugin
• /Library/Audio/Plug-Ins/HAL/IriunMic.driver
• /Library/LaunchDaemons/com.iriun.cmio.DPA.camera.plist
• /Library/CoreMediaIO/Plug-Ins/DAL/IriunCamera.plugin

These will likely require your admin password to delete. That's it. Enjoy your Iriun free life as it'll stop popping up as an input option.

Contents

• Know your Mac Pro Models
• CPU Upgrades
• GPU Upgrades
• OS Upgrades & OpenCore
• Firmware Upgrades
• Storage Upgrades
• RAM/Memory Upgrades
• ThunderBolt 2 to PCIe
• Thunderbolt 3 / USB 3.1c
• eGPUs
• Cooling
• Repairs
• Changelog

Introduction

To mark the first anniversary of my wildly successful blog post (garnering tens of thousands of views), The Definitive Classic Mac Pro (2006-2012) Upgrade Guide, I'm proud to announce a sequel. The Definitive Trash Can Mac Pro 2013 upgrade guide started in jest on social media as the guide no one wanted, seeing as the Mac Pro 2013. The 2013 Mac Pro is a tale of hubris for Apple, as it over-promised and under-delivered and is considerably less upgradeable than its predecessor. Is there a need or demand for such a guide? I don't know, but here we are, and while the origins are jocular, the rest of this guide is serious. While most users (and Apple engineers) probably prefer the moniker "cylinder," the trash can title stuck due to its obvious physical characteristics.

The Mac Pro 2013 has the dubious honor as the longest-produced Macintosh, besting the Macintosh Plus produced from 1986 to 1990 without an upgrade. The 2013 Mac Pro was conceived as the original Mac Pro's successor, eschewing the modularity for a (debatably) stylish and radical redesign. After a few positive reactions by publications for its foreign looks, it quickly became snubbed for its lack of upgradability and stability, and Apple's complete and absolute antipathy (verging on enmity) towards it.

The Mac Pro 2013 has been prone to an abnormal rate of failures due to heat, with a nameless Apple exec quoted as saying, "think we designed ourselves into a bit of a thermal corner if you will". Apple also took steps to extend its repair program, but problems persist. Despite the naysayers, the Mac Pro 2013 isn't without its fans (no pun intended), as at the time of its unveiling, it was a powerful, quirky computer with a diminutive form factor. Despite its limited upgradability, the computer is a modular design, and nearly every significant part can be replaced. Only the 2019 Mac Pro since it has allowed for the range of user serviceability (although the iMac 5k is a close second). It's the bridge to a bygone era where CPUs and, storage, and even GPUs were removable. Perhaps the 2019 Mac Pro a return to PCIe, but more than likely, 2013 will be the template.. Edit: The Mac Pro 2019 marks an expensive return to PCIe.

Know your Mac Pro Models

The Mac Pro line debuted in 2006 and has had eight major iterations by Apple's nomenclature: 1.1, 2.1, 3.1, 4.1, 5.1, 6.1, 7,1, and now 8.x. These are also generally referred to by the year 2006 (1.1, 2,1), less commonly 2007 (2,1), 2008 (3,1), 2009 (4,1), 2010-2012 (5,1), and 2013 (6,1). The other terms for these computers are "Cheesegrater" (2006-2012), "Trash can" (late 2013), and "Cylinder." For this guide's purpose, I will refer to the Mac Pro as a "trash can" as the 2013 (as does much of the internet). Please note This guide only covers the 2013 Mac Pro.

Please note This guide only covers the 2013 Mac Pro. For all other models, I've written a massive guide, The Definitive Classic Mac Pro (2006-2012) Upgrade Guide.

Configurations

Apple has only shipped a grand total of 3 base configurations with a fourth build-to-order option for the 12-core CPU. Apple has only made one minor change in the past six years to the Mac Pro 2013 by removing the original base configuration and lowering the remaining models' prices.

• Apple Mac Pro "Quad Core" 3.7 GHz, 12 GB of RAM, 256 GB SSD, and dual FirePro D300 2 GB of GDDR5 (4 GB total). Discontinued April 4, 2017,*
• Apple Mac Pro "Six Core" 3.5 GHz, 12 GB of RAM (16 GB after April 4), 256 GB SSD, and dual FirePro D500 2 GB of GDDR5 (4 GB total). Discontinued April 4, 2017,*
• Apple Mac Pro "Eight Core" 3.0 GHz, 12 GB of RAM (16 GB after April 4), 2256 GB SSD, and dual FirePro D500 3 GB of GDDR5 (6 GB total).
• Apple Mac Pro "Twelve Core"* 2.7 GHz, 12 GB of RAM (16 GB after April 4), 256 GB SSD, and dual FirePro D500 6 GB of GDDR5 (12 GB total). This is a build-to-order option only.

CPU Upgrades

Apple has never acknowledged the upgradeability of the Mac Pro CPU, but the Mac Pro 2013's CPU is not soldered, thus making it upgradeable. Only four CPU configurations were offered by Apple, E5-1620v2, E5-1650v2, E5-1680v2, and the E5-2697v2, but users soon discovered that the E5 v2 family was compatible. Unlike the previous Mac Pros, the Mac Pro 2013 was only offered in a single CPU configuration.

From personal observation, the E5-2697 can be found cheaper on eBay and local used markets (in the US) vs. the E5-2695 is considerably cheaper on aliexpress. This varies based on your local markets, as the European markets tend to be much more expensive than North America.

Credit to the CPU list goes to Mac Rumors forum member ActionableMango.

Useful Links

• Cylinder Mac Pro 2013 CPU Processor Upgrade
• MacRumors: Mac Pro CPU Compatibility List

GPU Upgrades

Yes, the Mac Pro's GPUs can be swapped out, but only three different GPUs were ever produced for it, the AMD FirePro D300 2 GB, D500 3 GB, or D700 6 GB. Apple has kept tight control on these (any official repairs require the GPUs to be returned to Apple), and thus few-to-none exist on the aftermarket, and the two higher GPUs are prone to failures thanks to a wattage ceiling. For most intents and purposes, it is cheaper to buy a Mac Pro 2013 than to track down two GPUs. Apple discontinued the entry-level Mac Pro 2013 that sported the D300. All-new Mac Pros sold after April 4th, 2017, have either a D500 or D700.

For other GPU options, see the eGPU section.

Useful Links

• iFixit: Mac Pro Late 2013 Graphics Card Replacement

OS Upgrades

The Mac Pro 2013 was supported on macOS 12 Monterey. It does not support Sidecar. Currently, OpenCore is a work in progress, fairly stable to get the Mac Pro 2013s up and running macOS 13 Ventura. For the latest information see, OpenCore Legacy Patcher and MacRumors: macOS 13 Ventura on Unsupported Macs Thread

The Mac Pro 2013s do not support AVX2 CPU instructions, which may create some incompatibilities with certain software.

Recently, I converted every Abstract Desktop background that shipped with 10.6 Snow Leopard, in Every OS X Snow Leopard Abstract Desktop - in Glorious 5k and 8k Resolution after being inspired by the wonderful 512 Pixel's Every Default macOS Wallpaper – in Glorious 6K Resolution post. I hand touched up the photos besides upscaling, sharpening, smoothing/blurring, retexturing, denoising, graining depending on the photos. Some had weaker sources than others due to focus/compression etc. For a long explination of how I made these, scroll to the bottom.

The above is an excerpt on a video I'm working on about OS X Snow Leopard, the most beloved OS X/macOS release, but I have other opinions ;) If you are at all curious, you can check out the Apple related vids I've made on the YouTubes.

All the photos are in their original 16:10, in the 5k (if there's such a resolution) 5120 x 3200 as I wanted users to be able to position the desktops as they see fit. Also adding Light / Dark desktops as I have time.

The iPhone wallpaper versions are 2000 x 2778 which should give wiggle room to position the backgrounds, see Apple'sChange the wallpaper on your iPhone for instructions on how to set your iPhone wallpaper.

Aurora

Download Aurora (iPhone)

Aurora Blue

Download Aurora Blue (iPhone)

Cirques

Download Cirques (iPhone)

Clownfish

Download Clown Fish (iPhone)

Dew Drop

Download (iPhone)

Earth

Download Earth (iPhone)

Earth Horizon

Download Earth Horizon (iPhone)

Evening Reflections

Download Evening Reflections (iPhone)

Flowing Rock

Download Flow Rock (iPhone)

Gentle Rapids

Download Gentle Rapids (iPhone)

Golden Palace

Download Golden Palace (iPhone)

Horizon

Download Horizon (iPhone)

Iceberg

Download Iceberg (iPhone)

Ladybug

Download Ladybug (iPhone)

Leopard Aurora

Download Leopard Aurora (iPhone)

Pond Reeds

Download Pond Reeds (iPhone)

Rock Garden

Download Rock Garden (iPhone)

Rocks

Download Rocks (iPhone)

Snow Leopard

Download Snow Leopard (iPhone)

Snow Leopard Prowl

Download Snow Leopard Prowl (iPhone)

Snowy Hills

Download Snowy Hills (iPhone)

Stones

Download Stones (iPhone)

Tahoe

Download Tahoe (iPhone)

Tranquil Surface

Download Tranquil Surface (iPhone)

Water

Download Water (iPhone)

Zebra

Download Zebra (iPhone)

Bonus

Pond Reeds (Light / Dark) Dynamic Wall Paper

Pond Reeds Dark Wall Paper

Download Pond Reeds Night (iPhone)

Aurora (Light / Dark) Dynamic Wall Paper

The Abstract Desktop Series from Snow Leopard

Check out Every OS X Snow Leopard Abstract Desktop - in Glorious 5k and 8k Resolution for the Abstract backgrounds from OS X 10.6.

Snow Leopard Videos I've made

Was Snow Leopard the greatest macOS? I made a video about that. Wonder what it's like to use Snow Leopard in 2024? I made a video about that too.

Note: I don't think the above video conveys how much I like Snow Leopard.

About these images

I do not own these images and thus am providing these free-of-charge (as I do with everything on my blog). These are images that shipped with OS X 10.6 Snow Leopard and upscaled for modern computers.

One of the very frequent tasks computers do is upscaling raster images, aka zooming in or increasing the size of an image. This happens when you zoom in on an image in an application like Preview or Photoshop, or zoom in on a web browser or pinch and zoom on your iPhone. Upscaling has come a long way in the past few years, with machine-learning-assisted upscaling algorithms. These use tricks like taking into account the hue/luminosity (color and its brightness) of surrounding pixels and filling in what it believes the new pixels should be. Before machine-learning, upscaling meant duplicating pixels (nearest neighbor) or duplicating pixels and creating transitions between the hue/luminosity (bilinear). If you have a sharp line, perhaps a sharp mountain silhouette against a sunset (or in my case, a cat whisker against the ground), the machine-learning algorithm will "notice" the sharp contrast between the two areas as it has been "trained" to do so. It will then infer that it should try and keep the sharpness rather than lose the detail, when it creates new pixels to fill the space from upscaling.

While none of these upscales are nearly as good as a high-resolution photo, ML upscaling provides a much better result in a pinch. The ML upscaling is much better at preserving details in the image, such as Mr. Orange's whiskers, his fur, and his pet tag, while providing nice soft results on the areas out of focus. ML (machine learning) upscaling has become popular with modern GPUs. In Windows, AMD and NVidia GPUs can take a videogame rendered at 1440p and upscale it to 4k as it takes less effort for the GPU to upscale the image than it takes to render an image with a lot more pixels, enabling the GPU to churn out higher frame rates and keeping more visual fidelity than using bilinear upscaling. Of course, machine learning isn't limited to uspcaling (or even images).

Machine learning can also be used to provide other types of "best guesses" like denoising images that have artifacts from compression, like JPEGs which are a "lossy format", meaning to save data, they alter the photos to consume less space. Lossy compression is used extensively for media as you do not need a byte-for-byte accurate representation of the original data. Depending on how much compression and what sort of codec (compression type), it will introduce evidence of the compression. Everyone is familiar with the effects of lossy compression, be it an image of low quality, a video with blocky noise, or an audio file that sounds garbled. Machine learning can undo some of this, but it's highly imperfect as it's trying to guess what the original source was trying to convey.

I first tried Gigapixel but ended up just using Pixelmator Pro. Each background was individually treated, using the ML upscaling but with a combination of multiple layers with sharpening, blurs, denoising, grain, and old fashion brush tools on layer masks. As a UX Developer, I split my time between graphics applications and good ol' fashioned coding as my job is to take static pixel images and turn them into code for interfaces for web and mobile applications, although I do often do "full-stack," which is a nice way of saying "everything". I've gone away from Adobe products years ago, with Sketch / Figma / Pixelmator Pro / Final Cut Pro partly out of personal preference and following the industry.

A great example of the power of just a few simple masks is the Ladybug image. First, I nabbed the image from the Snow Leopard Parallels virtual machine on my Mac Pro. Then I imported it into Pixelmator Pro. I converted it to 16-bit, then I created two layers of the exact same image. One I left alone. In the second layer, I used the ML Denoise to get rid of the JPEG artifacting. I used ML upscaling, which upscaled both layers. I masked the non-denoised layer in the Ladybug and used a tiny bit of grain and sharpening, and painted in the hairs on the grass blade. Then I used the blur tool on the larger out-of-focus areas on the denoised image. I went back and lightly burned the Ladybug's antenna. Once I was happy with the result, I downscaled the image from 8k to 5k and left the same color profile same as the original. My goal was to recreate the original as much to the spirit as the original image rather than, say, create an HDR (high dynamic range) version.

Each image was hand treated, although some images had better sources than others. I'm sure another artist with more time could bring these even better to life as I tried to keep roughly an hour-ish per image (as a UX developer, time budgets are a harsh reality for me). If you create your own versions, please reach out to me, as I'll happily link them.

I hope you enjoy, and if you're an Apple nerd and like mediocre camera work, and awkward on-camera presence, you can find me on YouTube and this blog.

If this blog looks familiar to you, you've probably seen, The Definitive Classic Mac Pro (2006-2012) Upgrade Guide or The Definitive Trashcan Mac Pro 6.1 (Late 2013) Upgrade Guide

There are two main routes to get up and running with MAME under macOS, OpenEMU, and SDLMame. OpenEMU is the vector most users will want to use, but it isn't Apple Silicon native (but will still run fine in Rosetta). I've made a video, but if you're like me, you probably prefer the written version.

OpenEMU and Mame

OpenEMU is a beautiful core-based emulator (akin to RetroArch) that supports many game consoles. It's designed to be the one-stop-shop for Mac emulation, and it does a fantastic job. It's gorgeous and easy to use..

Step 1: Download OpenEmu Experimental

Regular OpenEMU does not support the MAME core, so you need to get the experimental version fromopenemu.org. I'm unclear as to why after years, the MAME version is still experimental. If I had to guess, it's likely because there are a few missing assets (the control menu for MAME doesn't have a skeuomorphic graphic, for instance), and it doesn't filter out BIOS files.

Step 2: Safelist MAME (Big Sur 11.x and above)

OpenEMU is open source but it isn't signed, you'll see an error if you try an execute it.

Right-click OpenEMU, and select open to bring up the prompt that will let you run the application.

Step 3: Install the MAME core

Launch OpenEmu, open preferences, and select cores. Click the install button next to MAME.

Step 3: Install games

Installing games is pretty straightforward. Drag the .zip files into OpenEMU under the Arcade tab. MAME games are file name specific, unlike most other emulators, so do not change the name. Also some games have dependencies.

If the game you are trying to play has a dependency, it'll display an error when you try and launch the game, with the name of the missing file. You'll need to install that file by dragging it onto the window. Do not decompress MAME game .zip files.

Please do not ask me for game files. They exist in a legal grey area. Try a popular search engine or archive site like Archive.org for information about ROMs.

Step 4: Optional: NeoGeo Emulation

Emulating NeoGeo games requires one extra step, you'll need to get the NeoGeo BIOS Rom. Drag the Neo Geo zip into OpenEmu. You'll see the NeoGeo ROM in your game list, but ignore it and double-click the games as you normally would. I had some difficulty finding one that worked in more modern distributions of MAME.

SDLMAME / MESS

SDLMAME is a macOS port of MAME and MESS that features both native Intel and Apple Silicon support and (previously) has been updated much more frequently than OpenEMU's MAME core. For years, OpenEMU was stuck at MAME version .149.1, from 2013. As of writing this, OpenEMU is currently in lock-step with SDLMAME as both feature version 0.235, but OpenEMU is not Apple Silicon native. SDLMame is a port of MAME by @sdlmame_osx. It also features MESS which emulates many retro consoles as well.

The downside is that SDL Mame's UI is pretty minimal, and it requires being launched from the command-line but works great.

Step 1: Download the SDL2 Framework

Go to libsld.org, and download SLD2 for macOS and decompress it. Next, install the SDL2.framework into (in your root) Library/Frameworks. SLDL2 is a library for cross-platform development designed to provide low-level access to hardware such as I/O and graphics cards. Many Steam ports for macOS are based on this library.

Step 2: Download SDLMame/Mess

Go to sdlmame.lngn.net, and download Apple Silicon or Intel version depending on which type of CPU your Mac has. Decompress it into a folder of your choosing.

Step 3: Safelist SDLMame/Mess

This is where things get a little strange, as if you try and execute SDLMame, it'll give you the same warning as OpenEMU. Right like the mame unix executable and click open. It'll open the terminal and crash. This is fine. Close this terminal. The application has now been safelisted.

Step 4: Install ROMs

Drag your ROMs into the ROMs folder within SDLMame (If it doesn't exist, create one). Just like OpenEMU games may require dependencies. NeoGEO games require the Neo Geo BIOS.

Step 5: Launch SDLMame using the terminal

YYou'll need to navigate to the directory that SDLMame is in (or locate it). The easiest way is to use the change directory command then drag the folder onto the terminal. It should look something like:

    $ cd /path/to/sdlmame
    

Now that we're in the directory, we can check to see if we see the mame executable using list.

    $ ls
    

If you see mame, you're ready to go. Launch the mame executable.

    $ ./mame
    

SDLMame's UI might seem a bit rough but it does support the mouse. Be sure to use the configure options with Device Mappings and General Input to configure your game controls.

Happy Emulating!

Every macOS comes with some standout desktops, one of the better sets was Abstract series but they're from a bygone pre-retina/4k/5k/8k world. I really loved 512 Pixel's Every Default macOS Wallpaper – in Glorious 6K Resolution post. If you haven't checked it out, you should. I've used them and loved it. Feeling inspired, I decided to do the same to one of my favorite set of Mac desktop backgrounds, the Abstract series from OS X 10.6 - Snow Leopard.

I upscaled and denoised (originals had some JPEG artifacting) and hand touched up them tediously in 16-bit P3 (dithered down to 8-bit P3 JPEGs). I had to make the tough calls on cropping so rather than live with my choices, I've included the 8k originals in the 16:10 aspect ratio for those with 16:10 displays or looking to crop the desktops as they see fit. There's 8 total desktops and I'll probably do an iPhone update and possibly light/dark updates for a few. For now, enjoy!

Download Abstract 5k 16:9 Ratio - 26.6 MB

Download Abstract 8k 16:10 Ratio - 49.1 MB

Every OS X (macOS) 10.6 Snow Leopard Nature Desktop - in Glorious 6k Resolution

Oh hey, I also converted all the nature background from 10.6 to 5k too, check it out here!

Two nights ago, I decided to put my thoughts I've had for months now about the classic Mac Pro into a video. I've personally moved on and retired my Mac Pro 5,1 after spending the last 13 years using a cheesegrater Mac Pro as my primary computer. The Mac Pro enabled me to achieve a lot of my creative and professional achievements: recording two albums, contract work on a few TV shows, and a career change into web development full-time.

BBuying a new computer would be a fairly non-significant life event for most people, but I've spent the last three years updating and maintaining a guide on the classic Mac Pro. If you'd like, watch the video below or continuing reading as I'll provide a written summary below as I understand many prefer written words over video, as I'm one of those people too. My perspective is coming from a practical standpoint as I'd gauge most of my readers to be in the practical camp as opposed to the hobbyist camp where the utility isn't as important.

Leaving the classic Mac Pro behind...

A few short months ago, I bought a 2019 Mac Pro, replacing my classic Mac Pro, a 2010 dual CPU Mac Pro, outfitted with 96 GB of RAM, two NVMe drives, a Radeon VII, and two Sonnet pro series USB cards (USB-A and USBc), as well as a lot of internal storage. Originally, I really didn't think I'd ever been able to justify owning a 2019 Mac Pro. I could certainly afford one, but it's an irresponsible purchase as my needs for a computer do not require a Mac Pro 2019 as a UX/Web Developer. For nearly the last decade, I've used a slew of 15-inch MacBook Pros: 2013 (Retina), 2015, and still currently 2017. Then, the pandemic happened... and things changed for the classic Mac Pro.

Prices went wacky. During the pandemic, used electronics (and cars) skyrocketed in price. For reference, I bought my dual CPU Mac Pro 2010 in late 2017 for $800. I sold it (granted with better specs) for $900 in 2021. GPU pricing went through the ceiling, tripling in value. As I watched the prices climb, I had a crazy idea: if I sold my Mac Pros (two 5,1s, one dual CPU, and one single CPU), the Radeon VII and Vega 56 separately, quick back of the napkin math put me at roughly $1500 shy. I'd pull the NVMe drives and large SSDs and HDDs. I inventoried hardware I hadn't used in ages and added a semi-functional dual CPU tray for a 5,1, studio preamp, a set of Beyerdynamic headphones, an Oculus Quest 2, to name a few things. All said and done; it'd get me only a hundred dollars shy.... except I was about 15 years from my last eBay sale, I failed to calculate the eBay commission, which is heavy, roughly 18%.

I listed and sold my items, many extremely fast (I probably could have charged more for many items, but I was interested in a quick turnaround). I managed to be about $500-$700 short. With the money in hand, I balked several times ordering 2019 as $6000 is.... well... $6000. Inevitably I pulled the trigger.

#1: The classic Mac Pros cannot boot the latest macOS

The Mac Pros have been in a hackintosh-like state ever since Apple dropped official support for the classic Mac Pros after Mojave. Installing Catalina required either using DOSDUDE1 or, better, OpenCore. OpenCore is an open-source bootloader that does many things. The Mac Pro's implementation of EFI is not UEFI as it predates UEFI by a few years. This means a few inconsistencies. Apple's EFI and UEFI use different basic drivers for the pre-boot menu/screen. For Mac Pro users, non-Apple EFI GPUs will not output video until the full drivers have been loaded, which is late in the boot process. OpenCore fixes this by loading the correct drivers. It also emulates other Mac's UEFI, meaning macOS runs transparently without requiring software hacks to the OS itself. This is preferable to patching like DOSDUDE1 as you apply OS updates as there isn't a patch that can be overwritten. OpenCore also allows for various drivers and low-level tweaks, thus making hot-swappable Thunderbolt 3 possible on the classic Mac Pros as well as enabling full video decoding/encoding support.

While all of these minor miracles have greatly benefitted the cMP, with Big Sur 11.3, the Mac Pro hit an unlikely impasse. Apple, in a point release changed the PCIe bus handling with downstream effects to the classic Mac Pro. The cMPs can and will boot above macOS 11.2.3, but the PCIe bus, under certain circumstances, will cause kernel panics. Eventually, it'll corrupt the boot drive and require restoring.

This singular change influenced more than anything else that the clock was ticking for the classic Mac Pro. Apple typically provides security updates for a year to previous releases, and many applications generally support at least the previous version of macOS. Pixelmator Pro, for example, supports 10.15 Catalina, but not 10.14 Mojave. Xcode and Final Cut Pro X generally only supports one OS behind the current. This means the Mac Pro will be left out on the latest versions of some popular/major apps in two years.

Another huge hit is that Big Sur 10.4 added support for the 6000 series AMD GPUs. While it may be somewhat absurd to think about sticking an expensive brand new GPU into such an old computer, an AMD Radeon 6800 was supposed $579, which is less than the Radeon VII was pre-pandemic pricing as the VII MSRPed at $700 (and usually sold on the aftermarket for more). The 6800 might not be fully realized in such an old computer, but it certainly would perform much better than the VII. Thus, today's best GPUs are still limited at the 5700 XT and the Radeon VII for macOS.

#2 Windows 11 requires TPM 2.0

Despite Microsoft calling Windows 10 "the last version of Windows", MS is releasing Windows 11. There's one monumental change that has many people upset and requires Trusted Platform Module 2.0, a security feature that requires late generation Intel and AMD CPUs. MS lists that 8th generation Intel or AMD Zen 2, although 7th Generation Intel CPUs are "partially supported". Microsoft hasn't made it very clear on such issues. You can install Windows 11 on unsupported hardware with a bit of hacking, but that places Windows 11 in a hacked state, and it's unclear if this is a maintainable route and puts any unsupported computer in a precarious unsupported state. MS could easily require later generation instruction sets like AVX, and this would effectively drop support for the classic Mac Pro. Windows 10 will be supported until 2025, so the timeline isn't nearly as imminent as macOS.

#3 The GPU crisis

The biggest selling point of the classic Mac Pro is that it can use GPUs natively in a PCIe slot, whereas all other Macs sans 2019 required eGPU solutions. While GPUs aren't severely bandwidth-intensive as PCIe 2.0 16x vs. PCIe 3.0 16x barely makes a difference, the bandwidth-restricted world of Thunderbolt 3 incurs a higher performance penalty. The outcome resulted in Mac Pros performing better in pure GPU benchmarks over an iMac Pro outfitted with the same GPU in an eGPU enclosure.

However, due to the pandemic/bitcoin/whatever, GPUs are still absurdly expensive (but dropping in price as I write this). This makes the most attractive feature of the Mac Pro out-of-reach or at least justifiable. The GPU crisis also has affected the used Mac Pro market. Most Mac Pros are sold with non-metal compatible GPUs, which are incredibly out-of-date, cannot run macOS 10.14 and above. For the cost of many GPUs, a user could buy the Mac Mini M1 and see better performance for many workflows. The M1's iGPU will run circles around the incredibly out-of-date GPUs like the pitiful GT 120 or more recent ATI Radeon 5770, which brings me to my final point.

#4 the Mac Pro 2019 and Apple Silicon

If you have the coin, Apple has a vector for people who need discrete GPUs, PCIe slots, or a metric assload of RAM. The 2019 Mac Pro took the wind out of the cMP sails as pros finally could retire their Mac Pros without compromise. The 2019 is better than the Mac Pro 5,1 in every single aspect except price. I don't feel like there's much of an argument here. The 2019 Mac Pro is easily the most elegantly designed and powerful computer Apple has ever designed, besting its previous most over-engineered computers like the PowerMac 9600 or the Mac Pro 4,1 2009. The only sad part is that Apple is ending x86 support at some point in the future, nor did it embrace AMD architecture.

It's also impossible not to talk about the Apple Silicon. While it's a tired argument, the M1 only represents the entry-level for the Apple Silicon universe, and it's already better in many workflows than the classic Mac Pros could ever hope to be, especially in the realm of 2D design, compiling code, and general use. It isn't a sure-fire win as there are plenty of limitations like the limited RAM, the limited I/O, the ability only to drive two displays, and of course, the integrated GPU isn't up to the task of competing with discrete GPUs. Still, the M1 is a better computer for many workflows than the classic Mac Pro. We're about to see what the first professional line of Apple Silicon looks like, and if you've been following the rumors, pretty much everything I listed sounds like it'll be addressed to some degree in the MacBook Pro. There won't be many use cases left when this happens where a classic Mac Pro is a better choice than Apple Silicon.

Closing thoughts: The Apple Silicon future

In my article about my thoughts on my M1 I noted that we're likely going to see a future where we see a schizophrenic performance divide between x86 and Apple Silicon. This even seems more true as I've spent more time with my M1. I fully expect to see iMacs able to edit 8k video flawlessly, fly through machine learning/tensor flow, but incredibly ill-suited for 3D modeling/animation, gaming, advanced compositing. I've always felt a bit constrained with laptops, and thus my classic Mac Pro remained my primary computer despite frequent laptop upgrades. The M1 was the first time I started using a laptop over my desktop for some of my creative pursuits, including writing and video editing. Some of this is raw convenience of a mobile writing and editing station, but it was also practical: The M1 was faster in Final Cut Pro X than my classic Mac pro despite the 2 TB Samsung 970 Evo NVMe drive, the 96 GBs of RAM, and a Radeon VII.

The Apple Silicon future still remains questionable for the desktops where thermal budgets are high, computers are modular, and the competition is tough, but if you're able to do everything professionally you want from a laptop, will you really care?

Random Asides

Am I advocating you sell your Mac Pro? Not really. For me, I'm low on hobbyist ambitions. I don't dream of owning a den of old computers spanning my childhood. I'd rather just have one nice desktop and one really nice laptop. I've always been less interested in the nuts and bolts of a computer than really what can it do for me? Though are fundamental to really getting to what a computer can do for the user, the nuts and bolts are fundamental to really getting to what a computer can do. Thus, I ended up self-educating myself and assisting others in the same journey. If you want one really nice computer and have several old Macs and GPUs lying around, selling sooner than later will benefit you.

The big shift though isn't necessarily the classic Mac Pros aren't workable machines, rather we've hit that point where investing further into them is probably not a wise move. At this point, I'd recommend doing the minimal upgrades and saving for a new computer, again from a practical standpoint.

I purposely avoided gaming in this even though I do play some games, and when I do, I prefer to do it on my computer. If you're hyper interested in gaming, the classic Mac Pro is still the best route without breaking the bank to be able to dual boot to Windows and use a dedicated GPU, and play the absolutely mind-boggling catalog of PC games from numerous services/stores.

Bonus: Check out the Reaction Video from several Mac community members, including Jay (House of the Moth) who I've leaned on heavily on in the Definitive Mac Pro Upgrade Guide.

If you are here you are probably seeing the following error in Parallels in macOS 11 Big Sur and you haven't found anything working:

Network initialization failed
Your virtual machine will continue working normally but will have no network connection.

The fix

1. Go to Parallel's Network preferences and click "Restore to Default". This will update the entire config file which is important and remove errant configurations.
   
2. Quit Parallels, and open up the Activity Monitor and search for any Parallels tasks and quit them as well.
3. Go to /Library/Preferences/Parallels/network.desktop.xml with a text editor like Visual Studio Code, Atom, Sublime, BBedit, Webstorm etc that's permissions aware or use the terminal if you're comfortable, and use sudo nano /Library/Preferences/Parallels/network.desktop.xml. Locate the UseKextless entry and set its value to one. Save (depending on the application you're using it'll warn you that you need to use sudo to edit the filea and bring up a dialog. If using Nano, besure to ctrl + x and when prompted, answer yes to save changes).
   

Now relaunch Parallels. The eror should go away and networking should be restored.

As many users saw probably this week, macOS 12 (do we really want yearly OS updates... still?) Monterey was revealed to have Several macOS Monterey Features Unavailable on Intel-Based Macs. John Gruber of daringfireball cleverly observed these features all made use of the imaging pipelines or neural engine.

The chatter in the Mac groups is nail-in-the-coffin, or debating if the features really required an M1 (Spoiler: they don't, they really don't). However, what is truly revealing is how Apple will approach future iterations of macOS akin to iOS. Generally, this has (mostly) limited to iOS camera features that are one part lens but many, many parts software. Could the iPhone XS had night mode? My guess is most likely, but Apple will not spend the energy back-porting the feature to previous hardware configurations. Then, the features are intrinsically hardware related, like the iPhone 12's signal processor, used for Dolby Vision / Smart HDR 3 / temporal noise reduction.

What seems likely is that on macOS we could see old Macs left out of newer features, but not just the latest wifi protocol or faster storage like has always happened. Still, literal features, be it improvements to real-time color correction or perhaps a hypothetical ML library that quickly identifies and highlights elements in a photo.

PC World published Apple M1 vs. Ryzen 5000: MacBook Pro and Asus ROG Flow X13, compared article. While this is entirely leaning on the GPUs found in x86 laptops, it’s also leaning entirely leaning GPUs found in X86 laptops.

Thus far the biggest concern with my M1 is the GPU, I can edit 4k video and its fast long as I stay on the happy path, using Final Cut Pro X. It's easy to scoff at the PC World article for not including web browsing benchmarks, compile times, and other CPU bound tasks but the fact remains Apple's fight for total performance supremacy will be much tougher on the GPU front.

It's been awhile since I've linked a single article, but I absolutely loved the Popular Mechanics article, "How You've Been Conditioned to Love Conspiracy Theories"

"In addition to xenophobia, O’Leary fell victim to “proportionality bias,” the logical fallacy that believes the cause of an event should feel as important as its impact. Proportionality bias lies behind many of the most popular conspiracy theories. For instance, it feels disproportionate that JFK, the so-called most powerful man in the world, could have been assassinated by one disturbed individual, or that Princess Diana,—a powerful, famous, real-world princess—was killed in a car accident."

I wish I could say I had "proportionality bias" as part of my lexicon prior to this as I certainly intuitively understood that many conspiracies are a desire to assert order in a chaotic world and an artifact of the "Tiger in the brush" scenario". It stipulates that it was evolutionarily advantageous to spot false patterns than none at all. If you believed there was a tiger in the brush, then regardless if there is actually a tiger in the brush, you'd still be rewarded with a continuous existence. Thus, evolution tolerates, if not rewards, a baseline paranoia, and overactive imagination. It's this same process that allows many to find meaning in the noise, be it the infamous Virgin Mary in peanut butter to believing that 9/11 was an inside job.

About the only thing missing is that from my (albeit somewhat limited) observations, conspiracies act as currency among the indoctrinated which is its own chaotic agent and generally any rebutal is met as further evidence of the conspiracy as it's just that much deeper. It many ways, conspiracies have filled the void as Americans become increasingly less religious although clearly no more logical.

After a year of being stuck, tethered to social media, the claws of conspirancies turn family members into born-again believers and we're all the worse for it.

Largely ignored by the media outside of a few tech journals, the 3rd Gen Intel Xeon Scalable Processors were announced. Operating under the pretense that Bloomberg's reporting was accurate after predicting the latest Apple event, we can expect at least one more iteration of x86 Mac Pros along side Apple Silicon Mac Pros. Many Mac Pro enthusiasts lamented the Mac Pro's lack of PCIe 4.0, even if there isn't a massive point to it beyond future-proofing. There also was a bit of a disconnect. Intel has not had a PCIe 4.0 chipset until very recently, and Mac users were expecting the impossible out of Intel. GPUs simply do not benefit from PCIe 4.0 as the bandwidth requirements just do not require the increased bandwidth (AMD has shipped 8x PCIe 4.0 GPUs, which is equivalent to 16x 3.0), and there's barely a speed penalty still running the latest GPUs in PCIe 2.0. That leaves mostly SSDs as the sole component that benefits from PCIe 4.0, although I'd argue two things:

1. While not ideal, ASM2428 chipsets have allowed NVMe SSDs (based on the 4x standard) more bandwidth by address more lanes. Currently, in the PCIe 2.0 realm, users can achieve 3.0 speeds using an 8x or 16x slot (as well as multiple SSDs on motherboards that do not support bifurcation). It's likely we'll see similar chipsets allowing PCIe 3.0 cards that address more PCIe lanes to attain PCIe 4.0 speeds in regards to SSDs.
2. Latency and random read/write and caching mechanisms will have more perceivable benefits for daily use than increased bandwidth at this point, considering 3000 MB/s is only going to be a prohibited speed in extreme transfers.

The interesting specs

So what will the new Mac Pro specs look like?

• 10 nm, "up to 40 cores per processor" - This is using the Sunny Cove instead of the Cypress Cove variant, which is 14 nm.
• "46% more performance" - As of late, Intel has taken to dubious stats, and it is sad to see. Considering they're moving from 28 -> 40 core maximum, roughly 40% more cores, this isn't a surprising or impressive claim.
• " 8 channels of DDR4-3200" - Same as previous Xeons.
• "64 lanes of PCIe Gen4 per socket" - There's a lot of ways that PCIe lanes can be handled, but this is a beast amount of bandwidth, and with bandwidth switching certainly enough for workstations (servers are a different beast). Each CPU can negotiate a theoretical 128 GB/s of throughput via PCIe, although that might not be entirely accurate.

It's pretty easy to extrapolate that we're going to see a PCIe 4.0 Mac Pro with up to 40-Cores, roughly the same memory ballpark max-memory cap that'll still have the blistering prices we've seen before. I don't expect Apple to expend much energy redesigning the case, but we'll see the (likely) the final x86 Mac Pro setting a very high bar for Apple Silicon to follow. Still, mostly, the consensus is users want PCIe + upgradable memory and storage to continue.

Updating the firmware on a Mac Pro isn't difficult, but it is possible to "miss" firmware upgrades. This guide is for anyone looking to get to the latest (and most likely last) firmware released for the Mac Pro 5,1s, without having to install Mojave 10.14.x, or if you already have installed Mojave, or are looking to install Mojave. My first try, my firmware was stuck at 138.0.0.0.0.x even when running Mojave 10.14.6. Updating the firmware adds key funcitonality to the Mac Pro 5,1s, most notably native NVMe m.2 boot support. To learn more about Firmware and the Mac Pro 5,1s, see the Firmware Upgrades section of my Mac Pro Upgrade Guide.

These aren't the only instructions on the web, as MP5,1: What you have to do to upgrade to Mojave (BootROM upgrade instructions thread) for firmware upgrading. However, this the method I've found most reliable for users who are having firmware troubles.

Step 0: Remove unsupported GPUs

The biggest change for macOS Mojave is the deprecation of OpenGL and OpenCL. OpenGL has been a thorn in Apple's side for quite some time, as it's been nearly dead for years. Vulkan, the OpenGL successor, wasn't quite ready for primetime when Apple originally created Metal for iOS and thus decided to port it macOS. Despite the annoyingness of having to meet the requirements, it was a necessary evil. Mojave will not install if you have a non-metal supported GPU.

Note: some users are reporting they had to remove all PCIe cards sans their storage controller (SATA card) and GPU to install the firmware update. I did not. If you encounter issues, try removing additional PCIe cards.

Step 1: Have a 10.13 drive

Unfortunately, this is the biggest pain if you've already updated. You'll need a separate volume to boot into 10.13. Amazon and Newegg each have 120 GB SSDs for under $20 USD if you need a temporary drive to install macOS 10.13 on. (upside is you can buy a USB case and turn into a very fast USB 3.0 drive afterward or return it). You can get old versions of macOS via the Mac using DosDude1's installer if you can't access it. If you have no intention to upgrade to Mojave or already have it installed., don't worry. We won't be installing Mojave.

Step 2: Boot 10.13

The next step is pretty straight forward, boot into your install of macOS 10.13 if you haven't already.

Step 3: Download 10.14.6 Combined installer

Fortunately, firmware flashing does not require updating in a particular order. I went from 138.0.0.0.0 to 144.0.0.0.0 without any problems. There's several avenues for this, including the Mac App store, but when I used the Mac App Store route, I didn't get the combined OS installer (The Mojave installer + all the updates to Mojave). The easiest way to obtain the final combined update for Mojave is to use Dosdude1's installer. Much like before, download the OS DosDude1's installer, even though we have supported hardware but with the patcher for 10.14.

Clarification: You do not need to use the DosDude1 installer, as you can grab the update via the App store or other sources but I found this easier. Apparently this link was posted this on MacRumors and a few posters didn't read the full instructions and suggested that I was advocating using DOSdude1 on the OS. I am not. The Mac Pro 4,1/5.1 does not need DOSDude1, so do not run the patcher on Mojave. The Patcher just happens to be extremely reliable about fetching the correct version, and skips the hassle of the App Store.

1. Go to DOSdude1 Mojave patcher and download it

2. Launch the patcher.

   

   Depending on your security settings, your mac may suggest it's from an unverified developer. Go to the system prefs, Security and Privacy (general), and allow the app to open.

   

   You'll be bugged one last time.

   

3. The patcher should warn you that you are on supported hardware.

   

   This is fine, ignore the message. Within the patcher, select the download Mojave from the Tools menu.

   
4. Once the download has completed, quit DOSdude1's patcher.

Step 4: Launch the installer and click shut down

The installer should bring up a message about firmware and a shutdown message. This will not start the Mojave installer, only the firmware.

Step 5: Boot the Mac

Using the instructions in the previous image, press and hold the button until it blinks. If you do not have an EFI enabled GPU (see more about EFI in my Mac Pro Upgrade guide), you will not see any video output.

I trimmed down the video, as it took about 15 seconds of holding before the button flashed. After the button flashed, the internal speaker emmitted a long lowfi "boop" sound.

Step 5: Verify

Go to About this Mac, and click system report. Under the first screen, look for the "Boot Rom" text. This should list your firmware version. From here, you can continue using 10.13.6, upgrade, or boot to your 10.14 volume.

The 144.0.0.0.0 firmware works with any version of macOS your Mac Pro supports.

Updated: November 13th, 2019 MacRumors feedback

Updated: November 4th, 2019 based on Feedback from Mac Pro Users user group on Facebook.

Benchmarking a MacBook Air M1's SSD.

Awhile back, I made a video about USBc and the classic Mac Pro but lamented yet ago the terrible benchmarking on macOS. The first commenter on FaceBook pointed out that we finally have a good disk benchmark utility AmorphousDiskMark. While it isn't a direct port, it's heavily inspired by the famed and loved Windows utility, CrystalDiskMark.

So why am I always complaining about BlackMagic Disk Speed Test

BlackMagic's Disk Speed Test only tests one thing, continuous throughput. This is useful but only measures one aspect of an SSD, and doesn't necessarily mimic accurately how most disk interactions occur. Random Read and Write tests are as important, if not more so, as many SSDs can deliver fast maximum continuous read and writes but much less so for random small data blocks. CrystalDiskMark tests random reads and writes both as queued requests and single requests. The default depth is pretty high for the test. Usually, an OS wouldn't have that deep of a queue, but the Q1T1 does mimic a singular request. Also, CrystalDiskMark measures IOPS (Input/Output Operations-per-second), which is similar but also a different measure of disk speed.

Better but not perfect

There's plenty of aspects that aren't covered, such as latency, burst performance, power consumed, and mixed random read/writes, but this is a massive step in the right direction for gauging SSD performance on macOS. Oh yeah, and it's free.

Let's retire BlackMagic's Disk Mark and embrance Amorphous Disk.