cloudguy

jdb8167 said:

winston2010 said:

borps said:

So they have a CISC architecture that requires more transistors than a RISC design like the M1 and they are using a 10nm process while Apple is already at 5nm. Sounds like a plan. 

If I remembered correctly, Intel had a RISC-like core with CICS translation in "micro-code" to translate x86 instructions.  That was how Intel was able to increase frequencies so quickly back in the 90's

This is true. They translate x86 instructions into macro-ops that are closer to RISC than the original x86 opcodes. This is fine as far as it goes and served Intel and AMD well for the last 10-15 years. The problem is that Apple has shown that a true RISC architecture with a fixed length instruction length and a massive buffer to reorder instructions can now bring a substantial increase in instructions per cycle that wasn't feasible before. Apple is taking advantage of a huge transistor budget and a consistent ISA to create a very wide instruction decode step and an out of order architecture that is much more efficient than anything else in the industry.

This unique architecture allows Apple's M1 to have the fastest single thread execution while also sipping power. Where Intel and AMD need to clock up to 5 GHz to get good performance, the M1 clocks at just 3.2 GHz and is faster than the top of the line x86_64 (AMD64)  CPUs. This is relatively new and is enabled by using TSMC's fabrication to supply a very large number of transistors relatively inexpensively. Intel is going to have a very hard time replicating Apple's approach because the x86_64 architecture is not amenable to it.

I don't count Intel or AMD out of the race though. They both have very good engineers and Intel's semiconductor process engineers were the world's best until just recently when TSMC eclipsed them. It seems possible that Intel will rally. But they are going to have to do something very clever to get around the limitations of their preferred ISA. Going wide like Apple has in the A14/M1 is probably not a viable solution.

Intel's inability to get past 10nm has nothing to do with the limitations of their ISA. Good grief, AMD is at 7nm already and will join Apple at 5nm by the end of this year. Intel also has released several dual and quad core 10nm chips incidentally. Intel also 7nm dual and quad core chip designs ready to go, although I am not certain about hexacore and decacore designs. Intel's problems are solely about manufacturing. Which is why it is hilarious to see Apple fans take credit for "being ahead of Intel": Apple doesn't manufacture their own chips either. So it isn't Apple that is ahead of Intel but TSMC. Were TSMC's foundries to start having the same issues that Intel is, Apple would be equally impacted. 

Here is the actual factual reality here: if Intel had a competitive business need to, they could buy tons of 10nm and 7nm capacity from TSMC, Samsung, GlobalFoundries etc. for chips that would go on sale by the end of this year and do the same for 6nm and 4nm chips that would go on sale by 2023. But here is the deal: Intel has no competitive business need. Apple fans believing that M1 Macs pose a threat to Intel are merely folks with an inflated idea of Apple's actual importance, which is boosted by Apple's cheerleaders in the media writing columns on their MacBooks and iPads. The reality is that the 15-25 million Macs that Apple sells each year - the vast majority of which use Intel's less expensive Core i3 and Core i5 CPUs - only account for a tiny fraction of Intel's business.

Intel has no competition. It isn't Apple because Apple's market share has never risen beyond 15% ever. It isn't AMD because TSMC cannot manufacture enough chips to make AMD a viable competitor. And it isn't Qualcomm/Samsung because Microsoft has yet to take Windows on ARM seriously (because they have no reason to). ARM is a bigger threat to Intel's server business - especially since in the server arena you can do data science work on GPUs and AI work on TPUs - than it is to Intel's PC business. And again, were Apple or Qualcomm to somehow become a threat before Intel's foundries can catch up, Intel can quickly pivot by having Samsung manufacture 100 million 5nm CPUs in less than 18 months anyway.

People do realize this right? The problem has never been Intel's inability to design 10nm, 7nm and 5nm chips. It has been their inability to make them. Even if TSMC can't - or won't - make them, they can just get Samsung and GlobalFoundries to. (And either can actually make MORE 7nm and 5nm chips for Intel than TSMC can make for AMD). I know that a lot of folks would just rather believe "Apple is going to do to Intel what they did to Zune, the Walkman, Blackberry and Windows CE" but a much more accurate comparison would be what Apple DIDN'T do to not only Android but to ChromeOS (30 million Chromebooks sold last year against 23 million Macs).

Google has entirely separate app development teams for Android, iOS and PC. Features frequently arrive on one of the 3 platforms before they get to the other two, and some features aren't present on some of the platforms at all. So the idea that Google is going to update iOS at the same time that they update Android is specious to begin with. (Apple fans used to have great fun pointing out how new features at times came to iOS before Android as it was used as evidence that Google was losing money on Android. After the Oracle trial debunked this, the chirping stopped.) Comparing Google to Facebook also makes no sense: they are very different companies who - until recently - did not get along and saw each other as competitors for ad dollars until they joined forces due to current political forces started targeting them for breakup due to needing scapegoats for Hillary Clinton losing in 2016. And Google has far more apps to maintain than Facebook does anyway.

Also, Google is going to be forced to upgrade these apps eventually - or more like very soon - or else they will run afoul of their own Project Zero (Day), as most updates are due to security anyway. I can't imagine Google allowing Gmail, YouTube and Chrome going 60 days without security patches. 

EsquireCats said:

In the world of things that pay for your livelihood, home, etc, people are less likely to throw that away for simply having a bad vibe about a company.
It just sounds like he doesn't want to take the time to redevelop his skills for an ARM based platform because it was never meant to be a full time thing, and just wanted to take the opportunity to use his platform while he still has it. (Read: bitch while people are still listening.)

Sure he's welcome to his opinion, but his complaints don't stack up and many also apply to the Window/Linux platforms which he has no beef with.
As for independents, it's arguably the best time ever for an indy developer on Apple:

• Apple pumping money into indy developers for Arcade content
• Apple dropping the 30% cut to 15% for businesses making less than a million on the store
• Unifying the Apple platforms which allows smaller developers to have a wide customer base
• The opportunity to sell the same title again and again on the separate app stores for AppleTV, Mac & iOS, each without needing significant development/investment
• A more flexible approval process with more opportunity to challenge decisions

All of that has literally happen in the span of 18 months. What has -any other platform- done to make things easier for small developers in that timespan?

Yes, it is the best time ever for an indy developer on Apple but he is not an indy developer on Apple. That is what everyone is ignoring. He is a Linux developer who was able to augment his income by doing Mac ports because where in the past much of the work to do the Mac port could be carried over to Linux, now that isn't the case anymore. Let us go point by point.

1. He is not an Apple Arcade developer. He is a Linux PC developer
2. He does not have an app store account for these games because they are not his games. The actual owner of these games are the ones with app store accounts. The owners of these games are paying him to make Linux and macOS ports. They are not paying him for mobile ports. So any benefits from the 30%/15% whatever don't go to him anyway.
3. Again. Not an iOS, iPadOS or tvOS developer. Not even primarily a macOS developer. So this is irrelevant to him
4. See 2. and 3. He does not own these games. He cannot sell these games. He is merely a contract programmer being paid to make and maintain ports of these games to Linux and macOS. 
5. Again, this is applicable to the people who actually created and own Celeste and these other games. Not to him who creates ports of these games to Linux and macOS.

Please go back and read the part where he stated that he would consider reviving his macOS business if the owner of these games gave him royalty payments. The people who are getting those royalty payments now are the ones who will actually benefit from this.

When Apple announced these changes, a lot of people - including myself - stated that it would make macOS more like their other products to the benefit of developers of those other products but were sacrificing multiplatform developers in the process. Apple Silicon is great if you are already an iOS, iPadOS, tvOS and/or watchOS developer but a huge pain if you are a macOS/Windows/Linux/Android developer. Acknowledging the positive aspects of this is great but denying the negative that comes with it is merely deceiving yourself, especially since there is no need: the positives greatly outweigh the negatives.

As I have mentioned on this many times on previous posts, Apple is going to be a thing of its own, completely different from the rest of the industry. This is not a change. This is actually how Apple was from their founding in the 1970s until 2004. During that time, there were Macintosh people and DOS/Windows people. There were companies who used one or the other exclusively. And Apple's market share around that time was as low as 3%. When Apple switched to Intel, there was this period of convergence, which even extended to mobile with Android. Now Apple is going back to the divergence path. Good for people who are already on the Apple train, but it is going to leave the multiplatform people behind. Those folks are going to have to resort to multiplatform solutions like PWAs and cloud. (As I mentioned before, all these indie games should be on Stadia if they aren't already - Celeste is - and if they are on Steam you can import them via Nvidia GeForce Now).

PWAs and cloud are going to be the primary ways to achieve multiplatform capability going forward. But neither will help this guy's use case as he doesn't own the games and doesn't have the legal right to port them to the cloud. Maybe he can talk to the owners of these games about doing ports to Linux-based Stadia but that is all he can do.

Fidonet127 said:

elijahg said:

For a long time this kind of thing has been worrying me. I have some of the games on that list, and I hope it's not the start of a trend for games and other apps.

There is a disproportionately large number of games and cross platform apps available for macOS considering its market share, which is great news for us Mac users. But Apple doesn't make it easy to be an indie coder on macOS. They're so out of touch with the indie devs, and how common it is for indie devs to write apps as a secondary income to their main jobs. Apple just assumes devs have unlimited resources to follow their whims to the Next Big Thing™ and an expectation that devs will always follow along, they unfortunately seem to take them for granted - but in a lot of cases it was these very devs that stuck with Apple though its dark ages.

To name a few recent anti-developer Apple policies:

• Apple's regular deprecation of significant cross platform technologies (OpenGL)
• Their silence on deprecated technologies and APIs (little more than a warning that the "API is deprecated in <macOS version>")
• Announcement of something as the Next Big Thing (VR, external GPUs) and then silence on the subject, and eventual dropping of support.
• Onerous App Store rules with arbitrary application of those rules.

Also, and it's a big one - Apple's expectation that devs spend a disproportionate amount of time on Apple's proprietary APIs like Metal, for a platform whose marketshare is pretty tiny. With a lot of open source apps, engineering and games especially, OpenGL is key. Apple has always lagged far behind with OpenGL support, but a few macOS versions ago it was deprecated. OpenGL support makes supporting macOS little more than a tickbox. But the threat of OpenGL's removal resulting in a rewrite and subsequent maintenance of two graphics engine branches is simply untenable for most devs, so the result is macOS support is dropped. So ultimately Apple ends up harming Mac users, again.

I'd hardly call Metal a tiny market share as you have iPhones, iPads, AppleTVs, and Macs. You are talking over a billion devices. I'd say Apple is listening to devs, as they are making it easier to run Apps across their devices. With the ability to run iPad/iPhones apps natively on the Mac, plus Catalyst etc, this should enable more developers to develop for the Mac. The deprecation of OpenGL and requirement of Metal has been in effect for years now. Prior to that, Apple as warned about requirement of Metal and deprecation of OpenGL for years. 

Yes, this change will benefit more developers in general. However, this guy is a specific case. He is not an iOS developer. He is a Linux developer who got into developing for macOS on the side because of the similarity between Linux and macOS that existed at the time (Mojave). Now that Linux and macOS have almost entirely diverged, he will have to give up most of his macOS work in order to focus on Linux. In other words, he isn't looking to be a macOS developer or a mobile developer. He is strictly a PC Linux developer.

Apple is willing to sacrifice the interests of a small number of people like this in order to beneft a much larger number of iOS and especially iPadOS developers. So as I mentioned in a previous comment, the perfect person to take this over would be an independent iPadOS developer (though I said macOS developer previously). Which is probably what will happen. This guy isn't the developer/owner of these games. He is someone that is being paid by the actual game/platform owners to port the games to macOS - and Linux - on a per game contract fee basis. Humble and/or the devs are simply going to find someone else to make the iPadOS, tvOS and macOS ports while continuing to use this guy for the Linux ports. Or at least that is what we should hope happens. They have been using this guy to make both because it is cheaper. If they have to pay separate devs to make Linux and macOS, that will drive up the cost. 

Rayz2016 said:

longfang said:

He doesn’t want to develop for Apple Silicon, fine. But why make such a big deal about not doing something. Very quixotic behavior. 

I think he's hinting that Apple should give him a new ASi Mac.

Even if Apple did, it wouldn't solve 90% of his problems, which long preceded the switch to M1. It is just that the switch to M1 was the final straw that made it not worth exhaustion of his very limited resources anymore.