I personally would prefer if there were 5-6 competitive x86 CPU manufacturers (including ones not based in the US) and multiple fabs.
x86 is an overly complicated dead-end. It was great for its day, but the base 8086 instruction set is just out of date.
ARM showed that decoupling CPU architecture from an x86 monopoly allowed growth and innovation. Now, just everyone switching from x86 to ARM would be a repeat of the same problem with one body controlling it. I’m seeing RISC-V being the real future. Anyone can make a RISC-V chip. Its an open architecture. No need to pay a licensing fee to AMD/Intel or to ARM Holdings.
The magic of ARM is that there are more than just 2 licensees (technically 3 if you discount the 32-bit only licensee that was Cyrix).
With ARM, we have dozens of ARM implementations with each licensee focusing on their core strength, compute power, low power consumption, media rich processing, high density computing, extremely low cost, embedded systems, etc. This was the sin that AMD/Intel did to us by only allowing them to provide CPU solutions, and only when they felt like it.
I definitely have less exposure to enterprise CPUs. From the little that I have read, their MT performance (or even TCO) isn’t really as great as some of the early previews would lead one to believe…
Mind you, I am not saying ARM isn’t an excellent platform. Just look at Apple’s M and A series, but their approach also comes with its own set of tradeoffs.
I just don’t see ARM being a universal silver bullet (a straight line upgrade from x86) and with SoftBank trying to extract more cash out of ARM, things could get interesting.
From the little that I have read, their MT performance (or even TCO) isn’t really as great as some of the early previews would lead one to believe…
TCO is “total cost of ownership” a very important piece to that in the future is power consumption. Energy prices are rising. This isn’t just the electricity consumed by the CPU but also the cooling needed to exhaust the heat. Many of these highest performance x86 CPUs will cost substantially more to operate as the energy prices continue to rise.
I just don’t see ARM being a universal silver bullet (a straight line upgrade from x86)
Its not there yet, but with Intel fumbling on this one, leaving AMD the only leader in the space, trading one company being dominate over the other doesn’t really serve us well. What I’m pointing out is that its not a “straight line” upgrade, but its curving ever more toward a non-x86 future.
and with SoftBank trying to extract more cash out of ARM, things could get interesting.
I agree which is why I keep making references to RISC-V where I think the future will likely go instead. However, ARM showed (the industry as a whole) that we don’t need to stay with x86 forever as was the notion before. As in, “if we’ve successfully shown we can replace x86 with ARM, what would prevent us from replacing ARM with something else? Not much”.
No actually it never was. It was always a clumsy mess. The only reason IBM picked the X86, was because Intel also made the cut down i8088, that only had an 8 bit data-bus, which made the system easier and cheaper to make. And they also didn’t want it to be too powerful, so it could compete with more expensive IBM systems.
The X86 was excruciatingly slow compared to the competition. Even with the financial strength and developments Intel had available, it was still behind when Intel transitioned to 32 bit with the i80386.
The modest Arm with a tenth the transistors was 4-5 times faster than a full fledged 33 Mhz fully 32 bit 80386DX!
Motorola MC68000 In a Macintosh was about twice as fast at half the clock. And could do MIDI without problems, while the theoretically way more powerful i80386DX had problems executing the MIDI interrupts fast enough, even with extra fast ports installed for it!
Today X86 is considered pretty good on the desktop, because all the competition has disappeared. Alpha, Motorola, Sparc, MIPS, PowerPC.
X86 was never very good compared to any of those. It just enjoyed the benefit of the Wintel monopoly on desktop systems.
We got the worst OS with MS-Dos and later Windows, and we got the worst architecture with X86.
I think you’re looking at it from a pure technological view, but that is only half equation.
No actually it never was. It was always a clumsy mess. The only reason IBM picked the X86, was because Intel also made the cut down i8088, that only had an 8 bit data-bus, which made the system easier and cheaper to make.
The “cheaper to make” was the part that made it pretty good for its day.
Market penetration and ubiquity were key factors in the overall advancement of computing around the world. The explosion of progress occurred when there was mostly one computing architecture, and that writing software for it would mean a huge market with a long life. Most importantly, long enough to make back your initial investment and earn healthy profit.
The modest Arm with a tenth the transistors was 4-5 times faster than a full fledged 33 Mhz fully 32 bit 80386DX!
And with that performance advantage, why is it x86 continued to advance selling more and more units eventually becoming the standard for desktop and server computing? Market penetration.
Back then hardware and software ecosystems were closed. You could learn on a Wang, but that made you useless on VAX. Your SunOS on Sparc knowledge wouldn’t help you very much on Silicon Graphics IRIX on MIPS.
Contrast that with your DOS knowledge on IBM 5150 was almost identical Compaq Deskpro.
Today X86 is considered pretty good on the desktop, because all the competition has disappeared. Alpha, Motorola, Sparc, MIPS, PowerPC. X86 was never very good compared to any of those.
Most of those architectures you mention were workstation, server, or mainframe class CPUs and not desktop. Again, from a purely technical view, sure, they were better, but how good is a CPU that you can never afford to buy?
Even the Motorola (68000 series) and later the PowerPC (for desktops 601 etc) were only in computers that were far more expensive than their equivalent x86 counterparts. It wasn’t for a lack of computing power, but rather those brands wanted exclusive control of their hardware and would crush any attempt to make clones lowering the pricepoint. That did NOT serve the end users or the market, which is largely why I think they failed.
We got the worst OS with MS-Dos and later Windows, and we got the worst architecture with X86.
We got a single CPU architecture and OS compatibility for almost 40 years. If we hadn’t, we would have taken much longer to evolve to where we are today of being able to change out the underlying CPU with lighter weight changes for OS support. Today Linux will run on nearly every CPU architecture including the common x86, ARM, and now even RISC-V. It would have been a much longer path had we had multiple dominant computing architectures all vying for resources.
I remember standing in front of a wall of boxed video games sorting through them, getting excited to see a title, only to see it wasn’t for my platform. Tandy, Apple II, Atari, TI, Commodore, and all the various iterations in between! A game written for Commodore PET couldn’t run on Commodore VIC-20, and the VIC-20 game couldn’t run on the Commodore 64. X86 changed all that. The same game that ran on the 8088 could run on the 286, 386, 486, Pentium, etc. We needed all of that to get where we are today.
The “cheaper to make” was the part that made it pretty good for its day.
No it didn’t, it was so slow it wasn’t faster than 8 bit CP/M systems at the time. The original PC had very little advantage from being 16 bit, and everybody else went directly from 8 to 32 bit. But IBM was bigger than everybody else combined back then, and their support and the arrival of cheaper clones, made it an industry standard disregarding the platform was horrible, but because it was well supported.
And with that performance advantage, why is it x86 continued to advance selling more and more units eventually becoming the standard for desktop and server computing? Market penetration.
I already wrote that Intel was protected by the Wintel monopoly, later when mobile became a much bigger market, that monopoly did NOT help Intel, And Intel spend as much as Arms entire revenue on pushing Atom for an entire decade, and even had the production advantage back then. And despite that Intel was not able to compete against Arm, on platforms like Android that actually had X86 compatibility.
Most of those architectures you mention were workstation, server, or mainframe class
No, Arm was in desktop, but the company did not have the clout to compete.
Motorola was in Macintosh, Atari and Amiga.
PowerPC was in Macintosh and Playstation 3.
That the others were workstation and server does not change that among them all, Intel was inferior in every way.
I don’t understand how you can argue a point that X86 was ever any good, have you ever tried programming assembly on it and on any of the competitors?
Have you ever compared systems from back then on how well they actually worked? For sure the PC was awful. AND MS-DOS was the worst OS in existence at the time.
With Microsoft copying CP/M but removing security features, that has made MS-DOS and Windows the least protected and easiest systems to infects with viruses, causing a decades long nightmare.
Have you ever compared systems from back then on how well they actually worked? For sure the PC was awful. AND MS-DOS was the worst OS in existence at the time.
Yes, I lived through that period and have firsthand experience.
Most of those architectures you mention were workstation, server, or mainframe class
No
I think you missed the part of my post where I called out PPC 601 and Moto 68000 in desktops. PPC was also in workstation and server grade machines including IBM iSeries Midrange systems.
I don’t understand how you can argue a point that X86 was ever any good, have you ever tried programming assembly on it and on any of the competitors?
You’re still arguing technical superiority, when that isn’t the primary factor for folks that bought computers. Consumers didn’t want to throw away their entire computer and software library when going to the next iteration of a company’s product. PC Clones made PC computing affordable. Commodore with its Amiga fought against its only clone Atari ST. Apple quickly squashed any Mac clone makers. These companies got greedy because they wanted to sell hardware at a premium price and control their entire ecosystems, just like they before on prior platforms. They starved their pipeline of younger/poorer customers that would eventually be able to afford the premium products. PC had no such issue and won the computing war of the 80s and 90s.
x86 is an overly complicated dead-end. It was great for its day, but the base 8086 instruction set is just out of date.
ARM showed that decoupling CPU architecture from an x86 monopoly allowed growth and innovation. Now, just everyone switching from x86 to ARM would be a repeat of the same problem with one body controlling it. I’m seeing RISC-V being the real future. Anyone can make a RISC-V chip. Its an open architecture. No need to pay a licensing fee to AMD/Intel or to ARM Holdings.
I am personally not believer in the magical capabilities of ARM. It has its pros it has its cons.
The magic of ARM is that there are more than just 2 licensees (technically 3 if you discount the 32-bit only licensee that was Cyrix).
With ARM, we have dozens of ARM implementations with each licensee focusing on their core strength, compute power, low power consumption, media rich processing, high density computing, extremely low cost, embedded systems, etc. This was the sin that AMD/Intel did to us by only allowing them to provide CPU solutions, and only when they felt like it.
Is that really true though?
I believe there are only two high performance cores with ARM (Apple M series and ARM’s own X series), real-world benchmarks for Oryon are shit.
In terms of smartphone SoCs, you’re stuck with Qualcomm or Apple A series.
There is honestly not that much choice even though there may be many licensees.
And with ARM building out their own SoC, you’re going to have even more challenges with openness moving forward.
I think you’re only looking in the retail space.
Check out the Datacenter grade ARM CPUs like:
Ampere with their 192 core ARM CPUs.
Google Axion with their 72 core CPUs and more than 576B of addressable RAM.
Amazon Graviton 4 with its 96 core CPUs.
I definitely have less exposure to enterprise CPUs. From the little that I have read, their MT performance (or even TCO) isn’t really as great as some of the early previews would lead one to believe…
Mind you, I am not saying ARM isn’t an excellent platform. Just look at Apple’s M and A series, but their approach also comes with its own set of tradeoffs.
I just don’t see ARM being a universal silver bullet (a straight line upgrade from x86) and with SoftBank trying to extract more cash out of ARM, things could get interesting.
TCO is “total cost of ownership” a very important piece to that in the future is power consumption. Energy prices are rising. This isn’t just the electricity consumed by the CPU but also the cooling needed to exhaust the heat. Many of these highest performance x86 CPUs will cost substantially more to operate as the energy prices continue to rise.
Its not there yet, but with Intel fumbling on this one, leaving AMD the only leader in the space, trading one company being dominate over the other doesn’t really serve us well. What I’m pointing out is that its not a “straight line” upgrade, but its curving ever more toward a non-x86 future.
I agree which is why I keep making references to RISC-V where I think the future will likely go instead. However, ARM showed (the industry as a whole) that we don’t need to stay with x86 forever as was the notion before. As in, “if we’ve successfully shown we can replace x86 with ARM, what would prevent us from replacing ARM with something else? Not much”.
I guess we’ll see what happens. I remain unconvinced about ARM replacing x86.
Risc-V is indeed very interesting. Although the performance numbers I’ve seen require a lot more work.
I 100% agree.
No actually it never was. It was always a clumsy mess. The only reason IBM picked the X86, was because Intel also made the cut down i8088, that only had an 8 bit data-bus, which made the system easier and cheaper to make. And they also didn’t want it to be too powerful, so it could compete with more expensive IBM systems.
The X86 was excruciatingly slow compared to the competition. Even with the financial strength and developments Intel had available, it was still behind when Intel transitioned to 32 bit with the i80386.
The modest Arm with a tenth the transistors was 4-5 times faster than a full fledged 33 Mhz fully 32 bit 80386DX!
Motorola MC68000 In a Macintosh was about twice as fast at half the clock. And could do MIDI without problems, while the theoretically way more powerful i80386DX had problems executing the MIDI interrupts fast enough, even with extra fast ports installed for it!
Today X86 is considered pretty good on the desktop, because all the competition has disappeared. Alpha, Motorola, Sparc, MIPS, PowerPC.
X86 was never very good compared to any of those. It just enjoyed the benefit of the Wintel monopoly on desktop systems.
We got the worst OS with MS-Dos and later Windows, and we got the worst architecture with X86.
I think you’re looking at it from a pure technological view, but that is only half equation.
The “cheaper to make” was the part that made it pretty good for its day.
Market penetration and ubiquity were key factors in the overall advancement of computing around the world. The explosion of progress occurred when there was mostly one computing architecture, and that writing software for it would mean a huge market with a long life. Most importantly, long enough to make back your initial investment and earn healthy profit.
And with that performance advantage, why is it x86 continued to advance selling more and more units eventually becoming the standard for desktop and server computing? Market penetration.
Back then hardware and software ecosystems were closed. You could learn on a Wang, but that made you useless on VAX. Your SunOS on Sparc knowledge wouldn’t help you very much on Silicon Graphics IRIX on MIPS.
Contrast that with your DOS knowledge on IBM 5150 was almost identical Compaq Deskpro.
Most of those architectures you mention were workstation, server, or mainframe class CPUs and not desktop. Again, from a purely technical view, sure, they were better, but how good is a CPU that you can never afford to buy?
Even the Motorola (68000 series) and later the PowerPC (for desktops 601 etc) were only in computers that were far more expensive than their equivalent x86 counterparts. It wasn’t for a lack of computing power, but rather those brands wanted exclusive control of their hardware and would crush any attempt to make clones lowering the pricepoint. That did NOT serve the end users or the market, which is largely why I think they failed.
We got a single CPU architecture and OS compatibility for almost 40 years. If we hadn’t, we would have taken much longer to evolve to where we are today of being able to change out the underlying CPU with lighter weight changes for OS support. Today Linux will run on nearly every CPU architecture including the common x86, ARM, and now even RISC-V. It would have been a much longer path had we had multiple dominant computing architectures all vying for resources.
I remember standing in front of a wall of boxed video games sorting through them, getting excited to see a title, only to see it wasn’t for my platform. Tandy, Apple II, Atari, TI, Commodore, and all the various iterations in between! A game written for Commodore PET couldn’t run on Commodore VIC-20, and the VIC-20 game couldn’t run on the Commodore 64. X86 changed all that. The same game that ran on the 8088 could run on the 286, 386, 486, Pentium, etc. We needed all of that to get where we are today.
No it didn’t, it was so slow it wasn’t faster than 8 bit CP/M systems at the time. The original PC had very little advantage from being 16 bit, and everybody else went directly from 8 to 32 bit. But IBM was bigger than everybody else combined back then, and their support and the arrival of cheaper clones, made it an industry standard disregarding the platform was horrible, but because it was well supported.
I already wrote that Intel was protected by the Wintel monopoly, later when mobile became a much bigger market, that monopoly did NOT help Intel, And Intel spend as much as Arms entire revenue on pushing Atom for an entire decade, and even had the production advantage back then. And despite that Intel was not able to compete against Arm, on platforms like Android that actually had X86 compatibility.
No, Arm was in desktop, but the company did not have the clout to compete.
Motorola was in Macintosh, Atari and Amiga.
PowerPC was in Macintosh and Playstation 3.
That the others were workstation and server does not change that among them all, Intel was inferior in every way.
I don’t understand how you can argue a point that X86 was ever any good, have you ever tried programming assembly on it and on any of the competitors?
Have you ever compared systems from back then on how well they actually worked? For sure the PC was awful. AND MS-DOS was the worst OS in existence at the time.
With Microsoft copying CP/M but removing security features, that has made MS-DOS and Windows the least protected and easiest systems to infects with viruses, causing a decades long nightmare.
Yes, I lived through that period and have firsthand experience.
I think you missed the part of my post where I called out PPC 601 and Moto 68000 in desktops. PPC was also in workstation and server grade machines including IBM iSeries Midrange systems.
You’re still arguing technical superiority, when that isn’t the primary factor for folks that bought computers. Consumers didn’t want to throw away their entire computer and software library when going to the next iteration of a company’s product. PC Clones made PC computing affordable. Commodore with its Amiga fought against its only clone Atari ST. Apple quickly squashed any Mac clone makers. These companies got greedy because they wanted to sell hardware at a premium price and control their entire ecosystems, just like they before on prior platforms. They starved their pipeline of younger/poorer customers that would eventually be able to afford the premium products. PC had no such issue and won the computing war of the 80s and 90s.