@fishy_fis

So I'll give you the benefit of doubt - what exactly do you mean by better?

@megol

Quote:

Do you have any idea why MMX was the SIMD format for early x86?

From reading the Apollo site they've somehow managed to add 32 vector registers. That look like 32-bit width. I expected 128 bit. After all Gunnar is reported to say or provide graphs as to how Apollo core is better than PPC in some regards.

But I think the info on MMX is stil lacking. Gunnar says on his forum to look up MMX. But then AMMX differs from MMX. Looks confusng. When ever I look I can never find a page with basic info about this 080 thing. An assembler guide on AMMX doesn't even have basic info on registers last time I looked. Finding what SAGA actually is was the hardest. All this talk about but what it is? The answer was a page with new registers. There is more info now.

Quote:

I don't like the AMMX idea nor implementation for the Apollo core btw.

I don't like the idea of anything Intel inside 68K. You put X86 into 68K and then suddenly you've got a K86!

But, VMX may have not been much better, since it's RISC based. So is ARM. Guess there wasn't much else to base it on from the CISC world. Apart from building on the 68K with the common operations to match the 68K ISA and naming schemes.

@Hypex

Quote:

Hypex wrote: @megol Quote: Do you have any idea why MMX was the SIMD format for early x86? From reading the Apollo site they've somehow managed to add 32 vector registers. That look like 32-bit width. I expected 128 bit. After all Gunnar is reported to say or provide graphs as to how Apollo core is better than PPC in some regards. But I think the info on MMX is stil lacking. Gunnar says on his forum to look up MMX. But then AMMX differs from MMX. Looks confusng. When ever I look I can never find a page with basic info about this 080 thing. An assembler guide on AMMX doesn't even have basic info on registers last time I looked. Finding what SAGA actually is was the hardest. All this talk about but what it is? The answer was a page with new registers. There is more info now.

The 24 vector registers of AMMX are 64 bits wide, as are the 8 data registers that constitute the other 8 vector registers. Gunnar has said that the 16 address registers are also 64-bit though there's no way to access the high longword yet. Having the processor design in-house will be to their advantage if they want to implement 64-bit addressing on the 68090 or 68100.

Having the MPU (memory protection unit) separate from the MMU is central to his plan to add memory protection to the flat memory model of the Amiga but he needed to have OS development in-house to add that capability. Now he has it in his fork of AROS called ApolloOS.

I wish them good luck! It sounds like those who are with Gunnar (and can still put up with his shennanigans) have a plan to follow and a future ahead of them, if they can pull it off.

BTW, the Autodocs for AMMX are at http://www.apollo-core.com/AMMX.doc.txt for anyone interested.

@Samurai_Crow

Quote:

The 24 vector registers of AMMX are 64 bits wide, as are the 8 data registers that constitute the other 8 vector registers. Gunnar has said that the 16 address registers are also 64-bit though there's no way to access the high longword yet. Having the processor design in-house will be to their advantage if they want to implement 64-bit addressing on the 68090 or 68100.

He must have changed the specification. I was reading from this. It doesn't help that he keeps saying look up MMX rather than just giving basic specifications.

http://apollo-core.com/knowledge.php?b=1¬e=1897&z=ljj__N

Also, 68090, 68100? LOL. I thought this was all about the 68080!

Quote:

Having the MPU (memory protection unit) separate from the MMU is central to his plan to add memory protection to the flat memory model of the Amiga but he needed to have OS development in-house to add that capability. Now he has it in his fork of AROS called ApolloOS.

I hope it won't be an Amiga without a Workbench. It wouldn't be the same. Sounds complicated if there is an MPU and MMU. Most MMU support on Amiga is through the MMU libraries. Thought it would have been easier just to add support for the 080 there. If the project is still open.

Quote:

I wish them good luck! It sounds like those who are with Gunnar (and can still put up with his shennanigans) have a plan to follow and a future ahead of them, if they can pull it off.

Well they have been pulling it off for some time now so I would say so.

Quote:

BTW, the Autodocs for AMMX are at http://www.apollo-core.com/AMMX.doc.txt for anyone interested.

I think I've seen that before. What's all the Ps mean in the instructions? V would make sense to me. Bank switching? Oh no! It's going to look like the palette registers in AGA. Ax, Bx; Dx, Ex? Not again no! He really is turning the Motorola 68K into an Intel K86!

@thread

It appears matthey has spawned a second thread about SIMD units. (This probably should have happened a few pages ago.)

@thread

this sounds more dangerous for intel/amd:

https://www.nextplatform.com/2020/07/31/the-dollars-and-sense-of-nvidia-paying-a-fortune-for-arm/

_________________
retired

Quote:

OlafS25 wrote: something to smile https://www.youtube.com/watch?v=Z9IoSPT-suc old macintosh ad where they promote themselves as fighters for freedom seen today it is almost ironic

Fortnite parody video of old 1984 Apple ad after Fortnite was banned from the Apple store.

https://www.youtube.com/watch?v=ybN8NJqgZiY

Last edited by matthey on 14-Aug-2020 at 12:40 AM.

@matthey

Also Google is banning Fortnite for the same reason. Since before, Epic games are also in a conflict with Microsoft's Xcloud.

Quote:

TRIPOS wrote: Also Google is banning Fortnite for the same reason. Since before, Epic games are also in a conflict with Microsoft's Xcloud.

Steve Jobs in a Wall Street Journal interview at the launch of the App Store in 2008 said, “Those free apps cost money to store and to deliver wirelessly. The paid apps cost money, too. They have to pay for some of the free apps. We don’t expect this to be a big profit generator. We expect it to add value to the iPhone. We’ll sell more iPhones because of it.” Apple’s App Store is now a massive $519 billion developer ecosystem (number suggested by Apple under scrutiny). It’s also a key part of Apple’s growing services business, which is the second biggest revenue driver for the company, behind the iPhone.

“We treat every developer the same,” said Tim Cook in his House Judiciary Committee Testimony. “We have open and transparent rules. Those rules apply evenly to everyone.”

Transaction fees
Apple App Store 15%-30%
Google Play Store 15%-30%
Samsung Galaxy Store negotiable-30%
Credit Cards 1.3%-3.4%

I couldn't find the AMIStore transaction fees. The Apple App Store sales were $386 million on January 1, 2020 (single day sales). Apple's App Store probably had more sales revenue in a few seconds than the AMIstore did all year. Is that worse than the Raspberry Pi selling more units in an hour than the Amiga in a year?

Last edited by matthey on 14-Aug-2020 at 09:36 PM.

@AmigaBlitter

Quote:

AmigaBlitter wrote: @thread this sounds more dangerous for intel/amd: https://www.nextplatform.com/2020/07/31/the-dollars-and-sense-of-nvidia-paying-a-fortune-for-arm/

Counter argument from https://www.extremetech.com/computing/313471-arm-co-founder-sale-to-nvidia-would-be-a-disaster
ARM Co-Founder: "Sale to Nvidia Would Be a Disaster"

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

@fishy_fis

Quote:

fishy_fis wrote: @Hammer Quote: Classic Pentium's 64-bit bus was a good forward-thinking for 64-bit MMX SIMD (two INT32) and AMD K6's 64bit 3DNow SIMD (two FP32 or two INT32) At the time maybe. These days Intel are more or less begging people not to use it anymore due to it's awful performance (relative to modern tech). And for some reason the apollo team saw it a good thing to clone in recent times. Only Amiga makes it possible.

FYI, AMD64 (X86-64) has defined SSE2 to replace X87 (FPU) and MMX (INT SIMD).

Since K8, AMD's move to X86-64 has EoL (End of Life) MMX and X87 paths.

Last edited by Hammer on 15-Aug-2020 at 05:48 AM.

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

@matthey
Quote:

matthey wrote: Comparing something similar gives more useful information. In my comparison, the 68060 and Pentium use the same die size, clock speed, voltage, 8kiB ICache/8kiB DCache, are pipelined superscalar with in order execution and use CISC register-memory ISAs. The Pentium has a performance advantage in instruction fetch, data bus width and transistors used. The 68060 has a performance advantage in cache ways and a longer pipeline which is better for clocking up but never leveraged. The Pentium advantages primarily increase CPU cost and energy use while the 68060 advantages primarily increase transistors yet the Pentium uses 28% more transistors than the 68060.

Unlike Pentium Pro (P6, 1995), classic Pentium (P5) didn't include RISCy core design.
68K development was gimped by Motorola while X86 development has continued with Intel, AMD, and VIA. Your argument assumed X86 development was frozen.

Longer pipeline has higher latency. The real innovation is with medium length pipeline with high-clock speed designs.

~21 stage long pipeline has doomed AMD Bulldozer and Intel Northwood (Pentium 4).

AMD Zen is based on Jaguar's medium pipeline length.

Lowest latency with high compute power is important for reaching high frame rates with modern gaming.

Quote:

matthey wrote: The Amiga could use POWER and a PCIe raytracing card. The Amiga could even make this hardware a standard but it will be expensive in low quantity for no more than a few thousand Amiga customers. There wouldn't be a large enough installation base to support development or attract developers which is the problem we have now. In the end, you don't get what you want from the Amiga. The Amiga needs mass production to get back in the game and embedded partners are one way to do that. A single embedded partner may use or sell millions of processors allowing prices to be competitive with ARM but without the royalties or middlemen. Think Raspberry Pi with a reduced footprint. Of course there will be users like you wanting upgraded performance and features but it is necessary to start small and relatively simple. The Amiga 1000 could have been better with a 68020, several times the memory, a hard drive and expansion slots but it wouldn't have been because fewer people would have bought it. It was the cost reduced Amiga 500, which left a lot to be desired performance wise, that turned the Amiga around and then the affordable Amiga 1200 which nearly saved C=. With better management, the Amiga could have made an Amiga like the Raspberry Pi back in the '90s.

POWER is defeated by price and software stack from the re-heated X86-64 competition.

Hardware alone doesn't complete the solution i.e. the Blender 3D performance argument backed by raytracing middleware software stack. AMD's incoming "RDNA 2" has hardware raytracing but it has issues with software support i.e. Radeon Rays for Blender 3D needs updating for RDNA 2.

Intel is using its advanced 10 nm FinFET process tech at ARM's engagement segments i.e. mobile and server.

In terms of performance, ARM's Mali GPU was already defeated by Intel Xe GPU line.
Intel has confirmed Xe-HPG has hardware raytracing while IBM is not even in GPU hardware raytracing engagement.

Quote:

matthey wrote: User mode compatibility is very good although there are a few minor differences. Supervisor mode and the MMU changed significantly.

68020 and 68040 math library difference and the mess it created was stupid.


Quote:

matthey wrote: Some gaming snobs will turn up their nose at even 60 fps while others prefer the higher image quality. Raytracing requires specialized hardware but will likely become more popular and cheaper. It would be nice to bring raytracing to mass produced hardware but it may be too expensive or use too much power to make the 3D standard anytime soon.

Raytracing hardware will be to mass-produced hardware e.g. Sony has ordered +10 million PS5 units (AMD APU with Zen 2 and RDNA 2 based 40 CU based GPU) for it's launch window.

Sony's PS4 to PS5 move wouldn't be a problem.

DirectX12 Ultimate class raytracing hardware will be mass-produced from AMD, NVIDIA and Intel.


Quote:

matthey wrote: Sure. You want a 64 bit data bus if performance is more important than price. There is always a compromise as most customers don't buy the most expensive hardware possible. The 68060 made good choices which did not compromise performance or energy efficiency and made it more appealing for embedded use where it was successful. Sadly, it wasn't used much for desktops or laptops where Motorola/Freescale was pushing inferior low end PPC processors with even more compromises necessary for embedded use.

Bad management doomed both Commodore and Motorola.

Many 68K platform vendors didn't follow Motorola's PowerPC path and many 68K platform vendors didn't survive.

Products like Vampire is good for Amiga's retro-scene. I look at Vampire V2 and V4.

Last edited by Hammer on 15-Aug-2020 at 06:47 AM.

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

Quote:

Hammer wrote: Counter argument from https://www.extremetech.com/computing/313471-arm-co-founder-sale-to-nvidia-would-be-a-disaster ARM Co-Founder: "Sale to Nvidia Would Be a Disaster"

Co-founder Dr. Hauser has nationalistic bias. The sale of ARM to Nvidia would create uncertainty which could be reduced by purchase agreements to keep ARM headquarters in the U.K. and various neutrality clauses. For example, ARM could agree to provide licensing to everyone and not to raise licensing fees for a period of time. The down side to this is that ARM probably should raise fees to increase profitability unless there are significant cost savings from owning ARM. There is a big risk to Nvidia that competitors would drop ARM under their ownership anyway so I believe they want ARM for a reason. The first article suggested the primary reason for wanting ARM was server chips, this one AI and I have suggested CPU+GPU HSA SoCs.

@Hammer
Quote:

Hammer wrote: FYI, AMD64 (X86-64) has defined SSE2 to replace X87 (FPU) and MMX (INT SIMD). Since K8, AMD's move to X86-64 has EoL (End of Life) MMX and X87 paths.

Not in my opinion. MMX and x87 is fully supported even though it makes x86 chips take a bit more power as extended precision multipliers and dividers are larger than the standard double precision. MMX is a subset of SSE so removing it wouldn't save space either.

3DNow! is gone however. Kind of sad - it used a weird encoding as Intel probably didn't want to give them any instruction space. It however lives on in our hearts and in the PREFETCHW instruction, even Intel liked it so much that they added it. There were maybe some pressure from MS involved.

Quote:

Unlike Pentium Pro (P6, 1995), classic Pentium (P5) didn't include RISCy core design. 68K development was gimped by Motorola while X86 development has continued with Intel, AMD, and VIA. Your argument assumed X86 development was frozen. Longer pipeline has higher latency. The real innovation is with medium length pipeline with high-clock speed designs. ~21 stage long pipeline has doomed AMD Bulldozer and Intel Northwood (Pentium 4).

I doubt that. The problems of both those designs are well described. While Bulldozer have been described as AMD's speed demon equivalent to the Pentium 4 it's not really true.

Quote:

AMD Zen is based on Jaguar's medium pipeline length. Lowest latency with high compute power is important for reaching high frame rates with modern gaming.

Do you have a link to an analysis showing a reduction in length for Zen? The days when manufacturers themselves published pipeline diagrams are long gone.

Pipeline lengths have increased for some time now and they aren't the same problem as it were before with improved branch prediction and improved uop caches (shaves most of the front end latency). Data cache latencies have also increased but again with deeper OoO execution it's not as huge a problem.

The move to ARM hasn't slowed Apple so far as it reached a $2 trillion valuation which is a record for a U.S. publicly traded company.



The AAPL stock P/E around 30 looks pretty rich but earnings haven't been hurt by Covid-19 so the stock may go higher. I remember when the stock was a value play with a P/E around 10, more cash per share than a bank and even a respectable dividend for a tech stock. It's hard to believe Apple became one of the richest companies in the world after nearly going bankrupt while Commodore disappeared with the Amiga which was the best computer technology of the '80s.

"Lieutenant Dan got me invested in some kind of fruit company. So then I got a call from him, saying we don't have to worry about money no more. And I said, that's good!" - Forest Gump

Last edited by matthey on 20-Aug-2020 at 04:12 AM.

People, seriously.

Latest iMacs have the system SSD drive soldered to the motherboard.

This means you cannot replace it, neither for updating with a larger drive, nor it simply breaks up.

You must bring the computer to the repair centre for both operations.


Do you really think guys who don't care for a soldered system drive will pay attention to the processor of the whole thing?

@paolone

Quote:

Do you really think guys who don't care for a soldered system drive will pay attention to the processor of the whole thing?

The whole computer branch became a minor part of Apple, they will sold ARM now for their Desktops and Laptops and if they sell 50% less PCs no one cares. There will be a time when the iMac and Apple TV are the same devices, the Laptops will just be iPads with a keyboard and a MacPro will not be needed any longer.
There where already times when professional customers couldn't buy a MacPro and it didn't hurt Apple.

@amigang

FYI, Xbox Series S APU has AMD's 8 cores Zen 2 CPU with 20 CU RDNA 2 iGPU (4 TFLOPS shaders and hardware-accelerated raytracing ~4 TFLOPS).

Hopefully, Xbox Series S APU gets sold in the PC market space.

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

@megol

Quote:

Not in my opinion. MMX and x87 is fully supported even though it makes x86 chips take a bit more power as extended precision multipliers and dividers are larger than the standard double precision. MMX is a subset of SSE so removing it wouldn't save space either. 3DNow! is gone however. Kind of sad - it used a weird encoding as Intel probably didn't want to give them any instruction space. It however lives on in our hearts and in the PREFETCHW instruction, even Intel liked it so much that they added it. There were maybe some pressure from MS involved.

128 bit SSE SIMD supports both integer and floating-point formats which makes MMX instruction set redundant.

X87's FP80 vs SSE2's FP64

Intel Skylake: FP80(fld): 1 per 2 cycles (4 uops) vs. 1 per 0.5 cycles for FP32/FP64.

There's a throughput performance penalty for FP80.

If you need more precision than IEEE FP64, the options include so-called double-double (using a pair of double values to get twice the significand width but the same exponent range), or taking advantage of x87 80-bit hardware.

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

@paolone

Quote:

Latest iMacs have the system SSD drive soldered to the motherboard.

Yes, this removes the overhead, cost and complexity of connectors - you can totally expect other consumer product vendors to do the same. Oh wait, they already did...

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC