ZORRAM!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

@matthey

Quote:

matthey wrote: From a marketing perspective, it is obvious that customers will buy the 68030 instead of 68020+68851 when it is released. Motorola should have been up front telling customers that there would be no MMU for the 68020 but that the 68030 would follow quickly (6-18 months) if they needed a MMU. Most customers don't need a MMU and they get a cheaper 68020 sooner while the releasing a quality 68020 early builds customer confidence for customers waiting for the 68030. I suspect that Motorola waited to release the 68030 to try to recoup the development cost of the 68851 and sell off their inventories of the chip.

68030 prototype at 1 MHz
Quote:

I was in the Motorola 68000 design group from 1985 to 1996. The first project I worked on was the 68030, which was essentially the 68020 with an integrated MMU and virtual memory capability. We did have a full hardware breadboard of the 68020. The 68030 breadboard was 18 24"x24" wirewrap breadboards, each with a 23x14 array of 20 pin DIP sockets! The chips were mostly PALs and RAM/ROM, no bitslice. This was before software logic simulation was a real thing. Normally I would not have been allowed to take pictures of this, or even bring a camera into the building, but I gave a talk on the 68030 at Stanford in 1986 and I got permission …

I 'suspect' you are wrong about Motorola holding back the 68030 to sell more 68851's. I 'suspect' they were working hard out trying to produce an answer to the 386DX, and they got it into silicon as soon as they could.

Quote:

matthey wrote: if they delayed the release of the 68030 to recoup 68851 costs then this may have been the fatal mistake that caused the 68k to lose 2-3 years to competitors. I may have fixed what you accuse me of causing.

Look at their history - Motorola were always behind the competition. This wasn't a 'mistake', it was just the best they could do.

Quote:

Products of the late '80s required development in the mid '80s and development is much easier with hardware. Again, see CBM wanting a 68k MMU in 1984 likely for Anix.

The primary attraction of an MMU for personal (AKA 'home') computers in the 80's was to get around the need for larger amounts of RAM, and this is still true today. A modern PC with 4GB would be unsaleable without virtual memory because it would run out of RAM just getting to the desktop! Luckily for us the Amiga was designed to work efficiently without an MMU, so we didn't have suffer the extreme performance hit of virtual memory. When Amiga users didn't have enough memory they did the sensible thing - buy more RAM.

Quote:

the ARM based Raspberry Pi is taking desktop market share from the Intel desktop fortress

Not in any meaningful way.

Quote:

Linux didn't exist in the '80s but Unix did and 68k desktop customers were trying to move up into the lucrative workstation market too. They may have been successful if the 68030 with MMU was released a couple of years earlier like it should have been. CBM Amiga 3000UX (AMIX), Apple A/UX and Atari TT030 would have been much more competitive had they appeared before 1989-1990, primarily using a 68030.

... for a few years - assuming they could get competitive systems out that early. But they didn't, which means that in reality they couldn't. They couldn't because the workstation market wasn't just about running Unix. The idea of a company that made its mark in ultra-cheap 'toy' home computers breaking into the professional workstation market is ludicrous, and Commodore trying to do so with the A3000 was one of their biggest mistakes. If only they had put all their efforts into producing something like the A1200 earlier...

A few years later the workstation market collapsed when PCs became powerful enough to challenge them. Why? Because the answer to any desktop computer need is a PC! The Amiga managed to carve out a niche in video production for a while because the original machine was designed around video standards, which also drove the demand for faster 68k CPUs and FPUs for rendering. But we all knew the PC would muscle into this market too, as soon as it was able.


Quote:

Quad precision is absolutely useful for some science and engineering applications.

That's not us.

Quote:

It is possible to retain full extended precision but most compilers ignore the problem...

Problem? What problem?

Last edited by bhabbott on 19-Aug-2022 at 04:39 AM.

bhabbott Quote:

68030 prototype at 1 MHz Quote: I was in the Motorola 68000 design group from 1985 to 1996. The first project I worked on was the 68030, which was essentially the 68020 with an integrated MMU and virtual memory capability. We did have a full hardware breadboard of the 68020. The 68030 breadboard was 18 24"x24" wirewrap breadboards, each with a 23x14 array of 20 pin DIP sockets! The chips were mostly PALs and RAM/ROM, no bitslice. This was before software logic simulation was a real thing. Normally I would not have been allowed to take pictures of this, or even bring a camera into the building, but I gave a talk on the 68030 at Stanford in 1986 and I got permission … I 'suspect' you are wrong about Motorola holding back the 68030 to sell more 68851's. I 'suspect' they were working hard out trying to produce an answer to the 386DX, and they got it into silicon as soon as they could.

Nice story and pic but a 1986 prototype of the 68030 doesn't help the timeline much. Motorola had a 68020 CPU and 68851 MMU design in 1984 yet couldn't produce a 68030 with MMU until 1987 despite combining both in one chip being much simpler as the huge transistor savings shows. It should have been easier to design, test and produce the 68030 than the 68851 which should have been cancelled.

I suppose CBM got Amiga chipset improvements in silicon as soon as they could too. Jay Miner designing the Ranger chipset was only a minor distraction to their fine ECS advancement in 1990 that only took 5 years.

bhabbott Quote:

Look at their history - Motorola were always behind the competition. This wasn't a 'mistake', it was just the best they could do.

Motorola was ahead of the Intel competition with the 68000 and the 68020. They fell behind starting with the 68030, a minor but important update to the 68020, that easily could have and should have been out 2 years earlier. This would have allowed 68040 and 68060 development to begin sooner where the release of these more complex CPUs was late.

bhabbott Quote:

The primary attraction of an MMU for personal (AKA 'home') computers in the 80's was to get around the need for larger amounts of RAM, and this is still true today. A modern PC with 4GB would be unsaleable without virtual memory because it would run out of RAM just getting to the desktop! Luckily for us the Amiga was designed to work efficiently without an MMU, so we didn't have suffer the extreme performance hit of virtual memory. When Amiga users didn't have enough memory they did the sensible thing - buy more RAM.

Speaking of an "extreme performance hit", virtual memory isn't much good using a floppy for storage as the Amiga, Atari and Mac used in the '80s. One thing PC clones did often have sooner is a hard drive and this has synergies with a MMU for virtual memory.

bhabbott Quote:

Not in any meaningful way.

The desktop market is large so millions of Raspberry Pis used as low end PCs is barely noticeable. I doubt there are any statistics on RPi desktop users but the Raspberry Pi 4 and the all in one RPi 400 wedge computer are clearly aimed at the desktop market and likely surpass the embedded market sales for these devices. A x86-64 desktop isn't needed anymore except for playing demanding games and running Microsoft software. That's why a lot of the desktop market has been lost to mobile too.

bhabbott Quote:

... for a few years - assuming they could get competitive systems out that early. But they didn't, which means that in reality they couldn't. They couldn't because the workstation market wasn't just about running Unix. The idea of a company that made its mark in ultra-cheap 'toy' home computers breaking into the professional workstation market is ludicrous, and Commodore trying to do so with the A3000 was one of their biggest mistakes. If only they had put all their efforts into producing something like the A1200 earlier... A few years later the workstation market collapsed when PCs became powerful enough to challenge them. Why? Because the answer to any desktop computer need is a PC! The Amiga managed to carve out a niche in video production for a while because the original machine was designed around video standards, which also drove the demand for faster 68k CPUs and FPUs for rendering. But we all knew the PC would muscle into this market too, as soon as it was able.

It was logical to use desktop computers for workstations because the economies of scale are improved. That is why PC clones eventually took over and why Sun was looking for a larger scale desktop manufacturer like CBM with the Amiga 3000UX to produce their computers. Likewise, embedded economies of scale are better than desktop and we are seeing embedded market products like the RPi and mobile devices leverage that advantage to reduce prices and take desktop market share. The desktop computer may be doomed like the workstation. The only thing holding up embedded market hardware from fully taking over the desktop is the inferior performance of RISC CPUs compared to CISC CPUs. Also, the CISC x86-64 is not as good for embedded use as the 68k making it difficult to penetrate downward into the embedded market despite attempts like the Intel Atom. The 68k was unique in being able to scale from embedded hardware to desktops to workstations. It was the number one choice for the embedded market and workstations and would have been for the desktop as well had IBM chose the best available desktop CPU for the IBM PC.

bhabbott Quote:

That's not us.

Speak for yourself. I'm ACSE certified and the 'E' stands for Engineer.

bhabbott Quote:

Problem? What problem?

The default Unix ABI for the 68k passes floating point function arguments in double precision on the stack. FPU calculations are in extended precision by default though. The problem is that the extended precision is rounded to double precision sometimes. Functions may or may not be inlined so the precision can vary and the lack of consistency can cause problems. It would be possible to set the FPU to round to double precision all the time but rounding is extra work and slower on some 68k FPUs. The default x86 ABI has the same problem for floating point. The following GCC bug report includes hundreds of bugs which have been classified down to this bug but all work is suspended (the bug boneyard where x86 FPU precision related bugs come to live as zombies forever).

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=323

The discussion is funny. They did fix register spills so they are extended precision instead of double precision though.

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=30255

The problem is the default ABI is incompatible with extended precision. The solution is to pass function floating point args in FPU registers with any args not fitting in registers pushed on the stack in extended precision instead of double precision. The concept of extended precision was good but the benefits were never fully realized for the 68k and x86 FPU with the Unix ABI.

Last edited by matthey on 19-Aug-2022 at 05:49 PM.
Last edited by matthey on 19-Aug-2022 at 05:44 PM.

@bhabbott

Quote:

bhabbott wrote: @matthey Quote: matthey wrote: if they delayed the release of the 68030 to recoup 68851 costs then this may have been the fatal mistake that caused the 68k to lose 2-3 years to competitors. I may have fixed what you accuse me of causing. Look at their history - Motorola were always behind the competition. This wasn't a 'mistake', it was just the best they could do.

As Matt said, it was ahead with the 68000, 68020, and I would say the 68851 as well (which was also very very advanced for the time. But so much complex as well).

However those are the exceptions, I would say, because Motorola was behind (in terms of go to market and/or features and/or performances) with its 68008 vs 8088, 68010 vs 80186/80286, 68030 vs 80386, 68040 vs 80486, and 68060 vs Pentium.
Quote:

Quote: Products of the late '80s required development in the mid '80s and development is much easier with hardware. Again, see CBM wanting a 68k MMU in 1984 likely for Anix. The primary attraction of an MMU for personal (AKA 'home') computers in the 80's was to get around the need for larger amounts of RAM, and this is still true today. A modern PC with 4GB would be unsaleable without virtual memory because it would run out of RAM just getting to the desktop!

For today's standards it's already unusable with such limited memory.
Quote:

Luckily for us the Amiga was designed to work efficiently without an MMU,

In reality it wasn't lucky for us: the Amiga o.s. CANNOT make use at all of a PMMU. By design. Because of its message passing. This was why it's so fast, but also the reason that crippled its future.
Quote:

so we didn't have suffer the extreme performance hit of virtual memory. When Amiga users didn't have enough memory they did the sensible thing - buy more RAM.

LOL Are you serious?!? The solution was to buy much more EXPENSIVE RAM?!?
Quote:

Quote: Quad precision is absolutely useful for some science and engineering applications. That's not us.

Wrong. You should have said that it's not for YOU.


@matthey

Quote:

matthey wrote: I suppose CBM got Amiga chipset improvements in silicon as soon as they could too. Jay Miner designing the Ranger chipset was only a minor distraction to their fine ECS advancement in 1990 that only took 5 years.

I assume that you were sarcastic here, because 5 years to develop something (programmable resolutions) which chips like the 6845 (MDA, CGA) already had 10 years before was a little bit too much.
Quote:

The default Unix ABI for the 68k passes floating point function arguments in double precision on the stack. FPU calculations are in extended precision by default though. The problem is that the extended precision is rounded to double precision sometimes. Functions may or may not be inlined so the precision can vary and the lack of consistency can cause problems. It would be possible to set the FPU to round to double precision all the time but rounding is extra work and slower on some 68k FPUs. The default x86 ABI has the same problem for floating point. The following GCC bug report includes hundreds of bugs which have been classified down to this bug but all work is suspended (the bug boneyard where x86 FPU precision related bugs come to live as zombies forever). https://gcc.gnu.org/bugzilla/show_bug.cgi?id=323 The discussion is funny. They did fix register spills so they are extended precision instead of double precision though. https://gcc.gnu.org/bugzilla/show_bug.cgi?id=30255

Unbelievable: 22 years and still not fixed.

And how ridiculous is it, since we're talking about open source (they are supposed to fix bugs soon, as long as they are discovered).
Quote:

The problem is the default ABI is incompatible with extended precision. The solution is to pass function floating point args in FPU registers with any args not fitting in registers pushed on the stack in extended precision instead of double precision. The concept of extended precision was good but the benefits were never fully realized for the 68k and x86 FPU with the Unix ABI.

As I've said before, this might be ("automatically") solved once quad-precision will be introduced. Then it'll be enough to transparently replace quad-precision with extended-precision on 68K+FPU and x87.

@cdimauro

Quote:

I don't. Care to prove that I've forgotten it, dear liar? Anyway this is ANOTHER topic, which is NOT relevant to what we were discussing in THIS specific context. So, another Red Herring: you don't know where logic lives... In fact, and as I've already said, x86/x64 processors do NOT need such microcode update to work. Which is THE RELEVANT point.

In the real world, a BIOS update for the existing motherboard is required to support any new CPUs.

Ask any AMD X470 motherboard owners with Zen 2 and Zen 3 CPUs.

With Windows 11 and certain game DRM that uses TPM, my X570 motherboard has a UEFI CPU microcode update to fix the stuttering issues with certain games and AMD's built-in fTPM. The alternative fix is to install a 3rd party TPM hardware.

Reference: https://www.digitaltrends.com/computing/ryzen-tpm-stuttering-bug-windows-11-patch

The BIOSes that contain the bug fix have the AGESA version 1207 or 1.2.0.7 microcode. Microcode updates come directly from AMD and are sent to the company’s motherboard partners, and from there it’s up to them to distribute the new microcode in the form of new BIOS updates


Unlike Windows 11, Windows 10 doesn't assume TPM's existence.

I updated my X570 motherboard's UEFI BIOS for any potential Ryzen 7 5800X3D purchase.

The BIOS update management has higher complexity when supporting multiple CPU generations (more than two generations) on the same CPU socket.

I purchased 3rd party TPM for about $20 AUD since I couldn't wait for the fix.

---
Another example with Intel Z77 via MSI MPOWER motherboard

https://www.msi.com/Motherboard/Z77_MPOWER/support
2013-07-22, AMI BIOS 17.10 includes Update IVB CPU Micro Code.
2013-01-31, AMI BIOS 17.7 includes Update CPU Micro Code.
2012-12-19, AMI BIOS 17.6 includes Update CPU Micro Code.
2012-10-09, AMI BIOS 17.4 includes Update CPU Micro Code.
This is one of my motherboards.


https://rog.asus.com/motherboards/rog-strix/rog-strix-x299-e-gaming-model/helpdesk_bios/
BIOS updates for ASUS Strix X299-E for Intel's X299-based motherboards
2020/08/10, Version 3105 Update CPU Micro Code.
2019/10/04, Version 2002 adds support for Intel Core X-Series 10000 Processor Family
2018/03/20, Version 1301 Updated Intel CPU microcode, Improved system stability
2018/01/17, Version 1102, Updated CPU microcode.
2017/09/15, Version 0802, Update Intel X-series CPU (6-core and above) microcode
This is one of my motherboards before I sold it for ASUS ROG Strix X570-F


From https://videocardz.com/newz/reviewer-claims-amd-delayed-ryzen-7000-launch-by-two-weeks-due-to-bios-related-issues

AMD Zen 4's release date has been delayed for two weeks due to BIOS-related issues.

In any modern X86-64 CPU, the BIOS is an important part of the platform.

Reminder, AC68080 V2 has received its share of FPGA updates. Don't fool yourself.



After-sales support is very important.

You're the real-world liar.

-----

Back to the Amiga
https://aminet.net/package/game/shoot/jfduke3d
Duke 3D port bySzilard Biro
Vampire V2 users:
If the 060 executables don't work, please use the 040 executables. -Szilard Biro

You're the real-world liar.

Last edited by Hammer on 20-Aug-2022 at 10:17 AM.
Last edited by Hammer on 20-Aug-2022 at 10:14 AM.
Last edited by Hammer on 20-Aug-2022 at 10:14 AM.
Last edited by Hammer on 20-Aug-2022 at 10:11 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: I don't. Care to prove that I've forgotten it, dear liar? Anyway this is ANOTHER topic, which is NOT relevant to what we were discussing in THIS specific context. So, another Red Herring: you don't know where logic lives... In fact, and as I've already said, x86/x64 processors do NOT need such microcode update to work. Which is THE RELEVANT point. In the real world, a BIOS update for the existing motherboard is required to support any new CPUs.

In your parallel word.

My new PC with Intel's AlderLake 12900K booted and worked without problem at the first time, with the old firmware and without any update.
Quote:

Ask any AMD X470 motherboard owners with Zen 2 and Zen 3 CPUs.

That's a problem of those used that both this platform.
Quote:

With Windows 11 and certain game DRM that uses TPM, my X570 motherboard has a UEFI CPU microcode update to fix the stuttering issues with certain games and AMD's built-in fTPM. The alternative fix is to install a 3rd party TPM hardware. Reference: https://www.digitaltrends.com/computing/ryzen-tpm-stuttering-bug-windows-11-patch The BIOSes that contain the bug fix have the AGESA version 1207 or 1.2.0.7 microcode. Microcode updates come directly from AMD and are sent to the company’s motherboard partners, and from there it’s up to them to distribute the new microcode in the form of new BIOS updates Unlike Windows 11, Windows 10 doesn't assume TPM's existence.

It doesn't matter: Windows 11 works also without TPM. I've two Skylake PCs with it.
Quote:

I updated my X570 motherboard's UEFI BIOS for any potential Ryzen 7 5800X3D purchase.

Irrelevant.
Quote:

The BIOS update management has higher complexity when supporting multiple CPU generations (more than two generations) on the same CPU socket.

Irrelevant.
Quote:

Another example with Intel Z77 via MSI MPOWER motherboard https://www.msi.com/Motherboard/Z77_MPOWER/support 2013-07-22, AMI BIOS 17.10 includes Update IVB CPU Micro Code. 2013-01-31, AMI BIOS 17.7 includes Update CPU Micro Code. 2012-12-19, AMI BIOS 17.6 includes Update CPU Micro Code. 2012-10-09, AMI BIOS 17.4 includes Update CPU Micro Code.

Irrelevant.
Quote:

From https://videocardz.com/newz/reviewer-claims-amd-delayed-ryzen-7000-launch-by-two-weeks-due-to-bios-related-issues AMD Zen 4's release date has been delayed for two weeks due to BIOS-related issues.

Irrelevant.
Quote:

In any modern X86-64 CPU, the BIOS is an important part of the platform.

Nobody stated the contrary here. Unless you PROVE it, of course.
Quote:

Reminder, AC68080 has received its share of FPGA updates. Don't fool yourself.

So what?
Quote:

You're the real-world liar.

Don't change the cards on the table: I'm still waiting several PROOFs from what YOU stated before.

@cdimauro

Quote:

My new PC with Intel's AlderLake 12900K booted and worked without problem at the first time, with the old firmware and without any update.

Intel's AlderLake has a new LGA1700 motherboard and it's assumed it would out-of-the-box work with the correct pairing.

FYI, my ASUS X570-F/Zen 2/Windows 10 works out of the box and ASUS X570-E/Zen 3/Windows 10 works out of the box.


https://www.asus.com/Motherboards-Components/Motherboards/PRIME/PRIME-Z690-A/HelpDesk_BIOS/
For ASUS PRIME Z690-A.
2022/08/16, Version 1720 Update Microcode for next-generation Intel Processors.
2022/04/01, Version 1403 Update Intel Microcode
2022/01/27, Version 1003 Update Intel microcode.

In the corporate world, security is taken seriously.


https://www.tomshardware.com/news/intel-bios-update-disables-alder-lake-avx-512
Intel BIOS Updates Officially Disable Alder Lake's AVX-512 Capabilities


https://www.phoronix.com/news/Intel-May-2022-Security
Intel Releases New CPU Microcode For Latest Security Advisory (CVE-2022-21151)
The Intel for Linux 20220510 microcode release in addition to that security fix has various functional issues resolved too. This is their first Linux microcode CPU drop for Alder Lake processors while the updated CPU platforms range from Skylake and Valley View through Rocket Lake and Tiger Lake.

The updated Intel CPU microcode for Linux users can be found via GitHub while Windows users will likely find the microcode update coming down soon via a Windows update and/or via BIOS updates from motherboard vendors.


You're in a parallel world.

Quote:

It doesn't matter: Windows 11 works also without TPM. I've two Skylake PCs with it.

Read https://arstechnica.com/gaming/2021/09/riot-games-anti-cheat-software-will-require-tpm-secure-boot-on-windows-11/
Riot Games’ anti-cheat software will require TPM, Secure Boot on Windows 11.

You're in a parallel world.

Last edited by Hammer on 20-Aug-2022 at 10:32 AM.
Last edited by Hammer on 20-Aug-2022 at 10:31 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@cdimauro


[/img]Quote:

It doesn't matter: Windows 11 works also without TPM. I've two Skylake PCs with it.

From https://www.neowin.net/news/playing-valorant-on-windows-11-requires-tpm-20-and-secure-boot/
Riot Games' FTP competitive multi-player shooter Valorant has incorporated Trusted Platform Module (TPM) 2.0 and Secure Boot into its Vanguard Anti-Cheat game security software

[img]https://cdn.neow.in/news/images/uploaded/2021/09/1630826398_valorant_vanguard_tpm_2.0_(source-_anti-cheat_police_department_twitter).jpg[/img]

I consume modern PC games content. TPM matters for certain current games and in the future.

You're in dreamland.

Last edited by Hammer on 20-Aug-2022 at 11:51 AM.
Last edited by Hammer on 20-Aug-2022 at 10:35 AM.
Last edited by Hammer on 20-Aug-2022 at 10:34 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@cdimauro

https://overclock3d.net/news/software/windows_updates_have_fixed_most_of_alder_lake_s_gaming_issues_but_not_all_of_them/1
Intel and Microsoft fix most of Alder Lake's game compatibility issues.

This workaround effectively makes the Scroll Lock key a hotkey that disabled Alder Lake's E-Cores, turning Alder Lake into a P-Core only x86 processor.
-----
https://support.microsoft.com/en-us/topic/kb4093836-summary-of-intel-microcode-updates-08c99af2-075a-4e16-1ef1-5f6e4d8637c4
KB4093836: Summary of Intel Microcode Updates

This is the Intel microcode update example being distributed from the Microsoft Windows update service.


KB4093836 refers to https://www.intel.com/content/www/us/en/developer/articles/technical/software-security-guidance/advisory-guidance/processor-mmio-stale-data-vulnerabilities.html
Titled: Processor MMIO Stale Data Vulnerabilities


Processor MMIO Stale Data Vulnerabilities affected Intel CPU list
https://www.intel.com/content/www/us/en/developer/topic-technology/software-security-guidance/processors-affected-consolidated-product-cpu-model.html

06_97H, Stepping 5, Alder Lake S,

06_97H, Stepping 2, Alder Lake S,

Internet helps with CPU microcode update installs.

Security updates matter in the real corporate world.

From https://www.techarp.com/articles/amd-epyc-7000-series-tech-report/3/
Under "Cloud Datacenter and Enterprise Customers", I work in one of AMD's mentioned corporate partners.

You're in dreamland.

Last edited by Hammer on 20-Aug-2022 at 11:59 AM.
Last edited by Hammer on 20-Aug-2022 at 11:54 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@cdimauro




Quote:

That's why I've said that the Apollo's 68080 is more like Intel: it was Intel that reused the FPU to introduce its first SIMD unit, the MMX .

https://www.youtube.com/watch?v=fwo2hrzzaC8
RETROCENGO's Vampire V1200 issues.

You're an idiot.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer: again, you continue to edit and ADD stuff to already posted messages. Do you understand that you should only fix already written messages and write NEW ones for the new stuff? It that difficult for you, or the culprit is, again, the bad Mother Nature?

Quote:

Hammer wrote: Back to the Amiga https://aminet.net/package/game/shoot/jfduke3d Duke 3D port bySzilard Biro Vampire V2 users: If the 060 executables don't work, please use the 040 executables. -Szilard Biro

So what? V2 have a limited FPGA compared to V4 and subsequent. Do you know it? I think no, because you're the usual ignorant...

Anyway, could you please explain me why there are 040 and 060 executables? Is there any possibility that those processors where... INCOMPATIBLE (each other)? Do you know why?
Quote:

You're the real-world liar.

Care to PROVE why, dear mystifier?
Quote:

Hammer wrote: @cdimauro Quote: My new PC with Intel's AlderLake 12900K booted and worked without problem at the first time, with the old firmware and without any update. Intel's AlderLake has a new LGA1700 motherboard and it's assumed it would out-of-the-box work with the correct pairing. FYI, my ASUS X570-F/Zen 2/Windows 10 works out of the box and ASUS X570-E/Zen 3/Windows 10 works out of the box.

Do you understand that both things that YOU reported CONTRADICT your previous sentence on the topic? Here is it:
"In the real world, a BIOS update for the existing motherboard is required to support any new CPUs."

So, elementary logic at the hand, the update was NOT required.
Quote:

https://www.asus.com/Motherboards-Components/Motherboards/PRIME/PRIME-Z690-A/HelpDesk_BIOS/ For ASUS PRIME Z690-A. 2022/08/16, Version 1720 Update Microcode for next-generation Intel Processors. 2022/04/01, Version 1403 Update Intel Microcode 2022/01/27, Version 1003 Update Intel microcode.

Padding...
Quote:

In the corporate world, security is taken seriously.

Indeed. But it wasn't relevant. As usual...
Quote:

https://www.tomshardware.com/news/intel-bios-update-disables-alder-lake-avx-512 Intel BIOS Updates Officially Disable Alder Lake's AVX-512 Capabilities

I've the AVX-512 still enabled (when I want: there's a BIOS option) on my 12900K and works perfectly, because I've (specifically) bought an MSI mother board. This happens even if I flash a new BIOS in future, because MSI allows to select an older microcode which still supports the AVX-512.

Plus, I've one of the last Alder Lakes (directly coming from Santa Clara) without the laser cut to remove the AVX-512.

But, more important: this is your usual non-sense / off-topic. Hammer's PADDING...
Quote:

https://www.phoronix.com/news/Intel-May-2022-Security Intel Releases New CPU Microcode For Latest Security Advisory (CVE-2022-21151) The Intel for Linux 20220510 microcode release in addition to that security fix has various functional issues resolved too. This is their first Linux microcode CPU drop for Alder Lake processors while the updated CPU platforms range from Skylake and Valley View through Rocket Lake and Tiger Lake. The updated Intel CPU microcode for Linux users can be found via GitHub while Windows users will likely find the microcode update coming down soon via a Windows update and/or via BIOS updates from motherboard vendors.

See above.
Quote:

You're in a parallel world.

Don't recycle my statements: you lack creativity.

Specifically, don't use them if they aren't not true: see above. In fact, I can still use the old microcode which continues to enable the AVX-512.

That's because you're a complete ignorant. As usual. You talk of things that you've no clue at all, because you randomly search internet some stuff which might give you some points (which is never the case, of course, and as this thread clearly proves).

The main problem, however, is that you're talking with one which OWNS the specific product. So, I've experience and I know very well the situation.

If you were smart you would have avoided to talk of it, since I've explicitly and already told you that I own a 12900K. But you aren't smart and you have an irresistible desire to publicly show your ignorance and that you're a clown...
Quote:

Quote: It doesn't matter: Windows 11 works also without TPM. I've two Skylake PCs with it. Read https://arstechnica.com/gaming/2021/09/riot-games-anti-cheat-software-will-require-tpm-secure-boot-on-windows-11/ Riot Games’ anti-cheat software will require TPM, Secure Boot on Windows 11. You're in a parallel world.

First, in my real world I could AVOID buying those videogames and continue to use almost two thousand games of my libraries.

Second and, as I've already told you and you don't understand, I've an Alder Lake: so I've already a built-in TPM module. Running on Windows 11, of course. So, I can play those games, dear CLOWN!

Third, I could easily buy a couple of TPM modules for my Skylakes, and the problem would have been solved as well. CLOWN!

Forth, and that's the most important point, what you reported has NOTHING to do with the discussion. So, it's the usual Red Herring that you like a lot, since you're the king of logical fallacies...
Quote:

Hammer wrote: @cdimauro Quote: It doesn't matter: Windows 11 works also without TPM. I've two Skylake PCs with it. From https://www.neowin.net/news/playing-valorant-on-windows-11-requires-tpm-20-and-secure-boot/ Riot Games' FTP competitive multi-player shooter Valorant has incorporated Trusted Platform Module (TPM) 2.0 and Secure Boot into its Vanguard Anti-Cheat game security software [img]https://cdn.neow.in/news/images/uploaded/2021/09/1630826398_valorant_vanguard_tpm_2.0_(source-_anti-cheat_police_department_twitter).jpg[/img] I consume modern PC games content. TPM matters for certain current games and in the future. You're in dreamland.

Already answered above, PARROT!
Quote:

Hammer wrote: @cdimauro https://overclock3d.net/news/software/windows_updates_have_fixed_most_of_alder_lake_s_gaming_issues_but_not_all_of_them/1 Intel and Microsoft fix most of Alder Lake's game compatibility issues. This workaround effectively makes the Scroll Lock key a hotkey that disabled Alder Lake's E-Cores, turning Alder Lake into a P-Core only x86 processor. -----

This is a software fix on WINDOWS that has NOTHING to do with microcode etc.
Quote:

https://support.microsoft.com/en-us/topic/kb4093836-summary-of-intel-microcode-updates-08c99af2-075a-4e16-1ef1-5f6e4d8637c4 KB4093836: Summary of Intel Microcode Updates This is the Intel microcode update example being distributed from the Microsoft Windows update service.

Which is ONLY FOR SECURITY reasons. IF you have read it, of course, and which is NOT the case with you...
Quote:

KB4093836 refers to https://www.intel.com/content/www/us/en/developer/articles/technical/software-security-guidance/advisory-guidance/processor-mmio-stale-data-vulnerabilities.html Titled: Processor MMIO Stale Data Vulnerabilities Processor MMIO Stale Data Vulnerabilities affected Intel CPU list https://www.intel.com/content/www/us/en/developer/topic-technology/software-security-guidance/processors-affected-consolidated-product-cpu-model.html 06_97H, Stepping 5, Alder Lake S, 06_97H, Stepping 2, Alder Lake S, Internet helps with CPU microcode update installs.

Same as above.
Quote:

Security updates matter in the real corporate world.

Already answered above, PARROT!
Quote:

From https://www.techarp.com/articles/amd-epyc-7000-series-tech-report/3/ Under "Cloud Datacenter and Enterprise Customers", I work in one of AMD's mentioned corporate partners. You're in dreamland.

Hammer's usual non-sense / off-topic PADDING...
Quote:

Hammer wrote: @cdimauro Quote: That's why I've said that the Apollo's 68080 is more like Intel: it was Intel that reused the FPU to introduce its first SIMD unit, the MMX . https://www.youtube.com/watch?v=fwo2hrzzaC8 RETROCENGO's Vampire V1200 issues. You're an idiot.

And now again personal insults, which clearly shows that you're so desperate because you aren't able to sustain the discussion and you've to resort to them. Poor baby: go crying to your mammy.

Plus, it's the usual non-sense / off-topic AKA Hammer's padding, since it says absolutely nothing about the discussed topic.

Have I said that you're the official CLOWN of the forum? Ah, yes.

@cdimauro

https://www.cpsc.gov/Recalls/2022/ASUS-Computer-International-Recalls-ASUS-ROG-Maximus-Z690-Hero-Motherboards-Due-to-Fire-and-Burn-Hazards
ASUS's Intel Z690 motherboard recall!

ASUS ROG Z690 motherboards have a safety issue.


Name of Product: ASUS ROG Maximus Z690 Hero Motherboards

Hazard: A capacitor on the motherboard was installed in a reverse position which can lead to a short circuit, overheating or melting, posing fire and burn hazards.

Remedy: Replace

Recall Date: August 18, 2022

Units: About 10,000



Quote:

I've the AVX-512 still enabled (when I want: there's a BIOS option) on my 12900K and works perfectly, because I've (specifically) bought an MSI mother board. This happens even if I flash a new BIOS in future, because MSI allows to select an older microcode which still supports the AVX-512. Plus, I've one of the last Alder Lakes (directly coming from Santa Clara) without the laser cut to remove the AVX-512.

Intel has officially limited the AVX-512 deployment.

https://www.tomshardware.com/news/intel-nukes-alder-lake-avx-512-now-fuses-it-off-in-silicon
Intel nukes Alder Lake's AVX-512 Support, Now Fuses It Off in Silicon!

MSI's selective microcode has been defeated. You're late.

Quote:

This is a software fix on WINDOWS that has NOTHING to do with microcode etc.

Updated Windows CPU driver for Intel Alderlake restricts games from being scheduled into E-cores. This relates to Intel CPU performance issues.

Quote:

Care to PROVE why, dear mystifier?

https://www.youtube.com/watch?v=fwo2hrzzaC8&t=1795s
Vampire V1200 vs WARP 1260 Commodore Amiga 1200 Accelerator card comparison. Get wrecked.


https://aminet.net/package/game/shoot/jfduke3d

Vampire V2 users:
If the 060 executables don't work, please use the 040 executables.


Your defense for the Apollo team is BS. V1200 has legacy issues, hypocrite.


Quote:

Do you understand that both things that YOU reported CONTRADICT your previous sentence on the topic? Here is it: "In the real world, a BIOS update for the existing motherboard is required to support any new CPUs." Padding...

Again, https://www.asus.com/Motherboards-Components/Motherboards/PRIME/PRIME-Z690-A/HelpDesk_BIOS/
For ASUS PRIME Z690-A.
2022/08/16, Version 1720 Update Microcode for next-generation Intel Processors. (for Intel RaptorLake!)

You can't handle the truth.

You can't handle the timeline.

Quote:

But, more important: this is your usual non-sense / off-topic. Hammer's PADDING...

You're ignorant. You can't handle the timeline.

Last edited by Hammer on 21-Aug-2022 at 04:33 AM.
Last edited by Hammer on 21-Aug-2022 at 04:32 AM.
Last edited by Hammer on 21-Aug-2022 at 04:31 AM.
Last edited by Hammer on 21-Aug-2022 at 04:30 AM.
Last edited by Hammer on 21-Aug-2022 at 04:27 AM.
Last edited by Hammer on 21-Aug-2022 at 04:23 AM.
Last edited by Hammer on 21-Aug-2022 at 04:22 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Quote:

Hammer wrote: @cdimauro https://www.cpsc.gov/Recalls/2022/ASUS-Computer-International-Recalls-ASUS-ROG-Maximus-Z690-Hero-Motherboards-Due-to-Fire-and-Burn-Hazards ASUS's Intel Z690 motherboard recall! ASUS ROG Z690 motherboards have a safety issue. Name of Product: ASUS ROG Maximus Z690 Hero Motherboards Hazard: A capacitor on the motherboard was installed in a reverse position which can lead to a short circuit, overheating or melting, posing fire and burn hazards. Remedy: Replace Recall Date: August 18, 2022 Units: About 10,000

First, the usual non-sense / off-topic Hammer's PADDING.

Second, I already know it.

Third, it's a motherboard issue: it can happen on for boards for any processors. Whereas using the ASMedia chipset for the additional USB ports on AMD's Ryzen processor isn't an issue, but a proper decision decision which let several people experience issues with USBs due to the well know bad implementation of this chipset.
Quote:

Quote: I've the AVX-512 still enabled (when I want: there's a BIOS option) on my 12900K and works perfectly, because I've (specifically) bought an MSI mother board. This happens even if I flash a new BIOS in future, because MSI allows to select an older microcode which still supports the AVX-512. Plus, I've one of the last Alder Lakes (directly coming from Santa Clara) without the laser cut to remove the AVX-512. Intel has officially limited the AVX-512 deployment. https://www.tomshardware.com/news/intel-nukes-alder-lake-avx-512-now-fuses-it-off-in-silicon Intel nukes Alder Lake's AVX-512 Support, Now Fuses It Off in Silicon!

What wasn't clear to you about this:

"I've one of the last Alder Lakes (directly coming from Santa Clara) without the laser cut to remove the AVX-512."

PARROT?
Quote:

MSI's selective microcode has been defeated. You're late.

Again, what wasn't clear to you about this:

"I've the AVX-512 still enabled (when I want: there's a BIOS option) on my 12900K and works perfectly"

useless and hopeless PARROT?

You do NOT read what people write and repeat totally NON-SENSE stuff on things which were ALREADY KNOWN and, most important, ALREADY CLARIFIED. What's IQ level?!?

As said, Mother Nature was a very bad stepmother with you...
Quote:

Quote: This is a software fix on WINDOWS that has NOTHING to do with microcode etc. Updated Windows CPU driver for Intel Alderlake restricts games from being scheduled into E-cores. This relates to Intel CPU performance issues.

Which has NOTHING to do with their MICROCODE, right? It's a SOFTWARE-ONLY issue with WINDOWS that was solved.

Do you understand the difference? No, I don't think so, because it's proved that you're a reject of Mother Nature...
Quote:

Quote: Care to PROVE why, dear mystifier? https://www.youtube.com/watch?v=fwo2hrzzaC8&t=1795s Vampire V1200 vs WARP 1260 Commodore Amiga 1200 Accelerator card comparison. Get wrecked.

Which means NOTHING and, most important, PROVES NOTHING about your false sentences to me.

Useless and hopeless reject of Mother Nature.
Quote:

https://aminet.net/package/game/shoot/jfduke3d Vampire V2 users: If the 060 executables don't work, please use the 040 executables.

Already replied on the previous comment, reject:

"Anyway, could you please explain me why there are 040 and 060 executables? Is there any possibility that those processors where... INCOMPATIBLE (each other)? Do you know why?"

but you said nothing, eh? Who knows why, HOPELESS PARROT.
Quote:

Your defense for the Apollo team is BS. V1200 has legacy issues, hypocrite.

I don't defend the Apollo team. As I've ready said and you don't read and/or neither understand because of your evident intrinsic limits, I had a butcher fight with Gunnar.

Besides that, I only defend FACTs and REALITY.
Quote:

Quote: Do you understand that both things that YOU reported CONTRADICT your previous sentence on the topic? Here is it: "In the real world, a BIOS update for the existing motherboard is required to support any new CPUs." Padding... Again, https://www.asus.com/Motherboards-Components/Motherboards/PRIME/PRIME-Z690-A/HelpDesk_BIOS/ For ASUS PRIME Z690-A. 2022/08/16, Version 1720 Update Microcode for next-generation Intel Processors. (for Intel RaptorLake!) You can't handle the truth. You can't handle the timeline. Quote: But, more important: this is your usual non-sense / off-topic. Hammer's PADDING... You're ignorant. You can't handle the timeline.

No, the main problem here is that you're not able to understand even elementary logic.

Brief recap: you've written a sentence which has an ABSOLUTE value (read: NO exceptions allows):

"In the real world, a BIOS update for the existing motherboard is required to support any new CPUs."

which was easily dismantled by me before and, after that, even BY YOU.

Again, ELEMENTARY LOGIC. Something that you completely lack. And not only that: you are likely affected by functional illiteracy, since you're not even able to connect the points on what is written.

So, it's confirmed that you're a reject of Mother Nature.

This thread ...

_________________
Doing stupid things for fun...

@cdimauro

Quote:

Hammer: again, you continue to edit and ADD stuff to already posted messages. Do you understand that you should only fix already written messages and write NEW ones for the new stuff?

I do that too, usually because posting from a mobile device is fraught with autocorrection issues and thumb typos. For a time I used swipe input but it's basically learned how to suggest everything except the word I intended.

I clear the damn thing regularly but it seems to be on a mission to speak for me

_________________
Doing stupid things for fun...

@Karlos: Understood.

However I don't think that Hammer is posting from a mobile phone, according to the massive amount of things that he writes and the mistakes that he makes. He's likely using a PC and a keyboard & mouse. So, he's not justified.

@cdimauro

Quote:

First, the usual non-sense / off-topic Hammer's PADDING. Second, I already know it. Third, it's a motherboard issue: it can happen on for boards for any processors. Whereas using the ASMedia chipset for the additional USB ports on AMD's Ryzen processor isn't an issue, but a proper decision decision which let several people experience issues with USBs due to the well know bad implementation of this chipset.

You started the off-topic AMD vs Intel debate, hypocrite.

The USB disconnection issues have been resolved with AGESA 1.2.0.2 update.

My ASUS ROG Strix 570-E, Strix X570-F, and Gigabyte X570 UD motherboards are at AMD AM4 AGESA V2 PI 1.2.0.6b. Motherboard's software makes firmware updates easy.

Windows Update service also supplies P-Core/E-Core CPU driver and Intel microcode updates that include Intel Alder Lake CPUs.

I'm not against good aftersales support.

Quote:

"I've one of the last Alder Lakes (directly coming from Santa Clara) without the laser cut to remove the AVX-512."

That's a useless statement for the rest of the market. The PC market doesn't revolve around you!


Quote:

Again, what wasn't clear to you about this: "I've the AVX-512 still enabled (when I want: there's a BIOS option) on my 12900K and works perfectly" useless and hopeless PARROT? You do NOT read what people write and repeat totally NON-SENSE stuff on things which were ALREADY KNOWN and, most important, ALREADY CLARIFIED. What's IQ level?!? As said, Mother Nature was a very bad stepmother with you...

Your statement is useless when a new user purchases an Intel Adlerlake/Z670 motherboard combo as of today.

The PC market doesn't revolve around you!


Quote:

I don't defend the Apollo team. As I've ready said and you don't read and/or neither understand because of your evident intrinsic limits, I had a butcher fight with Gunnar. Besides that, I only defend FACTs and REALITY. Quote:

Your FACTs and REALITY are based on your centric self. The PC market doesn't revolve around you!

Apollo-Core has its own backward compatibility issues that require multiple FPGA firmware updates or FPGA V4 hardware update. You're fooling yourself when you presented they are saints for Amiga 68K legacy when Apollo-Core followed Motorola ColdFire's shortening FPU tactics.

No X86 CPU vendor has shortening X87 FPU compliance.

https://youtu.be/2QGFveBnNWU?t=1242
A1200 Vampire V2 RTG playing Duke Nukem 3D Atomic edition with jerky frame rates.
Date: October 2021


https://youtu.be/HTPWqEcjjds?t=132
Blizzard1260 equipped A1200 playing Duke Nukem 3D Atomic edition with smoother frame rates.
Date: June 2011

Again, the Apollo-core team is distracted by feature extensions.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Quote:

Again, the Apollo-core team is distracted by feature extensions

Other than full double and extended precision floating point, what is missing from the implementation? I realise those features will be critically important to some users, but for the guy with an 030 / no FPU or an LC040/060, thats not anything he'll miss.

I would personally prefer that the FPU implementation is brought up to at least full double precision but why should they do what some random person on the internet wants? If all you want to do is play quake with it, 32-bit single precision has you covered.

_________________
Doing stupid things for fun...

@Karlos

Right, it’s not an FPU, it’s a "quake enabler".

If it came at least with math libs for the OS that would’t do mental rounding errors, so at least simple math in arexx wouldn’t fail… but for all software that expects "proper" FPU present, expect awkward results.
For me, it’s a waste of space in the FPGA.

Last edited by kolla on 22-Aug-2022 at 11:43 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Pfftt. Like half the current retro crowd even know what Rexx is, let alone the Amiga version. Which in itself is a shame.

_________________
Doing stupid things for fun...