@ALL

And let's not forget DREAD
A Doom clone for Amiga 500 on Amiga 500 speeds using Amiga 500 hardware.

https://www.youtube.com/watch?v=UE4nOYc0RkA


Quote:
"005 AGIMA
11 months ago (edited)
Just played through the demo map on my stock A500 Rev6a with 512K slow mem expansion. I did find (thanks to a friend) that I had to unplug the external drive or it failed to load. That aside, I cannot believe I'm playing this on an A500. I'm simply STUNNED!!! No disrespect to the other Devs but this is better (so far) than Gloom or Alien Breed 3D (1) and it's running on OCS! This is some very clever black magic right here! Just.....WOW. "

Last edited by SHADES on 19-Aug-2022 at 12:38 AM.
Last edited by SHADES on 19-Aug-2022 at 12:36 AM.

_________________
It's not the question that's the problem, it's the problem that's the question.

@SHADES

The texture load is less due to 4-bits artwork and Dread's Blitter C2P works with 4-bit planes. Recent Dread builds have multi-level floors and Doom 1st level remakes.

Running on stock Amiga 500 1 MB RAM install base is impressive.

Retro PC also has improved performance with the Fast Doom fork.

The ease of programming with sufficient compute power vs development time is a major factor in commercial game development.

Gloom can run OCS/ECS on my old Amiga 3000/030 @ 25 Mhz, but it's closer to Wolfenstein 3D.

Last edited by Hammer on 19-Aug-2022 at 03:00 AM.
Last edited by Hammer on 19-Aug-2022 at 02:55 AM.
Last edited by Hammer on 19-Aug-2022 at 02:52 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)

Lou Quote:

I agree, this is not a thread about the Vampire ... but shouldn't you have included that in your benchmarks? ...as well as perhaps a Warp 1260 card which seems to be the cream of the crop. https://www.youtube.com/watch?v=fwo2hrzzaC8

The Vampire for the 1200 in the video has some serious issues and is an embarrassment compared to the Warp 1260. If Retrocengo could have found support for the issues he likely won't ever after posting this video (another needs verifying user on the apollo-core forum?). The Warp 1260 shows off the 68060 once it is uncorked by raising the clock speed and increasing the memory performance together (approaching twice the memory bandwidth of the common 68060@50MHz with 80ns memory from old 68060 accelerators). Considering the ancient chip fab process and only a 105MHz in order CPU, the 68060 impresses. After the 68040 had it's planned 50MHz max clock rate reduced to 40MHz due to heat, its nice to see the 68060 running significantly cooler at 105MHz (33.6C) than the ARM CPU (43.3C) and FPGA (62.9C) with their much newer fab processes. I don't think heat was as much of a problem for increasing the 68060 clock ratings as the shallow PPC pipeline designs the 68060 was threatening with its deep pipeline.

bhabbott Quote:

It's about whether packed pixels could have had a significant advantage on 1985-1992 hardware. Many Amiga fans have been saying that it would, some even suggesting the improvement would be huge.

cdimauro Quote:

There's no need of people's opinions: math is and should be enough to prove it.

I look forward to your mathematical proof but I think you will fail. I agree with bhabbott. I went back to the original Hypex post and carefully read and studied the data for awhile. Chip memory bandwidth is ~7MiB/s while Zorro III at 12-14MiB/s bandwidth is the bottleneck in writing to the memory of the graphics cards. Lets compare the WritePixelArray8 time/frame of the graphics card (12,528us/frame) to the Chunky2Planar time/frame of AGA (7427us/frame).

12,528/7,427=~1.69

7MiB/s (AGA chip memory) * 1.69 = ~11.81 MiB/s which is roughly the bandwidth of ZorroIII

The graphics card likely has a little bigger advantage but there is more function call overhead for WritePixelArray8() than Chunky2Planar(). It would be better to use the same function for both AGA and the graphic cards (possible with a switch?) and this should increase the performance of the graphics card if switching to Chunky2Planar() or decrease the performance of AGA if switching to WritePixelArray8(). Using the same graphics function, it should be possible to more accurately estimate the actual performance of the Zorro III bus and the bandwidth to the graphics card. My hypothesis is that the difference in fps is roughly due to the difference in write memory bandwidth to the different graphics memories. In addition to switching to the same graphics functions, the actual memory write bandwidth to the graphics cards should be measured with something like bustest.

http://aminet.net/package/util/moni/bustest

I believe data from the original post also verifies how cheap the C2P conversion is. The C2P conversion takes 1.3%-3.5% longer than the simple chunky copy method.

CopyMemQuick 12,528/12,106=~1.035 (~3.5% C2P overhead compared to CopyMemQuick)
Quick (MOVE16) 12,528/12,366=~1.013 (~1.3% C2P overhead compared to the MOVE16 copy)

The most efficient copy method should be employed so we take the best case of 1.3% of overhead for the C2P conversion which is, in my opinion, practically copy speed. It looks to me like the performance advantage of packed is not as great as the performance advantage of chip memory bandwidth not that CBM improved either. I still believe packed has advantages but I'm not convinced that "the improvement would be huge" or even a "significant advantage" in performance as bhabbot asserts. A 68060 in every Amiga certainly would have made a much bigger difference in performance as it is responsible for most of the game performance as well as allowing copy speed C2P conversion. Sadly, CBM didn't give us CPU performance either. That is the biggest reason why the Amiga became a budget retro PC already in the '90s and also why PPC Amigas are already retro without the budget part.

@matthey

Quote:

matthey wrote: bhabbott Quote: It's about whether packed pixels could have had a significant advantage on 1985-1992 hardware. Many Amiga fans have been saying that it would, some even suggesting the improvement would be huge. cdimauro Quote: There's no need of people's opinions: math is and should be enough to prove it. I look forward to your mathematical proof but I think you will fail.

As I've said, the article is already ready. So, I've already done the math and I've the numbers: I certainly cannot fail.
BTW, some numbers were already reported here, and not only by me. But nobody, even you, questioned them. Of course...

Anyway, it's vacation period and it wasn't possible to have a meeting with all my colleagues last weekend. Let's see this one. To me it's just a matter of pushing a button on WordPress and publish my article.
Quote:

I agree with bhabbott. I went back to the original Hypex post and carefully read and studied the data for awhile. Chip memory bandwidth is ~7MiB/s while Zorro III at 12-14MiB/s bandwidth is the bottleneck in writing to the memory of the graphics cards. Lets compare the WritePixelArray8 time/frame of the graphics card (12,528us/frame) to the Chunky2Planar time/frame of AGA (7427us/frame). 12,528/7,427=~1.69 7MiB/s (AGA chip memory) * 1.69 = ~11.81 MiB/s which is roughly the bandwidth of ZorroIII The graphics card likely has a little bigger advantage but there is more function call overhead for WritePixelArray8() than Chunky2Planar(). It would be better to use the same function for both AGA and the graphic cards (possible with a switch?) and this should increase the performance of the graphics card if switching to Chunky2Planar() or decrease the performance of AGA if switching to WritePixelArray8(). Using the same graphics function, it should be possible to more accurately estimate the actual performance of the Zorro III bus and the bandwidth to the graphics card. My hypothesis is that the difference in fps is roughly due to the difference in write memory bandwidth to the different graphics memories. In addition to switching to the same graphics functions, the actual memory write bandwidth to the graphics cards should be measured with something like bustest. http://aminet.net/package/util/moni/bustest I believe data from the original post also verifies how cheap the C2P conversion is. The C2P conversion takes 1.3%-3.5% longer than the simple chunky copy method. CopyMemQuick 12,528/12,106=~1.035 (~3.5% C2P overhead compared to CopyMemQuick) Quick (MOVE16) 12,528/12,366=~1.013 (~1.3% C2P overhead compared to the MOVE16 copy) The most efficient copy method should be employed so we take the best case of 1.3% of overhead for the C2P conversion which is, in my opinion, practically copy speed. It looks to me like the performance advantage of packed is not as great as the performance advantage of chip memory bandwidth not that CBM improved either. I still believe packed has advantages but I'm not convinced that "the improvement would be huge" or even a "significant advantage" in performance as bhabbot asserts.

First, bhabbot sentence was about the generic comparison about packed vs planar, and that's why he got my reply.

Second, the numbers for C2P cost were already reported (by me) from the original C2P routine, and aren't irrelevant as with the numbers that you're giving above. I mean: both the pure memory copy and calculations. They took a considerable number of raster lines for a 68060@50Mhz.
Quote:

A 68060 in every Amiga certainly would have made a much bigger difference in performance as it is responsible for most of the game performance as well as allowing copy speed C2P conversion. Sadly, CBM didn't give us CPU performance either. That is the biggest reason why the Amiga became a budget retro PC already in the '90s and also why PPC Amigas are already retro without the budget part.

The 68060 is a great processor for what it does, but it also arrived late on the market.

Commodore had its stupid makes, but Motorola wasn't second to it!

Last and not least, technology evolved much faster outside of the Amiga world. I mean: not only hardware but also software side.

It was the perfect storm for the Amiga, unfortunately...

@cdimauro

Quote:

Yup. Let's see if HyperX is interested on benchmarking them.

Oh no! More tests!

Quote:

No, I think that the approach is not correct. Of course, Doom is a 3D game, so it's very different from the 2D ones which were the more popular ones on the Amiga.

It's also somewhat backwards in some respects as it composites the screen by hand using only CPU. It's almost like older computers with limited hardware. Of course it did make use of available VGA features.

Quote:

However you're making tests on a unbalanced and much younger platform. I mean: an 68060@50Mhz running Doom greatly exceeds the specs of PCs running Doom when this game arrived.

It exceeds the specs of an average Amiga as well. Before PPC came on the scene such a setup was reserved for only the power user. And an Amiga user with the money to spend.

I don't recall many PCs having less specs than my A4000. Such as 33Mhz CPU clock. Most 80486 PCs I came across were 80486/DX-66. The MkII CyberStorm I have with 060 compatibility is dated 1996. And by then average PC was a Pentium,

Quote:

The bottleneck for your Amiga 4000 was/is the AGA chipset, and it was obvious that it creates its own problems even with your "beast". But your machine is much more powerful compared to a 386 or a 486, which were the most common CPUs used for running Doom. So, you're surpassing the AGA problems with a very powerful CPU + memory system.

If only it was more readily available back then. It would have cost more money to build. By the time my setup could be built PCs were more powerful.

Quote:

To be more clear, if your goal is to check the comparison between packed and planar, then you should use very very similar systems, with ideally only this feature which is different. Otherwise you aren't comparing packed vs planar, rather benchmark two different systems.

My A4000 setup is the only workable Amiga in the house.

I have an A1200 but I managed to break the IDE pins. I also have a DKB1240 card and SCSI but sold off my SCSI tower. So cannot presently attach a HDD to it. Can Doom run from floppy? Lol. If I find my CF card converter I may be able to test off that. But the slow CF card speed would possibly also slow the game down.

Quote:

So, the best test would be using systems where all is exactly the same, but with one using planar and the other using packed. Possible solution: changing WinUAE implementing an additional hardware configuration where there's only packed. Then you can test an Amiga 1000/500/2000/600 in cycle-exact mode for chipset & CPU, and it'll give a much more realistic result.

I've only got FS-UAE setup but it may have limitations. I tried to use Blizzard accelerator but it lacks the ROM. I can't select my model. But I don't trust emulation. Which is why I just got it working in UAE and once stable and working (and I stopped experimenting with ASM optimising) I hauled it over to my A4000 for the final testing results. It's hard to quantify as packed is simulated by simply treating the bitmap as packed. So really I'm cutting out C2P. But, packed may have given more bandwidth, which is used up by using all 8 bitplanes. I considered using 1 bitplane or even just 4 to save bandwidth and only display that but to me it seemed that would be cheating.

Quote:

Or, just implement single routines which simulate the display controller, or implement some primitives, and then you can benchmark them using exactly the same workload. However this will not benchmark an hardware system with packed and planar.

Yes, all sorts of ways about it. Which is why I decided to do it one way. Built around Doom. Suppose I could have used AB3D as well. Now the source can be compiled again. Or Quake which got an official port.

In any case it also provides results for RTG with my setup.

Quote:

20FPS is a very good result: very fluid, albeit not full-motion.

It is. Quite playable. Just five frames short, of half the average fast Amiga game. Although I spend more time watching demos of itself than playing. The normal demo without benchmark looks slightly faster and I've seen figures around 22FPS but it's not the timedemo that I used,

Quote:

And that's thanks to your "beast".

Yup!

Last edited by Hypex on 19-Aug-2022 at 03:50 PM.

@Hypex

Quote:

Oh no! More tests

You love it!

_________________
Doing stupid things for fun...

@Hypex

Quote:

Hypex wrote: @cdimauro Quote: [quote]So, the best test would be using systems where all is exactly the same, but with one using planar and the other using packed. Possible solution: changing WinUAE implementing an additional hardware configuration where there's only packed. Then you can test an Amiga 1000/500/2000/600 in cycle-exact mode for chipset & CPU, and it'll give a much more realistic result. I've only got FS-UAE setup but it may have limitations. I tried to use Blizzard accelerator but it lacks the ROM. I can't select my model. But I don't trust emulation. Which is why I just got it working in UAE and once stable and working (and I stopped experimenting with ASM optimising) I hauled it over to my A4000 for the final testing results.

Emulation can be trusted if you use 68000 + OCS/ECS with cycle-exact (on WinUAE).
Quote:

It's hard to quantify as packed is simulated by simply treating the bitmap as packed. So really I'm cutting out C2P. But, packed may have given more bandwidth, which is used up by using all 8 bitplanes. I considered using 1 bitplane or even just 4 to save bandwidth and only display that but to me it seemed that would be cheating.

For 3D games with 8-bits pixels width and resolutions like 256/320 pixels (horizontally) there's no difference between packed and planar, besides the C2P cost on Amigas.

That's why a packed vs planar comparison IMO doesn't make sense here (whereas it's fundamental for 2D games).


@matthey: the article publishing is delayed. We need to fix some technical issues on the site for handling comments by readers (we likely change the platform with a more modern one).

Anyway, I don't think that there's any urgent need to solve the question, right? We can survive some more weeks.

@matthey

Quote:

The Vampire for the 1200 in the video has some serious issues and is an embarrassment compared to the Warp 1260. If Retrocengo could have found support for the issues he likely won't ever after posting this video (another needs verifying user on the apollo-core forum?). The Warp 1260 shows off the 68060 once it is uncorked by raising the clock speed and increasing the memory performance together (approaching twice the memory bandwidth of the common 68060@50MHz with 80ns memory from old 68060 accelerators). Considering the ancient chip fab process and only a 105MHz in order CPU, the 68060 impresses. After the 68040 had it's planned 50MHz max clock rate reduced to 40MHz due to heat, its nice to see the 68060 running significantly cooler at 105MHz (33.6C) than the ARM CPU (43.3C) and FPGA (62.9C) with their much newer fab processes. I don't think heat was as much of a problem for increasing the 68060 clock ratings as the shallow PPC pipeline designs the 68060 was threatening with its deep pipeline.

I also spotted Retrocengo's V1200 issues last year (17, April 2021). This is why I stated the Apollo team has distractions e.g. AMMX, CoffinOS, SAGA.

Later in 2021, I purchased TF1260 (includes $40 68LC060 Rev 5, and I later purchased $200 68060 rev 1, reached 62.5 Mhz) despite its own flaws and I canceled my V1200 interest. I can reuse the 68060 CPU on another 68060 socket accelerator board.

Last edited by Hammer on 20-Aug-2022 at 12:00 PM.

_________________
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:

The 68060 is a great processor for what it does, but it also arrived late on the market. Commodore had its stupid makes, but Motorola wasn't second to it!

From https://techmonitor.ai/technology/motorola_plans_to_sample_the_68060_next_quarter
Date: April 19, 1994
Motorola Inc yesterday finally launched the long-promised 68060 follow-on to the 68040, claiming that it matches the performance of the Intel Corp Pentium at less than half the price – it costs $263 at 50MHz when you order 10,000 or more


68060 has half the asking price of Intel Pentium at +10,000 lots, but A1200 needs an adaptor card for 68060 with Fast RAM.

Commodore was bankrupt in Q2 1994.

https://www.amigareport.com/ar503/review12.html
By: Jason Compton

Phase 5's CyberStorm 060 is in the neighborhood of US $1500!



There's a problem after wholesale 68060 costs to adaptor card for 68060 with Fast RAM.
----


I don't have wholesale cost visibility for 68040.

Last edited by Hammer on 20-Aug-2022 at 12:31 PM.

_________________
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:

Amiga 4000 lacks a fast local bus for the video card and it doesn't help the overall frame buffer performance.

Of course for a dedicated video slot it should have had it. However my PicassoIV could keep up with the CyberVision. So it must offer some advantage over plain Zorro.

Quote:

486's fast VL-Bus concept has evolved into AGP and PCIe 16 lanes slots where the graphics card has a fast conduit with the host CPU. AGP is the PCI-era solution for a fast graphics slot.

PCI is what the Amiga missed. The best it had were ISA slots.

Quote:

In 1996, I have encountered a situation between upgrading my Amiga 3000 with CyberVision 64 and CyberStorm 68060 @ 50Mhz vs Buying an out-of-the-box PC clone with Pentium 150/S3 Trio 64UV graphics card/Yamaha Sonata 16bit sound card/1 GB Quantum Bigfoot IDE-based PC. Pentium 150 was overclocked to 166 Mhz with a 60 Mhz to 66 Mhz jumper.

Which did you choose?

Quote:

For Doom benchmarks from https://thandor.net/benchmark/32 Stealth 64 PCI 1MB PCI (with Pentium 100) scored 65.74 fps. Meanwhile, Diamond Viper V330 PCI (NVIDIA RIVA 128) scored 70.19 fps.

Good FPS.

Quote:

C= AGA's 19.6 fps is payable and performance is somewhere within the mid-range ISA SVGA cards.

Playable yes. At ISA speeds.

Quote:

ET4000AX 1MB ISA with Pentium 100 (and related 430VX PCI chipset and ISA bridge) has scored 27.10 fps.

It's faster than AGA but not by more than ten. So despite power increase in hardware it doesn't exactly kill it. By itself it fails to meet the Doom 35 standard.

Quote:

Amiga 1200 AGA couped with fast CPU has less bus mastering overheads when compared to Amiga 4000's Zorro III Super Buster chipset.

That could be the info I was looking for, thanks. I did a search but nothing came up related. About the A1200 being faster than the A4000 in some way.

Quote:

I'm in the not position to buy Amiga 1200 since my Dad purchased both an Amiga 3000 and 368DX-33+ET4000AX PC clone and around early 1992 to 1993 respectively. My Dad wasn't aware of Amiga 1200's Q4 1992 release window. My Dad wasn't happy about aging Amiga 3000's ECS after Amiga 1200's AGA release.

Yes that would have been annoying. And it was close as well. 1990 and 1992. if it only took 2 years for AGA what were they doing sine 1985!?

ECS should been standard in 1987 then. The only Amiga my dad bought was an A500. For my Christmas one year. Thanks to me there were only Amigas in the house. So all other systems I introduced him to were foreign to him. Windows. Linux. He couldn't use them as they weren't Amiga.

I found out a few years later I actually had an ECS Amiga. PD games that messed up the screen gave a clue. But OS1.3 didn't support ECS features like the screenmodes so it was rather useless for ECS.

Quote:

https://www.youtube.com/watch?v=lH7l_sU-mCk With 68060 @ 100Mhz, Amiga 1200 AGA vs RTG running Doom II. Amiga 1200 AGA frame rate is higher than your Amiga 4000 setup. The real-time frame rate counter is in the top right corner.

That is a fast one. It's double my CPU clock which would help. And the RTG card is newer. Something unavailable 20 years ago.

But, what I notice about these videos is that they don't do the actual speed test, and instead play it themselves. So it can't give a direct comparison to my tests. That may be more boring without the FPS counter and no sound but that's how it's done. Performing a timedemo. From DEMO1.

There's likely some test results like mine somewhere. Attached to a video could be good to make it more interesting. I don't know if there is an Amiga Doom speed database.

@Karlos

Quote:

You love it!

That's why I've been fiddling for two months. Then decided I need to stop. Of course my point was simple. To measure speed without C2P conversion.

I want to eventually open source my Doom mod. After I clean up the mess which was part of my plan. Then others can continue the testing! It could make it more useful with a faster C2P. My packed-planar mode isn't very practical.

@cdimauro

Quote:

Emulation can be trusted if you use 68000 + OCS/ECS with cycle-exact (on WinUAE).

Yes, but an A500 config is as practical as using DosBox to emulate Doom on a 286.

I did disable cache however and the speeds were more realistic. About the only time I've wanted slower emulation.

I needed too as I needed to use CodeProbe for testing at times. And the cache would stall it. Don't know if was the 060 emulation or if CPR didn't fully support later CPU models since it's on emulation.

Quote:

For 3D games with 8-bits pixels width and resolutions like 256/320 pixels (horizontally) there's no difference between packed and planar, besides the C2P cost on Amigas.

Yes, that is equal, though I did wonder if packed would have slightly less strain on the hardware and provide a small performance increase.

Quote:

That's why a packed vs planar comparison IMO doesn't make sense here (whereas it's fundamental for 2D games).

It's was a simple idea I just wanted to test. My results are limited to my "beast" config. But I was also interested in how native and RTG performed so my end results provide more.

Quote:

Anyway, I don't think that there's any urgent need to solve the question, right? We can survive some more weeks.

Not any more. You can relax since I posted mine! But, how many weeks, two more weeks?

@Hypex

Quote:

Of course for a dedicated video slot it should have had it. However my PicassoIV could keep up with the CyberVision. So it must offer some advantage over plain Zorro.

A local bus is needed for any video card.

The Amiga under Commodore/ESCOM lost its way and reverted to an old-school PC mindset while 486 gained VL-Bus (VESA Local Bus) which are slots designed for video cards.

PC GPU hardware evolution is largely driven by the gaming PC market competition i.e. high performance graphics are the heart of it. From the PC gaming market, PC GPU companies such as NVIDIA crushed SGI's hardware OpenGL business.

NVIDIA's legal battles also crushed SGI.

Quote:

PCI is what the Amiga missed. The best it had were ISA slots

Zorro III slot is not the problem, it's the Super Buster chip.

Super Buster chip needs to be upgraded to close the gap between real and theoretical numbers.
32-bit x 25 Mhz bus would yield a theoretical 100 MB/s while real work performance is about 15 MB/s


Quote:

Which did you choose

I selected PC, but I would have kept the A1200.

My purchase for A1200 in Y2020 was luck since it was advertised as a "sold for parts" wreak and I took the chance and paid $200 for it. A1200 was found to be working. A1200 was recapped and I fixed the timing issue. A fully working A1200's price is crazy.


Quote:

It's faster than AGA but not by more than ten. So despite power increase in hardware it doesn't exactly kill it. By itself it fails to meet the Doom 35 standard.

Some ET4000AX reached 33 fps. ET4000AX comes in either high-speed 16-bit VRAM or 32-bit DRAM SKUs. Your mileage may vary.

Overclocking the ISA bus is an option.

Quote:

That could be the info I was looking for, thanks. I did a search but nothing came up related. About the A1200 being faster than the A4000 in some way.

A4000's CPU cards would need gue chips between 68040 bus to 68030 bus.

A4000's motherboard wasn't designed for 68040, let alone 68060.
A4000's motherboard was an AA3000+ motherboard.

Quote:

Yes that would have been annoying. And it was close as well. 1990 and 1992. if it only took 2 years for AGA what were they doing sine 1985!? ECS should been standard in 1987 then. The only Amiga my dad bought was an A500. For my Christmas one year. Thanks to me there were only Amigas in the house. So all other systems I introduced him to were foreign to him. Windows. Linux. He couldn't use them as they weren't Amiga. I found out a few years later I actually had an ECS Amiga. PD games that messed up the screen gave a clue. But OS1.3 didn't support ECS features like the screenmodes so it was rather useless for ECS.

Amiga Ranger chipset with 128 colors (7-bit planes) and 4096 color palette should have been in 1987. 32-bit DRAM would be more than enough for 128 colors (7-bit planes) 320x200/256 resolution.
The original Amiga team worked on the Amiga Ranger chipset after the OCS.

1995 OCS (32 colors, 4096 color palette) release.
1987 Ranger (128 colors, 4096 color palette) release.
1990 AGA (256 colors, 16M color palette) release. AGA competed in March 1991. Mr Bill "IBM PCjr" Sydnes wasted time on ECS Amiga with faster CPUs development.
Commodore wasted A3000's faster 32-bit Chip RAM with 16-bit era ECS.
For the most part, ECS doesn't advance the Amiga gaming business.

A3000 is effectively redundant when A2500/030 has a similar capability. My Dad fell for the large model number on A3000 with a 68030 CPU. My Dad was one of those gamers who plays sports games.

ECS's 4 colors 640x480p mode is largely obsolete when VGA has 16 colors 640x480p mode. ET4000AX that was cost reduced IBM 8514/a clone was released in 1989.

My 1989 Amiga 500 has Rev 6 motherboard, ECS Agnus, Kickstart 1.3 and WB 1.3.2


Quote:

That is a fast one. It's double my CPU clock which would help. And the RTG card is newer. Something unavailable 20 years ago. But, what I notice about these videos is that they don't do the actual speed test, and instead play it themselves. So it can't give a direct comparison to my tests. That may be more boring without the FPS counter and no sound but that's how it's done. Performing a timedemo. From DEMO1

100 Mhz 68060 helps with compacting render time for the CPU stage while the AGA stage remained as is. AGA as a dumb frame buffer device is superior when compared to IBM's original VGA.

I don't have Warp1260 and my AMiga 1200/TF1260 only has 68060 rev 1 @ 62.5 and haven't applied the new firmware...

My preconfigured PiStorm/Pi 3a/Emu68/32 MicroSD is inbound from Poland. I plan to use Pi 3a to update the firmware on my TF1260. My Amiga 1200 also has Indivision AGA MK3 (with Graffiti).

https://www.youtube.com/watch?v=yfBOmXOKnKU
Amiga Graffiti add-on running Doom with 68030 @ 50Mhz.

http://wiki.icomp.de/wiki/Graffiti
Graffiti changes the Amiga bitplaned graphics into a chunky pixel mode. I'll get around to benchmarking it with Doom.

Last edited by Hammer on 20-Aug-2022 at 05:34 PM.

_________________
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: The 68060 is a great processor for what it does, but it also arrived late on the market. Commodore had its stupid makes, but Motorola wasn't second to it! From https://techmonitor.ai/technology/motorola_plans_to_sample_the_68060_next_quarter Date: April 19, 1994 Motorola Inc yesterday finally launched the long-promised 68060 follow-on to the 68040, claiming that it matches the performance of the Intel Corp Pentium at less than half the price – it costs $263 at 50MHz when you order 10,000 or more 68060 has half the asking price of Intel Pentium at +10,000 lots, but A1200 needs an adaptor card for 68060 with Fast RAM.

It was too late. When it arrived Pentium already reached double the frequency (100Mhz).
Quote:

Amiga Ranger chipset with 128 colors (7-bit planes) and 4096 color palette should have been in 1987.

It sucked on '87 with that specs.
Quote:

32-bit DRAM would be more than enough for 128 colors (7-bit planes) 320x200/256 resolution.

32-bit DRAMs weren't needed for that: the original chipset had enough bandwidth to display up to 256 colors at 320x200/256 with its 16-bit DRAMs.
Quote:

AGA as a dumb frame buffer device is superior when compared to IBM's original VGA.

Again? As I've already said, this comparison is completely non-sense: 1985 vs 1992...


@Hypex

Quote:

Hypex wrote: @cdimauro Quote: Emulation can be trusted if you use 68000 + OCS/ECS with cycle-exact (on WinUAE). Yes, but an A500 config is as practical as using DosBox to emulate Doom on a 286. I did disable cache however and the speeds were more realistic. About the only time I've wanted slower emulation.

It's very different with WinUAE: cycle-exact for 68000 + OCS/ECS is very realistic / accurate.
Quote:

Quote: For 3D games with 8-bits pixels width and resolutions like 256/320 pixels (horizontally) there's no difference between packed and planar, besides the C2P cost on Amigas. Yes, that is equal, though I did wonder if packed would have slightly less strain on the hardware and provide a small performance increase.

Only if you also have some 2D graphics on top of the 3D one: then the differences between packed and planar still applies, of course.
Quote:

Quote: Anyway, I don't think that there's any urgent need to solve the question, right? We can survive some more weeks. Not any more. You can relax since I posted mine! But, how many weeks, two more weeks?

Not two weeks, but I've no idea. It's a bad period: vacation time for many and the responsible has some family problems, so it's difficult for all member to meet.

But we aren't on a hospital, so no urgency.

@cdimauro

Quote:

Again? As I've already said, this comparison is completely non-sense: 1985 vs 1992...

Are you talking regarding VGA release?

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@cdimauro

Intel 820 chipset was released in November 1999, which is in the Pentium III era. This is during Intel vs AMD GHz race.

During Xmas year 2000, my PC was AMD K7 Athlon Thunderbird 1133 MHz.

Pentium II is an aging CPU during 1999 time period. My 1st AMD PC replaced my Intel Celeron 533/Intel 440ZX chipset. During Pentium II era, I was able assemble my own PCs from individual parts which keeps the cost relatively low. The competition among PC motherboard/ PC case/PC video card vendors was good. Constant new x86 SKU releases has downwards price pressures on lower clock speed x86 SKUs. The return of AMD vs Intel GHz competition makes the PC platform great again e.g. 6Ghz.

Last edited by Hammer on 20-Aug-2022 at 11:59 PM.
Last edited by Hammer on 20-Aug-2022 at 11:46 PM.

_________________
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:

t was too late. When it arrived Pentium already reached double the frequency (100Mhz).

68060 rev 1 at 62.5 Mhz overclocks works fine with a small passive heatsink. Amiga 1200 has a large metal RF shield that I could use to link with the heat pipes.

Classic Pentium requires an active cooling fan and a small heatsink. Give Pentium 60's fan and heat sink for 68060 rev 1, it will reach higher clock speeds. Intel has released CPU SKUs on the edge stability during the Mhz/Ghz race and GHz Pentium III has a product recall.

It was Motorola's decision to treat 68060 as a second-class CPU family below the PowerPC CPU family.

Quote:

It sucked on '87 with that specs.

128 color display is better than OCS's 32-color display and 64-color EHB workaround. Early Amiga 1000 NTSC does not have 64-color EHB mode.

The original Amiga engineers keep improving the Amiga chipset.

https://www.youtube.com/watch?v=GojpwZMBHz4
WIP arcade quality Final Fight port with 128 colors for the Amiga. This example runs on a dumb AGA frame buffer with the 68EC020 CPU handling soft blitter workloads on 32-bit Fast RAM.

Amiga OCS (after early Amiga 1000 chipset with missing EHB) already has 6-bit plane support via 64-color EHB.

128 color display Amiga Rander chipset is based on CSG's 1987/1988 chip fabrication capabilities.


Quote:

32-bit DRAMs weren't needed for that: the original chipset had enough bandwidth to display up to 256 colors at 320x200/256 with its 16-bit DRAMs.

In practice, it does not. Refer to the reasons why WIP arcade quality Final Fight port has conservative 128 color usage.

Extra memory bandwidth is required for audio and other non-graphics services. Chip RAM is not dedicated to graphics usage like on PC discrete graphics card.

Amiga 500's performance improves with an additional 16-bit Fast RAM i.e. less shared bus issues.

Quote:

Again? As I've already said, this comparison is completely non-sense: 1985 vs 1992...

PiStorm/Pi 3a/Emu68 can run Doom OCS at playable frame rates
https://www.youtube.com/watch?v=jo0dZ8eTErM


Vampire can run Doom OCS at playable frame rates
https://www.youtube.com/watch?v=J4gkxtiKQCE

Overclocked Blizzard 1260 66 Mhz with 64MB ram on an ACA500 with hacked firmware running Doom with Amiga 500 OCS.
https://www.youtube.com/watch?v=cwRQ2Iy_Y2Q


1987 Amiga Ranger chipset is close to Amiga AGA (completed in March 1991).

Last edited by Hammer on 21-Aug-2022 at 01:08 AM.
Last edited by Hammer on 21-Aug-2022 at 01:05 AM.
Last edited by Hammer on 21-Aug-2022 at 01:01 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)

@Hypex

Quote:

I have an A1200 but I managed to break the IDE pins. I also have a DKB1240 card and SCSI but sold off my SCSI tower. So cannot presently attach a HDD to it. Can Doom run from floppy? Lol. If I find my CF card converter I may be able to test off that. But the slow CF card speed would possibly also slow the game down.

I also broke my Witcher 508's pin, and I manage to solder a new pin.

For soldering work, I practiced on old non-Amiga PCBs e.g. PC gaming mouse that needs click button replacements.

I fixed timing issues on my Amiga 1200 rev 1d1 motherboard.

https://www.youtube.com/watch?v=_mbtfAa1fEQ
Fixing a pin on the Vampire example from GadgetUK164 - Retro Gaming Repairs & Mods.

Last edited by Hammer on 21-Aug-2022 at 01:22 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)

@pixie

Quote:

pixie wrote: @cdimauro Quote: Again? As I've already said, this comparison is completely non-sense: 1985 vs 1992... Are you talking regarding VGA release?

Yes.


@Hammer

Quote:

Hammer wrote: @cdimauro Intel 820 chipset was released in November 1999, which is in the Pentium III era. This is during Intel vs AMD GHz race. During Xmas year 2000, my PC was AMD K7 Athlon Thunderbird 1133 MHz. Pentium II is an aging CPU during 1999 time period. My 1st AMD PC replaced my Intel Celeron 533/Intel 440ZX chipset. During Pentium II era, I was able assemble my own PCs from individual parts which keeps the cost relatively low. The competition among PC motherboard/ PC case/PC video card vendors was good. Constant new x86 SKU releases has downwards price pressures on lower clock speed x86 SKUs. The return of AMD vs Intel GHz competition makes the PC platform great again e.g. 6Ghz.

Besides the usual padding, have you understood why I've reported the link to Pentium III's chipsets? If not, I can clarify...
Quote:

Hammer wrote: @cdimauro Quote: t was too late. When it arrived Pentium already reached double the frequency (100Mhz). 68060 rev 1 at 62.5 Mhz overclocks works fine with a small passive heatsink. Amiga 1200 has a large metal RF shield that I could use to link with the heat pipes. Classic Pentium requires an active cooling fan and a small heatsink. Give Pentium 60's fan and heat sink for 68060 rev 1, it will reach higher clock speeds. Intel has released CPU SKUs on the edge stability during the Mhz/Ghz race and GHz Pentium III has a product recall.

Hammer's padding. This has nothing do with the facts that I've reported: the 68060 was late and in the meanwhile the Pentium reached 100Mhz.
Quote:

It was Motorola's decision to treat 68060 as a second-class CPU family below the PowerPC CPU family.

Well known. Unfortunately. One of the long list of mistakes that Motorola did. Motorola wasn't second to Commodore for absurd and stupid decisions...
Quote:

Quote: It sucked on '87 with that specs. 128 color display is better than OCS's 32-color display and 64-color EHB workaround. Early Amiga 1000 NTSC does not have 64-color EHB mode. The original Amiga engineers keep improving the Amiga chipset.

Indeed, but this doesn't change the fact that 128 colors at 320x200/256 (and with 4096 palette) on '87 sucked. Competition already provided much better.
Quote:

https://www.youtube.com/watch?v=GojpwZMBHz4 WIP arcade quality Final Fight port with 128 colors for the Amiga. This example runs on a dumb AGA frame buffer with the 68EC020 CPU handling soft blitter workloads on 32-bit Fast RAM.

AGA is a different chipset compared to OCS and it's from '92: so SEVEN years after the OCS.
Quote:

Amiga OCS (after early Amiga 1000 chipset with missing EHB) already has 6-bit plane support via 64-color EHB.

I know very well, since I've worked to Fightin' Spirit, which heavily uses it.
Quote:

128 color display Amiga Rander chipset is based on CSG's 1987/1988 chip fabrication capabilities.

As said before, you don't need the Ranger chip to get the above.

And an EHB with 7 and 8 bitplanes without touching the rest could have been possible already on '85.
Quote:

Quote: 32-bit DRAMs weren't needed for that: the original chipset had enough bandwidth to display up to 256 colors at 320x200/256 with its 16-bit DRAMs. In practice, it does not. Refer to the reasons why WIP arcade quality Final Fight port has conservative 128 color usage.

See above, and you're talking about a game developed on AGA, so not OCS, with a 68020 instead of a 68000, and with an additional 64MB of fast RAM: an impossible monster on '85/87.
Quote:

Extra memory bandwidth is required for audio and other non-graphics services.

Not at all: those (audio, disk, sprites) take only a few cycles per scanline.
Quote:

Chip RAM is not dedicated to graphics usage like on PC discrete graphics card.

Yes, and I've used it also for running self-modifying code (removed on the final release of Fightin' Spirit).

But it doesn't change the picture.
Quote:

Amiga 500's performance improves with an additional 16-bit Fast RAM i.e. less shared bus issues.

On the exact contrary, and as I've already said several times, this is useless for 2D games (which was THE target for such machines).

As a game developer (Fightin' Spirit. USA Racing -> unreleased) I know very well the chipset AND its limits AND what's best/needed for the games to work as best as possible on that platform.
Quote:

Quote: Again? As I've already said, this comparison is completely non-sense: 1985 vs 1992... PiStorm/Pi 3a/Emu68 can run Doom OCS at playable frame rates https://www.youtube.com/watch?v=jo0dZ8eTErM Vampire can run Doom OCS at playable frame rates https://www.youtube.com/watch?v=J4gkxtiKQCE Overclocked Blizzard 1260 66 Mhz with 64MB ram on an ACA500 with hacked firmware running Doom with Amiga 500 OCS. https://www.youtube.com/watch?v=cwRQ2Iy_Y2Q

OK, and?
Quote:

1987 Amiga Ranger chipset is close to Amiga AGA (completed in March 1991).

No, AGA is way better (besides being an horrible patch over ECS). Unfortunately it was too late for industry's standards.

Anyway, not even AGA was what 2D games needed at the time.
The major problem with the Amiga chipset is that the Blitter (which is Amiga's strongest point/component) was left as it is: working with 16-bit and at 7Mhz.
The second, minor (but still important for the games!), problem was that Paula was left as it is: with only 4 audio channels (16-bit samples weren't that important). And the floppy which supported only normal 720/880kB disks (which means not only half the space available to store data: also half the speed to read/write them).

Commodore was a disaster for the Amiga...

@cdimauo
Quote:

Again? As I've already said, this comparison is completely non-sense: 1985 vs 1992...

Quote:

Are you talking regarding VGA release Quote: Yes

But from what I saw VGA only appeared after in 1987, that's why I got confused about the meaning.

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga