24 May 2024, 07:42 | #4701 | |
Registered User
Join Date: Aug 2020
Location: Sydney/Australia
Posts: 1,031
|
Quote:
|
|
24 May 2024, 08:15 | #4702 | |
Registered User
Join Date: Aug 2020
Location: Sydney/Australia
Posts: 1,031
|
"Access time" is different from "random read or write cycle time".
For example http://matthieu.benoit.free.fr/cross...s/HYB41256.pdf This is A500's DRAM example, For A500's DRAM, RAS has 150 ns, CAS has 75 ns, and "random read or write cycle time" is 260 ns. 260 ns translates to about 3.8 Mhz and OCS is clocked at 3.5 Mhz for a close match. AGA Lisa's double-pumped 3.5 Mhz (effective 7 Mhz) needs to be better than A500's DRAM quality, hence 140 ns for "random read or write cycle time". 140 ns translates to 7.14 Mhz. DRAM has an effective clock speed that is closely matched with the intended ASIC. AGA Lisa's 64-bit (32-bit double pump) x 3.5 Mhz is 28 MB/s. FP DRAM needs to be in 7 Mhz range quality. Memory interleaves can hide or mitigate memory latencies. I don't think Commodore hasn't mastered a memory interleave memory channel. Quote:
PS1 has NEC's SGRAM. Its development was around 1992-1993. For NDA road map and specification access, close partnerships with memory manufacturers are important. A modern example is the close relationship between NVIDIA and Micron with GDDR6X. Last edited by hammer; 24 May 2024 at 09:26. |
|
24 May 2024, 09:15 | #4703 | |
Registered User
Join Date: Aug 2020
Location: Sydney/Australia
Posts: 1,031
|
Quote:
C900 has two configurations i.e. one model with a Blitter with high-resolution monochrome graphics and another model has a MOS Technology 8563 video controller (recycled for C128's 80-column text mode). C900 was to be manufactured by Commodore West Germany. C900's two models had an Atari ST (no Blitter) vs Mega ST (with Blitter) or Plus4 (no hardware sprites) vs C64 (with hardware sprites) product segmentation. The C900 project was initiated in 1983 by Commodore systems engineers Frank W. Hughes, Robert Russell, and Shiraz Shivji. The fools don't understand 3rd party developers. Without Tramiel, Commodore offered EHB upgrades for A1000 owners with missing EHB mode. Last edited by hammer; 24 May 2024 at 09:28. |
|
24 May 2024, 09:47 | #4704 | |
Registered User
Join Date: Mar 2018
Location: Hastings, New Zealand
Posts: 2,719
|
Quote:
Furthermore the Los Gatos team had the wrong focus. They were designing machines for the classes, not the masses. The engineers had fallen in love with high-end workstations and wanted the Amiga to replicate them. Never mind that a 1024x1024 display would need a special monitor made just for that resolution (almost as silly as the A2024 'Hedley' monitor). Ranger was a boondoggle with impractical square expansion cards mounted horizontally, a cooling and maintenance nightmare. The A2000 was a much more sensible design. When the A2000 was ready for production the Los Gatos team had nothing to show - not even a breadboard prototype. Which design would you choose? The A2000 was basically an A1000 in a big box with Zorro slots. That meant it was fully compatible with existing Amiga software and would help grow the user base. It would also be attractive to those who wanted a more 'industrial' machine with excellent expansion capabilities. It could even be made IBM compatible - if you really needed that. The A2000-CR followed soon after, using the same chipset as the A500. Once again it grew the user base while having the potential to do much more. Several of my friends bought a base model A2000 with single floppy drive, which could be used just like an A500 or expanded with a hard drive, Fast RAM, accelerator card, graphics card etc. You could make it into anything you liked - and still play A500 games on it! The A2000 was the vehicle for NewTek's Video Toaster. Tons of specialized video production cards were made for it. Not many A2000s were sold compared to the A500, but video professionals would pay a fortune and be happy because it was still much cheaper than the dedicated stuff they would otherwise need. This alone helped many US Amiga retailers stay in business. |
|
24 May 2024, 10:11 | #4705 | |
Registered User
Join Date: Aug 2020
Location: Sydney/Australia
Posts: 1,031
|
Quote:
There's a reason why DRAM 260 ns (random read or write cycle time) and 3.5 Mhz were configured for A500. DRAM 260 ns for random read or write cycle time is effectively 3,846,153.846153846 hz or about 3.84 Mhz and Agnus has 3.5 Mhz clock speed. 1987-era VRAM example is NEC uPD41264C-12 with serial access of 40 ns time max and Row access of 120 ns time. https://www.ardent-tool.com/datasheets/NEC_uPD41264.pdf Pixel workload is usually sequential i.e. serial access mode. Need an ASIC that can keep up with 25 Mhz effective memory performance. |
|
24 May 2024, 10:13 | #4706 | |
Registered User
Join Date: Apr 2017
Location: France
Posts: 646
|
Quote:
I was alluding in terms of man willpower not tech decisions. The energies have to be oriented in the right direction and congruent. Management role. But Commodore seems to have been too much engineers driven for sure. Bill Herd deciding to do the C128 by himself because he saw an hole in the planning is just WTF and I said "Not that Jack should have stay". |
|
24 May 2024, 10:19 | #4707 | |
Registered User
Join Date: Mar 2018
Location: Hastings, New Zealand
Posts: 2,719
|
Quote:
The split definitely hurt Commodore badly. It was their main source of continuing financial problems and general turmoil. But they still managed to get some great machines out before finally succumbing a decade later. If Jack hadn't left would they have done better? The fate of the ST suggests otherwise. |
|
24 May 2024, 10:23 | #4708 |
Registered User
Join Date: Apr 2017
Location: France
Posts: 646
|
|
24 May 2024, 10:40 | #4709 | |||
Registered User
Join Date: Aug 2020
Location: Sydney/Australia
Posts: 1,031
|
Quote:
Slow RAM has the same bottlenecks as Chip RAM without DMA graphics/audio access i.e. CPU only address range. I have A500 Rev 5 with 8371 and A500 Rev 6A with 8372A. Quote:
IBM VGA wasn't for the masses until it was cost-reduced by the VGA cloners. A good cost reduction is an art in itself. You need two halves i.e. leading edge R&D and cost reduction teams. They would balance each other. NVIDIA's expensive Volta V100 led the RTX (emulated in software) and Tensor R&D before the Turing RTX masses. Quote:
Amiga OCS's 12-bit colors HAM's compressed color before J-MPEG were a wow factor. Your mindset is an "Atari ST" e.g. 512 color palette with just 16 colors, a slightly better Tandy Video I. https://en.wikipedia.org/wiki/Tandy_Graphics_Adapter ArtX (ex-SGI engineers, purchased by ATI), 3DFX (ex-SGI engineers, purchased by NVIDIA), and NVIDIA (key ex-SUN GX engineer) founders have the drive to bring workstation graphics to the masses. |
|||
24 May 2024, 11:42 | #4710 |
Computer Nerd
Join Date: Sep 2007
Location: Rotterdam/Netherlands
Age: 48
Posts: 3,839
|
|
24 May 2024, 14:36 | #4711 |
HOL/FTP busy bee
Join Date: Sep 2006
Location: Germany
Age: 46
Posts: 31,959
|
|
24 May 2024, 16:38 | #4712 |
Registered User
Join Date: Aug 2006
Location: Scunthorpe/United Kingdom
Posts: 2,087
|
|
24 May 2024, 16:46 | #4713 | |
Registered User
Join Date: Jan 2023
Location: Toronto
Posts: 419
|
Quote:
Considering the internal dissension over how Amiga was to be handled, I think Commodore did a fantastic job for the Amiga. They must have poured 40 or 50 million into the machine and advertising. The initial advertising wasn't productive, or smartly done, but it did cost a lot. Then they went a year or so with no advertising, because they didn't have any product, and it's only been in '88 that any advertising at all has started again. They had stopped making the 1000, and they expected the 500 and the 2000 to be ready a year earlier than they were. Of course, they weren't. I told Irving that they wouldn't be. What are you going to advertise if you can't deliver product? Now here's an interesting thought: Commodore didn't want to continue with the 1000. They thought it was too expensive and so instead of reducing the cost of production, they were trying to change it into a single unit, built-in keyboard thing, like the ST so it could compete better. At the same time. Atari was trying lo make their ST look more like a 1000, with a built-in drive and a separate keyboard. It was the stupidest thing. I got real exasperated with them. That's when I stopped taking it seriously. Now my only contact with Commodore is my keeping in touch with some of the designers. And I still go to the FAUG (First Amiga Users Group) meetings. It's a very close group. INFO: What would you have done differently if you had run Commodore? MINER: I would have cost-reduced the 1000; maybe given it a different name so it sounded like a new product. But I wouldn't have changed anything. It was a good machine compared to the 500. All it needed was a little cost reducing. I would have come out with it much sooner. I could have turned it around in a few months, rather than being caught without any product for over a year. Amiga lost so much momentum during that year because of lack of product and lack of advertising that I don't know if they'll ever recover. The software base wasn't built at all during that lime, and the IBM and Macintosh mushroomed. I wouldn't have bothered with IBM compatibility, which just slowed down the introduction of the machine. INFO: Do you see the IBM and the Macintosh as being in head-on competition with the Amiga? MINER: The Amiga is so far behind them in public recognition that it's sickening. First, there were bad ads, then the past year, there were no Amiga ads and people forgot about it. That's a shame because the Amiga is a better machine. The software people don't want to write for it partially because there are only 500,000 machines out there, and partially because Commodore treated their developers very badly. It's almost as if they didn't think they needed them at all. When the machine first came out, Commodore had very little service capability. There was literally no way to fix them. That was a big mistake. That cost a lot of public and dealer good will. Commodore would continuously have specials and undercut their dealers, or force dealers to take more machines than they wanted. Any dealer can tell you horror stories about Commodore. I'm not saying all of it was done intentionally, but it was done. Commodore wouldn't give machines to editors of large newspapers. They not only wouldn't give them machines, they wouldn't give them the time of day. At the same time Apple was giving machines to all the magazines and newspapers and schools. Commodore wouldn't even give two or three machines to major colleges. A couple to Stanford, or Berkeley, or MIT might have cost a few thousand, but that's peanuts lo what they spent on their stupid advertising, and the return would have been a lot better. That's the kind of mistake Commodore has made over and over again, and I'm not sure that they're not still making them. INFO: Do you see any hope for it? MINER: No. I'm sorry to say it, but no. Amiga is so far behind Macintosh and IBM now, and they've lost so much momentum and position, that I think it's going to be almost impossible to recover. Even if the Amiga goes South, a lot of the things that were done in the Amiga will continue to live, so people will have good machines. IBM's push for multitasking and Apple's push for sound and graphics were brought on by the Amiga. There's a lot of the Amiga in the Mac II, though not done as well, of course. For the next two or three years, Commodore may be able to stay competitive, but in the long run, they will be forced to come up with a new machine, and it will be hard to stay compatible with the Amiga. |
|
24 May 2024, 17:04 | #4714 |
Registered User
Join Date: Apr 2017
Location: France
Posts: 646
|
Thanks for pointing this very interesting interview.
I don't understand why you say "they went a year or so with no advertising, because they didn't have any product". Once the A1000 was on the production line, it was here, even if it was in small quantity due to production problems. So why the "they didn't have any product" ? You have a product, you can talk about it as long as it suits you. |
24 May 2024, 19:19 | #4715 | |
Registered User
Join Date: Jan 2023
Location: Toronto
Posts: 419
|
Quote:
So it sounds like it might be similar to what happened later when C= discontinued the A500 to make way for the A600. Just in this case when they discontinued the A1000 for the A500/A2000 they weren't actually ready. If it was true that Jay could have cost reduced the A1000 in mere months then things could have possibly been much different. |
|
24 May 2024, 19:35 | #4716 |
HOL/FTP busy bee
Join Date: Sep 2006
Location: Germany
Age: 46
Posts: 31,959
|
"The Amiga was killed by the marketing department of Commodore" Commodore had the machine, but weren't really interested in selling it because the C64 was selling like crazy in the mid 80s.
Bruce sometimes mentioned that we all should be grateful for any Amiga model we got, but I think it is not unfair to say that Commodore didn't use even half of the potential of any of the models (except the Amiga 500 which they handled mostly okay) they brought to market. |
24 May 2024, 20:56 | #4717 | |
Registered User
Join Date: Aug 2013
Location: Marseille / France
Posts: 1,510
|
Quote:
|
|
24 May 2024, 21:08 | #4718 | |
Registered User
Join Date: Oct 2020
Location: Bicester
Posts: 2,020
|
Quote:
the gap between it (the A500) and A1200 was too big and the A500+ probably should have been earlier with a 'bigger' upgrade than just the chip ram and os. Why not a 14mhz 68000 1mb chip (expandable to 2mb) and 1mb of 'real' fast ram? in say 89? that was not unthinkable was it? |
|
24 May 2024, 21:37 | #4719 | |
Global Moderator
Join Date: May 2013
Location: Setúbal, Portugal
Posts: 616
|
Quote:
No, not unthinkable. But a 14MHz 68000 (double that of the A500, Atari ST and SEGA Mega Drive. For a comparison, the X68000 came with a 68000@10MHz and the Neo-Geo with a 68000@12MHz) would make the Amiga A LOT more expensive. Maybe, right in 1987, the the A500 could have been the "baby Amiga" exactly as-was and then there would be a "grown up" Amiga as a modular machine with a 14MHz CPU frequency, with more RAM, a scan doubler and an Hard Disk Drive. AGA should have been 1990 tops. Ideally with an hardware C2P solution, even if the dreaded Akiko. But it should have been 1990 tops (ideally Christmas 1989). |
|
24 May 2024, 21:52 | #4720 |
Registered User
Join Date: Jun 2010
Location: PL?
Posts: 2,875
|
Doubling CPU clock, adding 64KB cache with CSG made cache controller probably will be not so expensive.
|
Currently Active Users Viewing This Thread: 6 (0 members and 6 guests) | |
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
A1200 RF module removal pics + A1200 chips overview | eXeler0 | Hardware pics | 2 | 08 March 2017 00:09 |
Sale - 2 auctions: A1200 mobo + flickerfixer & A1200 tower case w/ kit | blakespot | MarketPlace | 0 | 27 August 2015 18:50 |
For Sale - A1200/A1000/IndiAGA MkII/A1200 Trapdoor Ram & Other Goodies! | fitzsteve | MarketPlace | 1 | 11 December 2012 10:32 |
Trading A1200 030 acc and A1200 indivision for Amiga stuff | 8bitbubsy | MarketPlace | 17 | 14 December 2009 21:50 |
Trade Mac g3 300/400 or A1200 for an A1200 accellerator | BiL0 | MarketPlace | 0 | 07 June 2006 17:41 |
|
|