It started out as an interest into Bubsy 3D, but turned to this:
The early Playstation developers didn’t have access to hardware!
TL;DR: It seems very similar to Net Yaroze's LibPS
Any extra info or corrections appreciated.
Cheers
Mike.
Does anyone have any info about the early PSX 'OS' SDK?
-
Administrator Verified
- Admin / PSXDEV
- Posts: 2689
- Joined: Dec 31, 2012
- I am a: Shadow
- PlayStation Model: H2000/5502
Quite interesting, but at the time of Bubsy 3D, the DTL-H2000 development kit had a full Psy-Q SDK available which is limited to LibGS, LibGPU and LibGTE. The libraries are good and they are powerful, but the problem was that Sony wanted you to use these libraries without any manipulation or custom assembly routines. This was hard because the GPU and CPU weren't documented at all and Sony didn't allow access to them directly, so it was up to programmers to try and manipulate them to get more out of the console than what was developed by Sony Computer Entertainment. If Sony did release specifications about their GPU and CPU to some companies (such as Insomniac Games and or Naughty Dog) then it clearly explains how they managed to achieve such detailed and well polished games.
As an example, LameGuy64 here on the forums has written his own assembler GPU renderer which is better than Sony's version included with Psy-Q. It's faster, and it can handle a lot more sprites at once whilst maintaining 60 frames per second. This is however, due to the fact that the consoles GPU and CPU is now well documented online and information about it isn't so much a secret anymore as what it was back in 1994.
All-in-all, the included Psy-Q library is quite powerful, but what ruined Bubsy 3D was a sheer lack of quality due to poor time management. Using custom assembly routines is mostly to push more out of the console that what was anticipated, but it isn't required. It's more of an optimisation routine than anything because there are hundreds of games on the PlayStation which use the included Psy-Q library programs that are absolutely fantastic without any custom assembler routines.
As an example, LameGuy64 here on the forums has written his own assembler GPU renderer which is better than Sony's version included with Psy-Q. It's faster, and it can handle a lot more sprites at once whilst maintaining 60 frames per second. This is however, due to the fact that the consoles GPU and CPU is now well documented online and information about it isn't so much a secret anymore as what it was back in 1994.
All-in-all, the included Psy-Q library is quite powerful, but what ruined Bubsy 3D was a sheer lack of quality due to poor time management. Using custom assembly routines is mostly to push more out of the console that what was anticipated, but it isn't required. It's more of an optimisation routine than anything because there are hundreds of games on the PlayStation which use the included Psy-Q library programs that are absolutely fantastic without any custom assembler routines.
Development Console: SCPH-5502 with 8MB RAM, MM3 Modchip, PAL 60 Colour Modification (for NTSC), PSIO Switch Board, DB-9 breakout headers for both RGB and Serial output and an Xplorer with CAETLA 0.34.
PlayStation Development PC: Windows 98 SE, Pentium 3 at 400MHz, 128MB SDRAM, DTL-H2000, DTL-H2010, DTL-H201A, DTL-S2020 (with 4GB SCSI-2 HDD), 21" Sony G420, CD-R burner, 3.25" and 5.25" Floppy Diskette Drives, ZIP 100 Diskette Drive and an IBM Model M keyboard.
PlayStation Development PC: Windows 98 SE, Pentium 3 at 400MHz, 128MB SDRAM, DTL-H2000, DTL-H2010, DTL-H201A, DTL-S2020 (with 4GB SCSI-2 HDD), 21" Sony G420, CD-R burner, 3.25" and 5.25" Floppy Diskette Drives, ZIP 100 Diskette Drive and an IBM Model M keyboard.
-
gwald Verified
- Net Yaroze Enthusiast
- Posts: 313
- Joined: Sep 18, 2013
- I am a: programmer/DBA
- PlayStation Model: Net Yaroze
- Contact:
Ha thanks, cool!
Bubsy 3D would have been developed from 1994 through to 1996, using something called "PS-X OS".
Look it up, it reads like Net Yaroze (libgs with CD stuff, etc)!
In Net Yaroze we did have registers info and assembly access to the CPU (R3000.h) but nothing on the GPU nor the GTE, which is what I'm guessing is what Bill refers to as "macros" in the video.
We didn't even get details on the packets in the Order Table (OT), I'm sure it would have been trivial to reverse engineer it though and make your own draw calls, but would probably fail at Sony's Cert Check.
Access to CPU makes sense, as SEGA used their previous CPU's as co processors, ie 68k in the megadrive became the Saturn's audio chip and the same with the master system's SN76489 chip used as audio in the mega drive.
And Sony did just that, the original PS2, PS3 had their predecessor chips on board also.
I'm not sure when Psy-Q came out, but I'd guess to say it was after Bill, 1995?
And nobody would scrap a year of research and coding on a game.
I wouldn't say poor time management, I'd say they didn't have the talent (and Sony contacts) the larger developers Insomniac, Naughty Dog, Eidos, etc had and also being limited to TMD's (libgs).. well you can see why Bubsy 3D and early PS1 games are the way they are
Thanks again.
Bubsy 3D would have been developed from 1994 through to 1996, using something called "PS-X OS".
Look it up, it reads like Net Yaroze (libgs with CD stuff, etc)!
In Net Yaroze we did have registers info and assembly access to the CPU (R3000.h) but nothing on the GPU nor the GTE, which is what I'm guessing is what Bill refers to as "macros" in the video.
We didn't even get details on the packets in the Order Table (OT), I'm sure it would have been trivial to reverse engineer it though and make your own draw calls, but would probably fail at Sony's Cert Check.
Access to CPU makes sense, as SEGA used their previous CPU's as co processors, ie 68k in the megadrive became the Saturn's audio chip and the same with the master system's SN76489 chip used as audio in the mega drive.
And Sony did just that, the original PS2, PS3 had their predecessor chips on board also.
I'm not sure when Psy-Q came out, but I'd guess to say it was after Bill, 1995?
And nobody would scrap a year of research and coding on a game.
I wouldn't say poor time management, I'd say they didn't have the talent (and Sony contacts) the larger developers Insomniac, Naughty Dog, Eidos, etc had and also being limited to TMD's (libgs).. well you can see why Bubsy 3D and early PS1 games are the way they are
Thanks again.
-
Administrator Verified
- Admin / PSXDEV
- Posts: 2689
- Joined: Dec 31, 2012
- I am a: Shadow
- PlayStation Model: H2000/5502
That's strange. I thought Psy-Q came out in 1994 because it was available for the DTL-H2000 which was a development kit based on the 1992 or 1993 hardware design which pre-dates the launch hardware design (PU-7 and PU-8). I could be very wrong though, but I have never heard of nor seen such an SDK called "PS-X OS". Thanks for the information 'gwald'!
Development Console: SCPH-5502 with 8MB RAM, MM3 Modchip, PAL 60 Colour Modification (for NTSC), PSIO Switch Board, DB-9 breakout headers for both RGB and Serial output and an Xplorer with CAETLA 0.34.
PlayStation Development PC: Windows 98 SE, Pentium 3 at 400MHz, 128MB SDRAM, DTL-H2000, DTL-H2010, DTL-H201A, DTL-S2020 (with 4GB SCSI-2 HDD), 21" Sony G420, CD-R burner, 3.25" and 5.25" Floppy Diskette Drives, ZIP 100 Diskette Drive and an IBM Model M keyboard.
PlayStation Development PC: Windows 98 SE, Pentium 3 at 400MHz, 128MB SDRAM, DTL-H2000, DTL-H2010, DTL-H201A, DTL-S2020 (with 4GB SCSI-2 HDD), 21" Sony G420, CD-R burner, 3.25" and 5.25" Floppy Diskette Drives, ZIP 100 Diskette Drive and an IBM Model M keyboard.
Unfortunately we don't have any rips of extremely old PSYQ versions, just the libs with only Japanese documentation (something like 3.4 iirc). Having access to older sdk disks would definitively help in figuring out some dates there.
-
gwald Verified
- Net Yaroze Enthusiast
- Posts: 313
- Joined: Sep 18, 2013
- I am a: programmer/DBA
- PlayStation Model: Net Yaroze
- Contact:
Shadow wrote: ↑June 19th, 2018, 9:59 pm That's strange. I thought Psy-Q came out in 1994 because it was available for the DTL-H2000 which was a development kit based on the 1992 or 1993 hardware design which pre-dates the launch hardware design (PU-7 and PU-8). I could be very wrong though, but I have never heard of nor seen such an SDK called "PS-X OS". Thanks for the information 'gwald'!
https://en.wikipedia.org/wiki/SN_Systems
Off Topic
SN Systems' long association with the PlayStation line of consoles began in 1993, when PlayStation manufacturer Sony acquired Psygnosis, who were publishing SN Systems' tools at the time. While Sony had provided MIPS R4000-based Sony NEWS workstations for PlayStation development, Psygnosis disliked the thought of developing on these expensive workstations and asked SN Systems to create a PC-based development system.[1]
At the 1994 Winter Consumer Electronics Show, Psygnosis arranged an audience for SN Systems with Sony's Japanese executives, and Sony were impressed enough with their development tools that they decided to abandon their plans for a workstation-based development system in favor of using SN Systems' development system exclusively.[1]
At the 1994 Winter Consumer Electronics Show, Psygnosis arranged an audience for SN Systems with Sony's Japanese executives, and Sony were impressed enough with their development tools that they decided to abandon their plans for a workstation-based development system in favor of using SN Systems' development system exclusively.[1]
Date: 1994-01-06 – 1994-01-09 early 1994
Also, SN Systems would have made whatever Sony told them to make, ie Libgs
http://www.raphnet.net/electronique/psx ... tation.txt
Off Topic
Sony's acquisition of Psygnosis yielded another fruit as well: the development system. SN Systems, co-owned by Andy Beveridge and Martin Day, had been publishing its development software through Psygnosis under the PSY-Q moniker.
Sony originally had been planning on using expensive, Japanese MIPS R4000-based machines that would be connected to the prototype PS-X box. Having become accustomed to developing on the PC, Psygnosis gave Beveridge and Day first crack at creating a PlayStation development system that would work with a standard PC.
The two men worked through Christmas and New Year's, around the clock, eventually completing the GNU-C compiler and the source-level debugger. Psygnosis quickly arranged a meeting for SN and Sony at the Winter CES in Las Vegas, 1994. Fortunately, Sony liked the PSY-Q alternative and decided to work with SN Systems on cendensing the software onto two PC-compatible cards. Thus, an afordable and, more importantly, universally compatible PlayStation development station was born.
Sony originally had been planning on using expensive, Japanese MIPS R4000-based machines that would be connected to the prototype PS-X box. Having become accustomed to developing on the PC, Psygnosis gave Beveridge and Day first crack at creating a PlayStation development system that would work with a standard PC.
The two men worked through Christmas and New Year's, around the clock, eventually completing the GNU-C compiler and the source-level debugger. Psygnosis quickly arranged a meeting for SN and Sony at the Winter CES in Las Vegas, 1994. Fortunately, Sony liked the PSY-Q alternative and decided to work with SN Systems on cendensing the software onto two PC-compatible cards. Thus, an afordable and, more importantly, universally compatible PlayStation development station was born.
Who is online
Users browsing this forum: No registered users and 1 guest