Quantcast

Page 2 of 3 FirstFirst 123 LastLast
Results 16 to 30 of 40

Thread: busting 32x performance myths

  1. #16
    Sports Talker
    Join Date
    Sep 2013
    Posts
    40
    Rep Power
    0

    Default

    May I ask for a Saturn binary that tests both VDP1 and VDP2?
    no prob
    http://www.hyakushiki.net/misc/satspd.bin
    I mean, the 68000 is not gonna be any faster so there's nothing they could do about the 68000 hogging the cartridge bus against the SH-2 when it got there first. And if it's running code from cartridge area, it's gonna be doing it all the time.
    Well, I think the 32x could have enforced an access window in between 68k accesses, kind of like how the Amiga shares memory bandwidth. Although that's not what concerns me, I'm just saying they should have made it less of a bottleneck for the SH2, when the 68k is not using it. Even PCE hucards are faster.

    edit: now that I've found this additional piece of documentation https://segaretro.org/images/2/2c/SH...are_Manual.pdf I'm wondering if fooling with the bus state controller could tweak the performance a bit (although the 32x doc said not to do this)
    Last edited by bakemono; 05-30-2019 at 02:50 AM.

  2. #17
    Raging in the Streets Sik's Avatar
    Join Date
    Jan 2011
    Posts
    4,031
    Rep Power
    72

    Default

    Quote Originally Posted by bakemono View Post
    Well, I think the 32x could have enforced an access window in between 68k accesses, kind of like how the Amiga shares memory bandwidth. Although that's not what concerns me, I'm just saying they should have made it less of a bottleneck for the SH2, when the 68k is not using it. Even PCE hucards are faster.
    That's… precisely what it's doing. If a SH-2 wants to access the cartridge while the 68000 tries to access it, then the SH-2 will be given immediate access as soon as the 68000 is done, and if the 68000 wants to access again before SH-2 is done, then the 68000 has to wait. The docs explicitly say that whoever gets in first wins.



    The 68000 is really that slow from the SH-2's viewpoint (don't forget the latter spends less cycles per instruction and also runs at over 3x the clock speed, as well as having a cache to cope with most common accesses). And the PCE used a CPU that spent less cycles on a bus access than the 68000, if you wonder why the hucards are faster. The 68000 is a well known cycle waster, as it was designed in a time where CPU speed was starting to surpass RAM speed significantly and it was better to spend cycles that way to get more complex instructions.

    For the record, the Jaguar suffers from a similar problem: the two DSPs sharing a bus with the 68000 was an awful idea because the latter loves to hog the bus.

    Quote Originally Posted by bakemono View Post
    edit: now that I've found this additional piece of documentation https://segaretro.org/images/2/2c/SH...are_Manual.pdf I'm wondering if fooling with the bus state controller could tweak the performance a bit (although the 32x doc said not to do this)
    I'm pretty sure that's used for the RAM on the 32X and you really really don't want to touch it (・ω・;)

  3. #18
    Sports Talker
    Join Date
    Sep 2013
    Posts
    40
    Rep Power
    0

    Default

    Quote Originally Posted by Sik View Post
    That's… precisely what it's doing.
    Not quite. On the Amiga, the chipset can access memory more often than the 68k does by itself. Every two cycles. It can leave every other slot open for the CPU, and then the CPU only waits when it executes instructions that don't finish in an even multiple of four cycles, while the chipset still gets its business done. On AGA the chipset does two consecutive accesses during its turn.
    The 68000 is really that slow
    That's not relevant to my point though. Even if the 68000 used 5 minutes to read a word, that's no reason for the SH2 to also behave in such a way

  4. #19
    Raging in the Streets Sik's Avatar
    Join Date
    Jan 2011
    Posts
    4,031
    Rep Power
    72

    Default

    The Amiga chipset also has predefined access patterns that have been factored into the wait states they added to the 68000. That doesn't apply to the SH-2s which spend most of their time away from the ROM space and where ROM accesses are completely arbitrary depending on the program running on them.

    The truth is that they probably expected programmers to try avoiding having both sides accessing ROM at the same time and the approach they took was just a fallback for the cases where it's really needed (or for when ROM accesses aren't common enough to become a big issue).

    Also I should mention that going by the tech bulletins, they seemed to be still making some modifications to the hardware three months ahead of release…

  5. #20
    Mastering your Systems Shining Hero TmEE's Avatar
    Join Date
    Oct 2007
    Location
    Estonia, Rapla City
    Age
    30
    Posts
    10,095
    Rep Power
    110

    Default

    Sega didn't opt to add an access buffer in the bus management chip on 32X side which would do fastest possible ROM cycles right away and store it for 68K to work on so that SH2 can get on the bus immediately after rather than wait entire slow 68K bus cycle to get its turn. But we got full CPU access durations on the ROM.
    Death To MP3, :3
    Mida sa loed ? Nagunii aru ei saa "Gnirts test is a shit" New and growing website of total jawusumness !
    If any of my images in my posts no longer work you can find them in "FileDen Dump" on my site ^

  6. #21
    Wildside Expert vexatious's Avatar
    Join Date
    Oct 2014
    Posts
    139
    Rep Power
    8

    Default

    Good job on your testing program.

    I'm still in progress of assembling 74ls612 extension RAM cartridge. Just want to post image to show I haven't forgotten about it but still waiting for parts in mail. Still have to possibly order another controller to interface the RAM with the 74ls612 and core 68000. So far $8.20 after shipping for two 72 POS SKT simm sockets, and $8.75 after shipping for new 60ns 64 megabytes 16X32 non-parity EDO 5v RAM.



    edit
    Received blank pcb from France yesterday 10-16-2019. About $5.00 after shipping.
    BACK

    FRONT

    The two large holes make it useful for mounting other hardware (like cut-outs from breadboard and/or simm sockets).

    edit 10-26-2019
    Received 74ls612 about four days ago.
    About $4.50 after shipping from Asia (shipping lasted about 20 days).


    edit 10-28-2019
    10 pieces of DIP 40 IC socket arrived. Around $4.50 after shipping from china.


    This is everything ordered. Just need some capacitors and maybe chunk of bread board and I can start assembling.
    Last edited by vexatious; 10-28-2019 at 08:57 PM. Reason: More information. New information. Update.

  7. #22
    Wildside Expert vexatious's Avatar
    Join Date
    Oct 2014
    Posts
    139
    Rep Power
    8

    Default

    ComradeOj said, "I have some pages with good info saved".

    @ComradeOj

    Do you have a backup mirror of the old Sega FTP from about 1994-1995? There was some older GNU stuff and tools; some of it might have been used for early E3 demos for Metal Head and other titles before being catalogued. There were some older non version number programs for Macintosh developers too (think 3Dedit, sound editor and others).

    edit:
    please see my post-thread here http://www.sega-16.com/forum/showthr...rom-about-1994 Thanks.
    Last edited by vexatious; 12-09-2019 at 05:58 AM.

  8. #23
    Comrade as in friend. Master of Shinobi ComradeOj's Avatar
    Join Date
    Dec 2012
    Location
    New Mexico, USA
    Age
    25
    Posts
    1,418
    Rep Power
    61

    Default

    Quote Originally Posted by vexatious View Post
    ComradeOj said, "I have some pages with good info saved".

    @ComradeOj

    Do you have a backup mirror of the old Sega FTP from about 1994-1995? There was some older GNU stuff and tools; some of it might have been used for early E3 demos for Metal Head and other titles before being catalogued. There were some older non version number programs for Macintosh developers too (think 3Dedit, sound editor and others).

    edit:
    please see my post-thread here http://www.sega-16.com/forum/showthr...rom-about-1994 Thanks.
    No, sorry. I just have some 32X programming manual scans. They seem have screenshots and references to programming tools, but it's only just manuals. The sound stuff in particular has screenshots of what I think are old Macintosh programs.

    I have a mirror of it on my website here, if you want to look: https://mode5.net/32x-DDK/32Xtoc.html
    Modded consoles:
    Master System (v7040) with s-video & direct AV out
    Model 1 with 10mhz overclock & halt switches
    Model 1 with 10mhz 68010
    Model 2 VA2.3 with unfiltered Mega Amp, & s-video
    Model 3 VA1 with compatibility fixes & s-video
    32X with s-video
    Visit my web site at www.mode5.net
    Or my collection of homebrew Genesis games, programs, and music on SEGA-16!

  9. #24
    Master of Shinobi
    Join Date
    Sep 2012
    Posts
    1,314
    Rep Power
    38

    Default

    I finally got around today to build a bootable iso of this and run it on the Saturn. Results:

    on a JPN VA0 (1.00) machine:

    37616 KB/sec cached read
    22752 KB/sec SDRAM read
    6080 KB/sec loRAM read
    10928 KB/sec sparse SDRAM read
    1728 KB/sec sparse loRAM read
    10976 KB/sec sparse SDRAM uncached
    6208 KB/sec sparse loRAM uncached

    24704 KB/sec SDRAM write
    13872 KB/sec VDP2 write
    12816 KB/sec SDRAM-VDP2 copy
    5088 KB/sec loRAM-VDP2 copy

    2416 KB/sec VDP2 read
    2176 KB/sec VDP1 read
    8448 KB/sec VDP1 write


    on a PAL VA7 machine:

    45168 KB/sec cached read
    27472 KB/sec SDRAM read
    7312 KB/sec loRAM read
    13184 KB/sec sparse SDRAM read
    2080 KB/sec sparse loRAM read
    13264 KB/sec sparse SDRAM uncached
    7488 KB/sec sparse loRAM uncached

    29840 KB/sec SDRAM write
    17296 KB/sec VDP2 write
    15568 KB/sec SDRAM-VDP2 copy
    6128 KB/sec loRAM-VDP2 copy

    2928 KB/sec VDP2 read
    2624 KB/sec VDP1 read
    10176 KB/sec VDP1 write

    The PAL machine runs at 50Hz vs the JPN machine at 59,82Hz (as repoted by OSSC), this corresponds to the roughly 20% speed boost on all figures: 20% less video frames to draw roughly equals 20% more time available for doing CPU work.

  10. #25
    WCPO Agent chinitosoccer's Avatar
    Join Date
    Dec 2008
    Location
    Uruguay
    Posts
    926
    Rep Power
    26

    Default

    So what does all of this gibblerish means? thats impossible to have a 32x port of Mortal Kombat and WWF arcade running at a fremerate higher than 24-30fps?
    Last edited by chinitosoccer; 03-04-2020 at 08:45 AM.

  11. #26
    Hedgehog-in-Training Hedgehog-in-TrainingSports Talker waterclaws6's Avatar
    Join Date
    May 2018
    Location
    Virginia
    Posts
    43
    Rep Power
    0

    Default

    It possible to have a great 32X port MK and WWE the arcade game at 60fps, however the 32X is not fast enough to software render multiple big sprites at high framerates without help from the genesis and a lot of time getting good code from the system. The 32X is basically giving the genesis a 486 overdrive cpu with crippling bottlenecks. At least MK2 and NBA Jam were good ports on the 32X.
    TRUST BUT VERIFY IS ALWAYS THE RIGHT THING TO DO.

  12. #27
    WCPO Agent chinitosoccer's Avatar
    Join Date
    Dec 2008
    Location
    Uruguay
    Posts
    926
    Rep Power
    26

    Default

    Both MK2 and WWF arcade are crap on the 32X when they run at 25-30fps while they do it aT 60fps on the MD and SNES versions.

  13. #28
    Master of Shinobi
    Join Date
    Sep 2012
    Posts
    1,314
    Rep Power
    38

    Default

    Steve once said that Mortal Kombat 2 is already pushing the limits of what the 32x can do. I'm assuming fillrate-wise.

  14. #29
    Hero of Algol
    Join Date
    Aug 2010
    Posts
    7,861
    Rep Power
    180

    Default

    Quote Originally Posted by chinitosoccer View Post
    So what does all of this gibblerish means? thats impossible to have a 32x port of Mortal Kombat and WWF arcade running at a fremerate higher than 24-30fps?
    Quote Originally Posted by chinitosoccer View Post
    Both MK2 and WWF arcade are crap on the 32X when they run at 25-30fps while they do it aT 60fps on the MD and SNES versions.
    Quote Originally Posted by zyrobs View Post
    Steve once said that Mortal Kombat 2 is already pushing the limits of what the 32x can do. I'm assuming fillrate-wise.
    SNES' MKII actual animation framerate is way lower and far more fragmented (torso animates when lower body is still and so on) than both the 32X and MD ports.
    The 32X port isn't locked at 30 fps either, it runs well for the most part but has noticeable slowdowns in spots, like in the Acid Pool stage.

    32X's WWF Wrestlemania: The Arcade Game is a very different port: while it is locked at 30 fps, it runs at higher resolution, has bigger and better animated characters, animated crowd at all times (even in the 4 wrestlers matches), far less slowdown, transparent shadows and clearer voice samples playback when compared to its MD counterpart. It's a major overhaul in all technical areas, unlike the MKII port.
    The SNES port of WWF doesn't even have a mode with 4 wrestlers on screen and with 3 it's already slowing down to a crawl and having to completely cut the in-game narration to offload the CPU; to even compare the 32X version to it is madness.


    Both ports had room for improvement, but especially the MKII one which is clearly a half-assed job in many aspects.
    One of the things they overlooked was the voice samples: the 32X version is still using the same MD ones and then adding some others (with non-normalized volume, so you can notice a difference between the old and "new" ones, it's a mess).
    So the Z80 keeps accessing the ROM very often to stream those big uncompressed PCM samples (15 kHz for MKII!!!!) and it's a cardinal sin in terms of performance to do it with 32X games.
    They certainly didn't "maxed out" the 32X with such a sloppy version but they surely made it struggle.
    The parallax implementation in some stages, such as the Acid Pool is also far from optimized.
    If after those optimizations fillrate was still an issue, they could still have fragmented the animation of the characters a bit more like Sculptured Software did for the SNES.
    There were many known tricks they could have used to improve it.
    Last edited by Barone; 03-06-2020 at 08:16 PM.

  15. #30
    End of line.. Hero of Algol gamevet's Avatar
    Join Date
    Jan 2008
    Location
    Dallas, Texas
    Posts
    8,945
    Rep Power
    127

    Default

    WWF on the SNES is on a much smaller cart. It doesn’t even have 2 of the characters, because of the cut down cart. It’s a horrible port.

    Have you checked out the latest Digital Foundry Retro, Barone? They’re doing a 2 part segment where they play every 32X game. They also compare stuff like MKII across all 3 platforms. Oops! I just realized that video is 11 months old.😉
    Last edited by gamevet; 03-09-2020 at 07:36 PM.
    A Black Falcon: no, computer games and video games are NOT the same thing. Video games are on consoles, computer games are on PC. The two kinds of games are different, and have significantly different design styles, distribution methods, and game genre selections. Computer gaming and console (video) gaming are NOT the same thing."



Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •