Caladrius Blitter - Should the TT have a Blitter?

General discussions or ideas about hardware.
Post Reply
User avatar
Icky
Site Admin
Site Admin
Posts: 4450
Joined: Sun Sep 03, 2017 10:57 am
Location: UK

Caladrius Blitter - Should the TT have a Blitter?

Post by Icky »

Caladrius Blitter - Should the TT have a Blitter?

While working on the Phoenix BLUE (the FPGA Blitter for the ST with @ijor's core) I started wondering: why doesn't the TT have a Blitter? Could I take the Phoenix BLUE as a base and make a TT version? That idea became the Caladrius Blitter.

Caladrius (another mythical bird).

Why didn't Atari fit one, does anyone know?
My best guess is that a 32 MHz 68030 with caches was already quick at doing blits in software. The ST Blitter is a 16-bit chip with a 24-bit address range, built for the ST bus. Fitting one would have meant a new 32-bit part, more cost and more board space.

What's already in TOS 3.06
Digging through the reconstructed TOS 3.06 sources (thanks to the late Thorsten Otto for those), a surprising amount of Blitter support is still there:
  • The boot-time probe is still there. At start-up the BIOS still does a tst.w $FFFF8A00 behind a bus-error trap, exactly like ST TOS. On a stock TT nothing answers, so TOS decides there's no Blitter.
  • XBIOS 64 Blitmode() is still there. On a stock TT it returns 0. Fit a Blitter and it returns 2, "present but off", which is what SYSINFO reports if there were a TT Blitter.
  • The VDI still has the plumbing. The line-A tables for switching between software and hardware drawing routines are all present. In 3.0x, though, the hardware routines are compiled out: vdidefs.h enables them for TOS 2.x and 4.x only. In their place are ten stubs that do nothing but rts. Switch the Blitter on and those stubs replace the drawing code, and that's when I got two bombs.
  • The desktop still has the menu item. On a 68030 the desktop just relabels it: the Options menu's "Blitter" object is the same one that shows "Cache", and the "Blitter" text is still in the ROM.
  • The routines exist for the 68030. Falcon TOS 4.0x has its Blitter code back, and that code already flushes the 68030 data cache around each blit. So putting it back into 3.06 was mostly a build switch rather than a rewrite.
So TOS 3.06 can see a Blitter but can't use one. Whether that's a deliberate plan or just code shared with ST TOS 2.0x? If you fit one and switch it on, you get two bombs.

So, is it even worth it?
Before spending money on hardware I wanted an answer, so:
  • I added a Caladrius Blitter to Hatari based on the hardware research: a 32-bit, 68030 bus-master version of the ST Blitter that starts hidden, so stock TOS 3.06 stays safe.
  • I patched TOS 3.06 (as a build option on the reconstructed sources) to put the Blitter routines back in.
  • I added a Blitter item to the desktop Options menu next to Cache, as on the ST.
Along the way I found a bug in Atari's own Blitter text code that switches the 68030 caches off whenever the Blitter is on, so that's fixed in the patch too.

SysInfo showing the TT has a Blitter present but turned off:
sysinfo-stock-tos306-caladrius-visible.png
sysinfo-stock-tos306-caladrius-visible.png (314.38 KiB) Viewed 139 times

The test
GEMBench 6.31 on the emulated TT (TOS 3.06, 640x200x4, caches on, run from ST-RAM). The reference is a standard TT with Cache on; the second run has Cache and Blitter on.

gembench-caladrius-tos306-blitter-on.png
gembench-caladrius-tos306-blitter-on.png (316.26 KiB) Viewed 139 times

The results
They're interesting. Everything the screen does gets faster: blitting is 3.6x, text roughly 2x, scrolling 1.8x, and the display average is 170%. The CPU, maths and memory tests stay at 100%, as you'd expect. So even next to a cached 68030, a Blitter makes a real difference to GEM.

A caveat: this is Hatari, using its STE Blitter timing on an emulated TT. It shows the gains are real and how much of GEM benefits, but the real numbers will come from real hardware. That's the next step for the Caladrius Blitter.

Showing the two options side by side in the patched TOS:
Screenshot 2026-11-15 at 17.58.47.png
Screenshot 2026-11-15 at 17.58.47.png (112.65 KiB) Viewed 139 times
User avatar
JezC
Posts: 2909
Joined: Mon Aug 28, 2017 11:44 pm

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by JezC »

Interesting to see where you go with this next - and I'm curious to see where you will connect the physical blitter in the TT (now that I have one I'm weighing up which upgrades are likely to be most useful).
ijor
Posts: 861
Joined: Fri Nov 30, 2018 8:45 pm

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by ijor »

Icky wrote: Sun Nov 15, 2026 6:42 pm The results
They're interesting. Everything the screen does gets faster: blitting is 3.6x, text roughly 2x, scrolling 1.8x, and the display average is 170%. The CPU, maths and memory tests stay at 100%, as you'd expect. So even next to a cached 68030, a Blitter makes a real difference to GEM.

A caveat: this is Hatari, using its STE Blitter timing on an emulated TT. It shows the gains are real and how much of GEM benefits, but the real numbers will come from real hardware. That's the next step for the Caladrius Blitter.
This is so cool!

Note that I don't know anyhing about the TT architecture, but I do suspect that the result is completely biased by how Hatari (or any other emulator, for that matter) Blitter emulation works. When a Blitter operation is started, Hatari performs the whole Blitting operation directly without any emulation. In the worst case it might emulate the bus mastering sequence (not sure, but doubt this is really emulated in a TT). But the actual Blitting is performed natively by the host processor, only the memory access is (sort of) emulated.

OTOH, when TOS performs the blitting, Hatari needs to emulate every single 680X0 instruction. I would expect it to be much slower.

Either way I find this a very clever and cool idea.
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com
User avatar
HigashiJun
Posts: 2508
Joined: Fri Jun 19, 2020 7:21 am
Location: Tokyo

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by HigashiJun »

Very interesting !
User avatar
Icky
Site Admin
Site Admin
Posts: 4450
Joined: Sun Sep 03, 2017 10:57 am
Location: UK

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by Icky »

@JezC: it goes in the TT-RAM slot, with a pass-through connector so an existing TT-RAM card stacks on top. The slot doesn't carry the bus-arbitration or async termination lines, so it also needs a small tap: five wires to one resistor pack (RP101) and one to the bus-grant output of the SCSI DMA chip. Taking the grant there puts the Blitter last in the DMA chain, where the Blitter sits on an ST too.

@ijor: I agree the numbers need real hardware. I think in Hatari's blitter.c: GEMBench times in emulated time, and Hatari charges each Blitter bus access as emulated cycles, so running the blit natively on the host doesn't in itself inflate the result. The bigger bias is that it applies STE Blitter timing on an emulated TT. The cycle-exact Blitter path is only used with a 68000, so on a TT those cycles probably count at 32 MHz, while real ST-RAM on a TT comes in 250 ns slots shared with video. My guess is the 3.6x is optimistic, and the 030 side isn't cycle-exact either. So it shows which parts of GEM would benefit, but not by how much.
ijor
Posts: 861
Joined: Fri Nov 30, 2018 8:45 pm

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by ijor »

Icky wrote: Mon Nov 16, 2026 12:38 pm @JezC: it goes in the TT-RAM slot, with a pass-through connector so an existing TT-RAM card stacks on top. The slot doesn't carry the bus-arbitration or async termination lines, so it also needs a small tap: five wires to one resistor pack (RP101) and one to the bus-grant output of the SCSI DMA chip. Taking the grant there puts the Blitter last in the DMA chain, where the Blitter sits on an ST too.
Blitter is actually the first device in the DMA chain, not the last one. It is connected directed to the CPU. GLUE, when Blitter is installed, is the last device in the DMA chain.

This doesn't mean it won't work like that, as last in the DMA chain, although probably this was never tested.
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com
User avatar
Darklord
Site sponsor
Site sponsor
Posts: 1707
Joined: Wed Sep 20, 2017 1:41 pm
Location: Prestonsburg
Contact:

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by Darklord »

This makes me curious about something. I have a Pak 68/3 accelerator board
in my STacy - 68030 running at 40mhz.

It has a modified version of TOS v3.06, courtesy of Holger Zimmerman.

It has the "Cache" drop-down menu item available.

So my question is, does turning the cache off here actually disable the
Pak board's L1 and L2 caches?

The topic just made me curious. :)
Welcome To DarkForce! www.darkforce.org "The Fuji Lives.!"
Atari SW/HW based BBS-Telnet:darkforce-bbs.dyndns.org 1040
User avatar
Icky
Site Admin
Site Admin
Posts: 4450
Joined: Sun Sep 03, 2017 10:57 am
Location: UK

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by Icky »

ijor wrote: Mon Nov 16, 2026 4:30 pm Blitter is actually the first device in the DMA chain, not the last one. It is connected directed to the CPU. GLUE, when Blitter is installed, is the last device in the DMA chain.

This doesn't mean it won't work like that, as last in the DMA chain, although probably this was never tested.
We'll have to do some tests to see where it should be on chain in the TT.
User avatar
Cyprian
Posts: 572
Joined: Fri Dec 22, 2017 9:16 am
Location: Warszawa, Poland
Contact:

Re: Caladrius Blitter - Should the TT have a Blitter?

Post by Cyprian »

The blitter for the TT sounds great. Will it be 16 or 32bit, and how do you plan to connect it to the mobo?



Regarding Hatari, I'm not sure if it emulates the TT ST-RAM bus correctly, and I wonder if it's clocked at 8Mhz or the CPU clock (32MHz).
Maybe it would be worth to measuring how fast it can move data in ST-RAM, from what I remember, for 16bit blitter it should be 4MB/s.
Atari: V4sa TT030 Falcon030 Mega ST Mega STe 1040 STe 520 STm 800XL 65XE Jaguar Lynx I/II 7800 Portfolio / VME: ATW800/2 Mach64 Mach32 ET4000 CrazyDots PAMNet / SM146 SC1435 SM124 / Antonia2 ATSpeedC16 TF536 PAK68/3
Megafile30 Atari1050 DDDHDD XCA12 SF314 / AVGCART SUBCART SidecarTridge ACSI2STM FujiNet LynxMultiCard SkunkBoard CosmosEx SatanDisk UltraSatan USBFloppyDriveEmulator SDrive SIO2PC Eiffel
http://260ste.atari.org
Post Reply

Return to “HARDWARE DISCUSSIONS”