exxos's Atari, Amiga & retro forum

Since the Netusbee with uIP tool is really finnicky and doesn't work with the TF534 I decided to give PARCP-USB a try instead. Happy to say it works perfectly! :)

https://joy.sophics.cz/parcp/parcpusb.html

Hi,
may i ask You to backport Your current firmware to TF520?
Wanna try to run it on ST....
Thanks ;)

agranlund wrote: 03 Nov 2019 16:29 I build EmuTOS with -m68030 -DSTATIC_ALT_RAM_ADDRESS=0x01000000

The biggest performance boost comes from letting it know about fastram at compile time, this allows it to put a large chunk of its internal variables there instead of in ST-RAM (the Vampire V2 + V4 optimized builds do the same thing)

Is this something that hasn't been done on the TT030, and is it something it could benefit from?

may i ask You to backport Your current firmware to TF520?

I was under the impression that card already works fine on the Atari?
Since it doesn't have Fastram nor IDE I don't think there is really anything to backport to it - those were mainly the changes I did for the TF534 on Atari (along with some fix to make it compatible with Blitter).

The TF530 however, should probably work the same as TF534 on Atari, but I don't have a 530 so I cannot say for sure.
You can get built firmware binaries for both the 534r2c and 530r3 in the /bin folder:
https://github.com/agranlund/tf534

I want buy a TF520 but someone told me it not work stable under Atari TOS 2.06. The Speeder must be run synchronous to the Atari ST system/busclock to work stable.

I have uploaded Maprom v1.4 to github: https://github.com/agranlund/tftools

- Added support for roms at 0xFC0000
- Mapping system vectors and supervisor stack to fastram
- Mapping some of tos internal variables to fastram (TOS 2.06 only)

This change makes TOS 2.06 quite a bit faster compared to previous version.
It also also adds support for KAOS TOS 1.04.

TOS 2.06 with NVDI and Maprom 1.4: TOS 2.06 with Maprom 1.4: KAOS TOS 1.04 with Maprom v1.4:
EmuTOS runs a tiny bit faster than with the previous version but not really enough to warrant making new gembench screenshots.

There we go!
Upping the clock to 50Mhz finally tipped it over the 1000% average compared to how the machine started life :D
There really aren't that many options for accelerating Ataris out there so I am super grateful that this card exists and is open source.


Nembench v2.1:
percentages are against Falcon030

Integer multiply (16bit) -> 1.921 Mips (~313%)
Integer divide (16bit) -> 1.135 Mips (~313%)
Linear (stalled) integer -> 24.975 Mips (~313%)
Interleaved (piped) integer -> 24.975 Mips (~313%)

16bit read (100% hit) -> 24.630 MByte/sec (~313%)
16bit write (100% hit) -> 14.492 MByte/sec (~240%)
32bit read (100% hit) -> 49.261 MByte/sec (~313%)
32bit write (100% hit) -> 28.901 MByte/sec (~433%)

Linear 32bit read (ST-Ram) -> 3.919 MByte/sec (~73%)
Linear 32bit write (ST-Ram) -> 3.921 MByte/sec (~60%)
Linear 32bit copy (ST-Ram) -> 1.964 MByte/sec (~60%)

Linear 32bit read (FastRAM) -> 30.131 MByte/sec (~566%)
Linear 32bit write (FastRAM) -> 27.976 MByte/sec (~433%)
Linear 32bit copy (FastRAM) -> 13.988 MByte/sec (~433%)

:bravo: You will kill GB6 if those numbers get much higher! :lol:

impressive

How about this as the next challenge for the Mongrel?
:coffee: :ball: :-D