English Amiga Board


Go Back   English Amiga Board > Support > support.Other

 
 
Thread Tools
Old 04 July 2021, 06:39   #1
RDP
Registered User
 
Join Date: Jun 2019
Location: Tokyo
Posts: 66
MuRedox

I'd like to try out MuRedox. I've installed:

mmulibs
MuRedox

No matter what I do I get the fpsp.resource not found error. I have a upgraded Warp Engine with a LC060, in my libs I have no 040 library in libs, only 060 loaded from libs. Some libs are:

(OS 3.1.4 from Kickstart ROMS)

68060 46.5 THOR
mmu 46.22 THOR


I'm stumped!


thanks,
RDP
RDP is offline  
Old 04 July 2021, 10:27   #2
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 3,214
That's correct and expected behavior, and this cannot work in your setup. The fpsp.resource is the collection of FPU support functions that come with the 68040 or 68060.library, and that implement the transcendental functions (such as FSIN or FCOS) on top of the common algebraic (FADD/FSUB) instructions. The resource hence requires a FPU to operate. Since the FPU is lacking, there are no FPU functions to build on, therefore there is no fpsp.resource.

Note that MuRedox is *NOT* a FPU emulator. It rather hot-patches non-implemented FPU instructions to routines that implement them on top of the existing FPU code.

If you want to do mathematics on the 68K, your best option are the mathieee libraries. They are always present, work regardless of whether the FPU is present or not, and if the FPU is present, they make use of it. They also make use of the fpsp.resource if this is present, not requiring the Redox patch code.

IOWs, the math functions that implement mathematics without a FPU are already in the system, and part of the math libraries. Just use them, or run programs that use them.
Thomas Richter is offline  
Old 04 July 2021, 15:24   #3
RDP
Registered User
 
Join Date: Jun 2019
Location: Tokyo
Posts: 66
A response from the boss himself! Thank you.

I've read your message a few times now, and I believe I understand what you are saying is:

1. MuRedox is only applicable for FPU users. This is very clear.

2. My base OS install of 3.1.4 has all that I need, don't touch anything. This I'm 90% positive is what your saying. If I'm right, then thanks for the instructions, I'll not touch things that don't require tweaking.

Note: The purpose of my original request was not to try and emulate an FPU, but just to try and ensure my system is going as fast as possible. For instance changing icon.library to the Peter K. version has made a remarkable improvement.


many thanks!
RDP

Last edited by RDP; 04 July 2021 at 15:44.
RDP is offline  
Old 04 July 2021, 20:08   #4
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 3,214
Quote:
Originally Posted by RDP View Post
1. MuRedox is only applicable for FPU users. This is very clear.
That is correct. To be even more precise, it is only for users of the 68040 and 68060 integrated FPU, as the 68882 FPU implements all instructions.


Quote:
Originally Posted by RDP View Post
2. My base OS install of 3.1.4 has all that I need, don't touch anything. This I'm 90% positive is what your saying. If I'm right, then thanks for the instructions, I'll not touch things that don't require tweaking.
That's also correct. The 3.1.4 plus the math.libraries installed there are your best option. They will provide mathematics, though implement them on top of the CPU.


This means that they are necessarily slower, though they will work.


PS: There are multiple "improved" math libs on Aminet - don't thrust them. They will also require a FPU and thus will just crash you on the LC CPU. They lack the genercity of the Os math libs which fall back to the CPU code if an FPU is absent, so there is nothing to be gained there for a FPU-less system.


Don't worry. There is not really much that needs the FPU on the Amiga anyhow. The LC CPU is fine.



PPS: The math libs of 3.2 are a bit better (precision improved), but they are not faster (speed decreased). The real advantage is that they round "IEEE aware" (i.e. follow the IEEE 754 standard that defines how floating point math works).
Thomas Richter is offline  
 


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 11:39.

Top

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Page generated in 0.10987 seconds with 12 queries