Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
|
|
|
|
Poster | Thread | matthey
| |
Re: ARPi - AROS on Raspberry Pi Posted on 27-Aug-2018 1:39:57
| | [ #1 ] |
| |
|
Elite Member |
Joined: 14-Mar-2007 Posts: 2019
From: Kansas | | |
|
| Quote:
bison wrote: Quote:
I don't agree that "it's not safe to use a system without memory protection on the internet". Anything coming through the internet is under complete control of the internet software. Anything unsafe which is allowed is the fault of the internet software not memory protection. |
I'm not sure what you mean by "the internet software" -- can you clarify this point?
|
Internet software is any software used on the internet including the network stack. Take a simple example of a program that sets the Amiga clock from an atomic online clock. How is malicious code going to enter, execute or crash the system if the network stack and clock setting program are robust and bug free? Even in a dangerous scenario of Javascript being allowed to run in a browser, it is possible to keep the run time environment isolated if performance is sacrificed. Every address and instruction can be checked before an instruction is executed. As long as the programs are robust and bug free there is no problem. The problem is that some internet programs are large, like browsers, and the larger the program the more difficult it is to avoid bugs. The Amiga did have some large robust programs without memory protection back when there was money and developers. Unfortunately, the quality (and quantity) of Amiga software keeps declining as the number of users and developers decline.
Quote:
Quote:
If we can not use the 4 cores but only one of the ARM Cortex-A53, 1.2GHz How fast or slow will that be? |
Faster than any Commodore-era Amiga, I would think. Would be kind of a waste of resources only using one core, but at least it's a very inexpensive waste of resources, since the board is about $45 US.
|
The Cortex-A53 is an energy efficient 64 bit, 4 core, 8 stage in order RISC CPU that is far from high performance. The Tabor Freescale QorIQ P1022 (with PPC e500v2 cores) is a higher performance 32 bit, 2 core, 7 stage OoO RISC CPU. Both have similar clock speeds. The QorIQ P1022 CPU probably has higher single core integer performance than the Cortex-A53 but the non-standard and uncommon SPE/FPU APU and book E departure from AIM PPC makes it difficult to extract non-integer performance. I expect floating point and SIMD workloads to be faster on the Raspberry Pi. I think the Cortex-A53 would be much easier to program and is better supported. AArch64 is an improved PPC in my eyes (similar 32 bit fixed length encoding with ~32 GP registers but easier to use, better code density, more powerful addressing modes, better 32/64 bit support, etc.).
Of course the Cortex-A53 at 1.2GHz would outperform any C= era Amiga CPU. The 68060 could probably come close in single core integer performance per MHz for programs completely in the caches. It wouldn't take much of a rework and die shrinks to outperform a Cortex-A53. The 68060 outperformed many of the RISC CPUs of its time while using less resources and often less energy.
|
| Status: Offline |
| |
|
|
|
[ home ][ about us ][ privacy ]
[ forums ][ classifieds ]
[ links ][ news archive ]
[ link to us ][ user account ]
|