Click Here
home features news forums classifieds faqs links search
6018 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
» Home
» Features
» News
» Forums
» Classifieds
» Links
» Downloads
Extras
» OS4 Zone
» IRC Network
» AmigaWorld Radio
» Newsfeed
» Top Members
» Amiga Dealers
Information
» About Us
» FAQs
» Advertise
» Polls
» Terms of Service
» Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
85 crawler(s) on-line.
 15 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 Birbo:  7 mins ago
 Frank:  16 mins ago
 g0blin:  33 mins ago
 Fl@sh:  54 mins ago
 paolone:  1 hr 1 min ago
 AP:  1 hr 5 mins ago
 Trixie:  1 hr 18 mins ago
 gonegahgah:  1 hr 29 mins ago
 AmigaOldskooler:  1 hr 43 mins ago
 Rob:  1 hr 46 mins ago

/  Forum Index
   /  Amiga General Chat
      /  Arne and Gunnars ASM course Twitch video
Register To Post

Goto page ( 1 | 2 | 3 | 4 Next Page )
PosterThread
Overflow 
Arne and Gunnars ASM course Twitch video
Posted on 1-Aug-2021 21:40:51
#1 ]
Super Member
Joined: 12-Jun-2012
Posts: 1628
From: Norway

Arne and Gunnar (Apollo/Vampire) is going to do a step by step ASM course/school.

The first broadcast was yesterday, and the next will be in four weeks I believe.

https://www.twitch.tv/videos/1104004078

Ofcourse, waiting for 4 weeks for the next video might feel a long wait, so Photons course is a nice distraction while waiting for the next broadcast.

https://www.youtube.com/watch?v=p83QUZ1-P10

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Arne and Gunnars ASM course Twitch video
Posted on 2-Aug-2021 19:16:53
#2 ]
Elite Member
Joined: 9-Jun-2004
Posts: 11871
From: Norway

@Overflow

Let’s summarize, what we have learned.

Wow 1 hour, of PowerPC is bad,
SAGA is fun because limited.
680x0 is fun because its complex, and slow.
(of cause you don't need a good FPU..)
and don’t need Ghz, don’t need RAM.
flat memory is good, because waste address space.
you don’t need MMU.
and C/C++ programs does not exist.

Last edited by NutsAboutAmiga on 02-Aug-2021 at 07:19 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
agami 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 5:46:03
#3 ]
Cult Member
Joined: 30-Jun-2008
Posts: 629
From: Melbourne, Australia

Another thing we learned is that @NutsAboutAmiga is a cynic

 Status: Offline
Profile     Report this post  
BSzili 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 6:58:27
#4 ]
Regular Member
Joined: 16-Nov-2013
Posts: 429
From: Unknown

@agami

Maybe he is, but is his summary inaccurate? I've only listened to the first 30 minutes, but it ticked off a few points from his list.

_________________
This is just like television, only you can see much further.

 Status: Offline
Profile     Report this post  
IridiumFX 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 14:03:55
#5 ]
Member
Joined: 7-Apr-2017
Posts: 57
From: London, UK

well, I went through the whole video.

Disclaimer: I have no vampire, no plans to buy one, I am not even registered to the website.

Gunnar spoke of PPC in terms of number of instructions to achieve the same as the 68K. It's not "PPC is bad" but, in the exact same words as he used, "68K is easier to program". Prove him wrong ?

He keeps comparing or mixing ECS/AGA and SAGA, which is less than ideal, his point is "SAGA is fun because it's like AGA just less limited".

on the FPU he felt the pain, yes, he acknowledges the total mess of not having room in the V2 cards and having partial implementations left behind.

MMU ... was in the specific context of video drivers for Shapeshifter (in case the point was missed). No idea if his description of the Shapeshifter video driver behaviour is accurate, someone more informed may pick up. Yet this is not a "you don't need an MMU" by a long shot.

all of the GHz/RAM/C/C++ points made are really coming from some obsession of the writer. I think the technical definition is FUD more than being cynical.

This is Gunnar making an introductory asm coding video for his own product. The three takehome points are: introductory, asm, own product. I see nothing wrong, tbh.

I suggest watching the video again, in case of doubts, rather than going nuts.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 16:32:33
#6 ]
Elite Member
Joined: 9-Jun-2004
Posts: 11871
From: Norway

@agami

"cynic" by definition.

1. believing that people are motivated purely by self-interest; distrustful of human sincerity or integrity.

2. concerned only with one's own interests and typically disregarding accepted standards in order to achieve them.

I have no self interests here, I did not prepare a 5 pages, to explain way PowerPC / AmigaONE’s are bad, in order to sell my own product. I have no products on sale.

What I’m trying do there is satire:

“the use of humour, irony, exaggeration, or ridicule to expose and criticize people's stupidity or vices, particularly in the context of contemporary politics and other topical issues.”

not everyone get satire, and not everyone understand irony.

Last edited by NutsAboutAmiga on 03-Aug-2021 at 04:41 PM.
Last edited by NutsAboutAmiga on 03-Aug-2021 at 04:33 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Rose 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 16:35:15
#7 ]
Cult Member
Joined: 5-Nov-2009
Posts: 927
From: Unknown

@NutsAboutAmiga

It might be understood as satire/irony if you weren't known for your crusade against classics.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 16:50:23
#8 ]
Elite Member
Joined: 9-Jun-2004
Posts: 11871
From: Norway

@Rose

I’m not on crusade on classic, the only crusade is topic of writing code that bangs the hardware, my believe is that should always have a software API, in between the hardware and software, I believe allows for flexibility and allows users to take advantage of addons, highend sound card, graphic cards, and USB controllers, or other upgrades like scsi / sata / sas or whatever. If take over the OS, and stop multitasking, then you kill all drivers and upgrades.

I'm not the only one saying this, even real Classic users thinks this.

https://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=44227&forum=25&start=20&viewmode=flat&order=0#843395

I written many times that think its good have product a low price point for Amiga users, and believe I also have argued for way chunky is good thing, two things the vampire offers, so don’t tell me I’m on crusade agents the vampire,

Last edited by NutsAboutAmiga on 03-Aug-2021 at 05:26 PM.
Last edited by NutsAboutAmiga on 03-Aug-2021 at 05:25 PM.
Last edited by NutsAboutAmiga on 03-Aug-2021 at 05:07 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
matthey 
Re: Arne and Gunnars ASM course Twitch video
Posted on 3-Aug-2021 22:53:04
#9 ]
Super Member
Joined: 14-Mar-2007
Posts: 1145
From: Kansas

NutsAboutAmiga Quote:

Let’s summarize, what we have learned.

Wow 1 hour, of PowerPC is bad,


I was thinking Gunnar's primary point was that the 68k is easier to program than PPC before IridiumFX posted it. This is very much true. Care to write us a 3D projection demo in PPC assembler? You may say that it would be more difficult using the OS but it isn't much. I wrote a 68k assembler projection demo/test program many years ago that is 100% system friendly and really only cost a few kB of code at most.

Gunnar talks about PPC needing many more (simple) instructions which it often does. The larger code is less readable and wastes caches/memory. RISC processors like PPC and ARM cheapened the hardware to make development of cores faster and moved complexity into the compiler which gave inferior performance. These are common limitations of RISC processor designs with a high performance goal. PPC has a further problem of the ISA not being human friendly as the plan was that optimizing compilers would solve all performance problems. The compiler didn't solve complex problems for RISC or VLIW processors despite huge investments from large companies. ARM processors were easier to use and often simpler and cheaper than PPC processors even though this often gave inferior performance except for low memory bandwidth embedded systems using Thumb2 due to code density similar to the 68k. PPC died quickly when ARM AArch64 improved performance with more complexity for performance, standardization, better code density and an easier to use ISA.

Gunnar talks about limitations of the VMX/Altivec PPC SIMD unit. PPC has alignment limitations and data transfers between units are limited. CISC style LOAD+OP instructions reduce instructions and registers needed as well. While AMMX avoids some performance reductions of Altivec, it it not the dream 68k SIMD unit that Gunnar makes it out to be. Altivec has wider registers for more operations per SIMD instruction and supports floating point. The Vampire assembler course Tie fighter demo uses floating point in the FPU where real time 3D projections and transformations would very much benefit from floating point support in the SIMD unit. It's interesting that FSIN and FCOS FPU instructions are used which are likely not available in the FPU but rather trapped and calculated in software. Using floating point and the missing FSIN and FCOS instructions is nicer and easier to use than fixed point math and tables but this isn't helped by the SIMD unit and even the FPU is limited in the FPGA. In my opinion, this is where Gunnar is less than honest by omitting major weaknesses of the Apollo core in FPGA using an ISA designed for an FPGA.

NutsAboutAmiga Quote:

SAGA is fun because limited.


Once again IridiumFX pointed out that SAGA is less limiting and easier to use than AGA making it more fun. I agree. SAGA and other enhanced Amiga chipsets work well in FPGA, can provide good compatibility, can provide enhancements planned for AA+ and AAA with fewer restrictions and can support modern I/O. Where is the limitation?

NutsAboutAmiga Quote:

680x0 is fun because its complex, and slow.


A 680x0 core in an FPGA is at slower clock speeds because of routing and resource limitations in a FPGA. The Motorola/Freescale 680x0 chips are at slower clock speeds because of the old chip fab process used. Complexity can improve performance which is why modern desktop processors often have more complexity than the 68060 and use many times the number of transistors. The 1994 68060 was ~2,500,000 transistors while a 2019 AMD Ryzen 9 3900X is ~9,890,000,000 transistors which is ~3956 times the number of transistors.

NutsAboutAmiga Quote:

(of cause you don't need a good FPU..)


A cut down 68k FPGA FPU seems to do ok as that is what is used in the demo. Never mind that it is likely trapping FPU instructions which made the code easier to write and is chopped to a 64 bit FPU or less.

NutsAboutAmiga Quote:

and don’t need Ghz, don’t need RAM.


Gunnar says that more memory is less limiting than back in his hacker days. More memory could easily be added to the Vampire boards but offers diminishing advantages especially considering the performance of the FPGA CPU core which lacks GHz.

NutsAboutAmiga Quote:

flat memory is good, because waste address space.


Flat memory is good for performance. Would you rather waste address space or performance on lower performance hardware?

Quote:

We see that the total cost of memory management includes the execution of the virtual memory system (adding roughly 5-10% to total execution time), the cost of taking interrupts to handle the memory-management events (another 5-10%), and the increased number of cache misses seen by the application (another 5-10%). This adds up to a significant cost that becomes more noticeable as caches and reorder buffers increase in size.


http://tnm.engin.umich.edu/wp-content/uploads/sites/353/2017/12/1998.10.A_look_at_several_memory_management_units_ASPLOS-VIII.pdf

NutsAboutAmiga Quote:

you don’t need MMU.


The Amiga demonstrates that an MMU is not needed.

NutsAboutAmiga Quote:

and C/C++ programs does not exist.


Yet 68k fanatics can program in C better than PPC fanatics can program in PPC assembler. The 68k fanatics can usually debug C code better as well.

NutsAboutAmiga Quote:

I’m not on crusade on classic, the only crusade is topic of writing code that bangs the hardware, my believe is that should always have a software API, in between the hardware and software, I believe allows for flexibility and allows users to take advantage of addons, highend sound card, graphic cards, and USB controllers, or other upgrades like scsi / sata / sas or whatever. If take over the OS, and stop multitasking, then you kill all drivers and upgrades.


I agree. There is nothing wrong with using assembler to show how easy it is to program the Vampire at a low level but it is a poor example to encourage programmers to bang the hardware when it is less necessary than when the Amiga was introduced. The Vampire USB support is a good example of the flexibility of USB lost by mapping particular device types on particular ports to particular hardware registers so a device driver is unnecessary. The hardware is less professional when shortcuts like this are taken. Unfortunately, the CPU ISA has similar hackish shortcuts and FPGA optimizations. This is the result of one man making the decisions instead of a knowledgeable team.

 Status: Offline
Profile     Report this post  
megol 
Re: Arne and Gunnars ASM course Twitch video
Posted on 4-Aug-2021 20:25:33
#10 ]
Regular Member
Joined: 17-Mar-2008
Posts: 355
From: Unknown

@matthey
1998 is 23 years ago, that paper is very old.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Arne and Gunnars ASM course Twitch video
Posted on 4-Aug-2021 21:24:57
#11 ]
Elite Member
Joined: 9-Jun-2004
Posts: 11871
From: Norway

@matthey

there is no advantage to writing code, which only works on one CPU, it must then be better to write code that works on several types of CPUs.

PowerPC instructions are acronyms, so if know what etch letter stands for its not hard to read.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
saimo 
Re: Arne and Gunnars ASM course Twitch video
Posted on 4-Aug-2021 21:33:57
#12 ]
Elite Member
Joined: 11-Mar-2003
Posts: 2367
From: Unknown

@NutsAboutAmiga

Quote:
the only crusade is topic of writing code that bangs the hardware, my believe is that should always have a software API, in between the hardware and software, I believe allows for flexibility and allows users to take advantage of addons, highend sound card, graphic cards, and USB controllers, or other upgrades like scsi / sata / sas or whatever. If take over the OS, and stop multitasking, then you kill all drivers and upgrades.

"Crusade" bothers me. You're entitled to have your own view and express/promote it, but it should not be forced onto others. Your points about system-compliant software are valid, but that's only one side of the story: also hardware-banging software has its advantages - namely, achieving what would be unthinkable through the OS, better/maximum exploitation of the hardware resources, the fun of programming the hardware directly, the satisfaction of getting the machine to do incredible things

_________________
RETREAM - retro games for Amiga, Commodore 64 and PC

 Status: Offline
Profile     Report this post  
Fl@sh 
Re: Arne and Gunnars ASM course Twitch video
Posted on 4-Aug-2021 23:37:21
#13 ]
Regular Member
Joined: 6-Oct-2004
Posts: 229
From: Napoli - Italy

When you drive a motorbike for first time it's more difficult to manage vs a simple bicycle.
The same is when you approach to a powerpc for first time, coming from an old 68k.



_________________
Pegasos II G4@1GHz 2GB Radeon 9250 256MB
AmigaOS4.1 fe - MorphOS - Debian 9 Jessie

 Status: Offline
Profile     Report this post  
kolla 
Re: Arne and Gunnars ASM course Twitch video
Posted on 5-Aug-2021 5:06:42
#14 ]
Super Member
Joined: 21-Aug-2003
Posts: 1809
From: Trondheim, Norway

Why don't everyone just write asm for VAX instead.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
kolla 
Re: Arne and Gunnars ASM course Twitch video
Posted on 5-Aug-2021 5:39:12
#15 ]
Super Member
Joined: 21-Aug-2003
Posts: 1809
From: Trondheim, Norway

@matthey

Quote:

The Amiga demonstrates that an MMU is not needed.


"THE AMIGAOS IS NOT INTENDED FOR USE IN THE OPERATION OF NUCLEAR FACILITIES, AIRCRAFT NAVIGATION, COMMUNICATION SYSTEMS, OR AIR TRAFFIC CONTROL MACHINES IN WHICH CASE THE FAILURE OF THE AMIGAOS COULD LEAD TO DEATH, PERSONAL INJURY, OR SEVERE PHYSICAL OR ENVIRONMENTAL DAMAGE."

Meanwhile, a drone is flying around on Mars, running Linux, with an MMU.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
OlafS25 
Re: Arne and Gunnars ASM course Twitch video
Posted on 5-Aug-2021 12:49:10
#16 ]
Elite Member
Joined: 12-May-2010
Posts: 5873
From: Unknown

@NutsAboutAmiga

If it would be recognizeable as humour yes... perhaps you should write something like (/irony) next time to make it obvious

The problem is you are very against 68k and vampire and pro PPC, and that without any humour or irony. I did not see your comments as being funny but making Gunnar look as a stupid and fanatic talking down PPC and selling his product (as you wrote). Not very funny

 Status: Offline
Profile     Report this post  
matthey 
Re: Arne and Gunnars ASM course Twitch video
Posted on 5-Aug-2021 23:45:40
#17 ]
Super Member
Joined: 14-Mar-2007
Posts: 1145
From: Kansas

megol Quote:

1998 is 23 years ago, that paper is very old.


The study is rough too but the performance issues are obvious. The hardware used for virtual memory is similar and the cache sizes studied large enough to see that the performance problem is not going away today.

Modern hardware improvements on high end hardware
L2 TLB and MMU superpages (many big disadvantages but highly effective at reducing TLB misses)
L3 caches (decreased cost for cache misses expensive)

Modern hardware challenges that have grown worse
interrupt and exception costs have increased with aggressive OoO and parallel processing
memory has increased roughly 1000 times (more strain on TLB and DCache especially)

Quote:

Main memory has grown exponentially in size over at least the last decade and, as cause or consequence, the memory requirements of applications have proportionally increased [20]. In contrast, TLB coverage has lagged behind. The TLB is usually fully associative and its access time must be kept low, since it is in the critical path of every memory access [13]. Hence, TLB size has remained relatively small, usually 128 or fewer entries, corresponding to a megabyte or less of TLB coverage. Figure 1 depicts the TLB coverage achieved as a percentage of main memory size, for a number of Sun and SGI workstation models available between 1986 and 2001. Relative TLB coverage is seen to be decreasing by roughly a factor of 100 over ten years. As a consequence, many modern applications have working sets larger than the TLB coverage. Section 6.3 shows that for many real applications, TLB misses degrade performance by as much as 30% to 60%, contrasting to the 4% to 5% reported in the 1980’s [2, 24] or the 5% to 10% reported in the 1990’s [17, 23]. Another trend that has contributed to this performance degradation is that machines are now usually shipped with on-board, physically addressed caches that are larger than the TLB coverage. As a result, many TLB misses require access to the memory banks to find a translation for data that is already in the cache, making misses relatively more expensive.


https://www.usenix.org/legacy/events/osdi02/tech/full_papers/navarro/navarro.pdf

Superpages are complex and waste lots of memory while only dealing with TLB misses. An L3 cache helps with the (mostly data) cache miss problem but each level of cache added gives diminishing returns while being more expensive than the previous level of cache. Then there is the interrupt and exception overhead which is getting worse as flushing the pipeline of highly parallel processing OoO processors gets more expensive. Modern computing has been slow to adapt to this reality. Look no further than IEEE 754 for floating point which still requires exceptions when it makes little sense for an SIMD unit let alone a FPU which would often be better off propagating errors through parallel calculations (propagated NaNs). This is another example where the technology standard has not changed much since the 1990s despite performance issues on modern hardware.

NutsAboutAmiga Quote:

there is no advantage to writing code, which only works on one CPU, it must then be better to write code that works on several types of CPUs.


Everyone could write for one virtual processor/hardware or intermediate code but these exist and usually aren't popular because they are less efficient than real hardware. If your intent is AmigaNowhere then that is already as available as it will never be.

kolla Quote:

Why don't everyone just write asm for VAX instead.


VAX code could have been an intermediate language but it has endian issues and is variable length byte encoded like x86. Sticking to all big or little endianness and using a 16 bit variable length encoding would have been better. Then again, simplify it a little to execute better on real hardware and you have the 68k. Problem solved.

kolla Quote:

Meanwhile, a drone is flying around on Mars, running Linux, with an MMU.


That's not the Mars Pathfinder that crashed with a priority inversion error?

https://www.cs.unc.edu/~anderson/teach/comp790/papers/mars_pathfinder_long_version.html

I guess that was VxWorks which is more reliable and better for real time systems than Linux (still garbage in garbage out). At least VxWorks didn't crash from interrupts like Linux with a MMU did in real time tests. It looks like the CPU is an IBM RAD6000 (POWER ISA) based on a RS/6000 workstation/server core and it looks like it has a MMU.

https://en.wikipedia.org/wiki/IBM_RAD6000
https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.533.3218&rep=rep1&type=pdf

The OS may be VxWorks 5.x which may *not* use the MMU though.

https://ntrs.nasa.gov/api/citations/20080040872/downloads/20080040872.pdf

NASA had no plans to discontinue no MMU real time OSs for its low end systems which included ColdFire hardware in 2007.

Last edited by matthey on 05-Aug-2021 at 11:55 PM.

 Status: Offline
Profile     Report this post  
Jose 
Re: Arne and Gunnars ASM course Twitch video
Posted on 6-Aug-2021 1:13:38
#18 ]
Cult Member
Joined: 10-Mar-2003
Posts: 956
From: Unknown

@kolla
"
""THE AMIGAOS IS NOT INTENDED FOR USE IN THE OPERATION OF NUCLEAR FACILITIES, AIRCRAFT NAVIGATION, COMMUNICATION SYSTEMS, OR AIR TRAFFIC CONTROL MACHINES IN WHICH CASE THE FAILURE OF THE AMIGAOS COULD LEAD TO DEATH, PERSONAL INJURY, OR SEVERE PHYSICAL OR ENVIRONMENTAL DAMAGE."

Meanwhile, a drone is flying around on Mars, running Linux, with an MMU.
"

So what ? That's just legal protection from liability. NASA used Amigas and it included managing rockets...

Last edited by Jose on 06-Aug-2021 at 01:13 AM.

_________________

José

 Status: Offline
Profile     Report this post  
kolla 
Re: Arne and Gunnars ASM course Twitch video
Posted on 6-Aug-2021 20:24:37
#19 ]
Super Member
Joined: 21-Aug-2003
Posts: 1809
From: Trondheim, Norway

@matthey

Quote:

kolla Quote:

Meanwhile, a drone is flying around on Mars, running Linux, with an MMU.


That's not the Mars Pathfinder that crashed with a priority inversion error?


Very much stuck in the 90ies, huh?!

No, I meant NOW, in 2021, and I meant literally FLYING - the Mars Helicopter. The "kids" at NASA have all grown up with Linux, and so it is their operating system of choice when RT is not a requirement.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
matthey 
Re: Arne and Gunnars ASM course Twitch video
Posted on 7-Aug-2021 2:37:38
#20 ]
Super Member
Joined: 14-Mar-2007
Posts: 1145
From: Kansas

kolla Quote:

Very much stuck in the 90ies, huh?!


Well, it takes 6 months to get to Mars and the exploration vehicles can be operational for a long time (15 years for the Opportunity rover). The important computers are specially manufactured for reliable embedded use so look outdated compared to modern high performance computers hear on earth.

kolla Quote:

No, I meant NOW, in 2021, and I meant literally FLYING - the Mars Helicopter. The "kids" at NASA have all grown up with Linux, and so it is their operating system of choice when RT is not a requirement.


You are talking about the experimental demonstration of technology drone helicopter that arrived with the Perseverance called the Ingenuity? VxWorks didn't work on the mobile phone SoC according to one of the software engineers Tim Canham.

Quote:

The helicopter is not part of the main mission of Perseverance, but rather a “demonstration of technology” to prove it is possible to fly on Mars. For this reason, it was built with commercially available components (“off-the-shelf”, in industry jargon) and low cost, instead of solutions specially developed for use in space.

“We wanted a powerful and compact processor for the helicopter, and the best candidate was a Snapdragon 801 card that we found. But there is no version of VxWorks for it, ”said Canham.


https://olhardigital.com.br/en/2021/02/22/ciencia-e-espaco/linux-chega-a-marte-a-bordo-da-ingenuity/

Ingenuity is *not* mission critical so they could use an OS which is less reliable, less responsive and more wasteful of performance and energy like Linux. The ARM based SoC can be clocked up to 2.2GHz while the Perseverance main computer only runs at a max clock of 200MHz using a PPC BAE RAD750 (PPC G3) and VxWorks OS. The RAD750 CPUs (there is a 2nd on Perseverance for redundancy) cost in the neighborhood of 10,000 times more than the Snapdragon but reliability has a price tag.

The bloat of the Ingenuity computer system is out of this world. It uses inefficient Linux underneath, a flight software framework written in C++ on top and even interpreted Python for scripting that uses garbage collection. Just shrink the chip die size and turn up the CPU clock to make a real time system. I'm surprised they didn't use a Raspberry Pi in it as the flight software runs on it. Would it have looked too much like a toy with a RPi?

 Status: Offline
Profile     Report this post  
Goto page ( 1 | 2 | 3 | 4 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle