Poster | Thread |
Rose
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 21:47:13
| | [ #81 ] |
|
|
 |
Cult Member  |
Joined: 5-Nov-2009 Posts: 982
From: Unknown | | |
|
| @wawa
Quote:
n the third post under the above link gunnar explains advantages of his cpu in comparison with different previous 68k implementations. we can keep pavlors caveat in mind, this being publicity excercise, and nevertheless everybody can decide for himself if he trust gunnar, or not.. |
If info is coming from Gunnar you might need to take it with this....

Yes, it's salt.... |
|
Status: Offline |
|
|
tangoone
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 21:59:06
| | [ #82 ] |
|
|
 |
Regular Member  |
Joined: 2-Jul-2014 Posts: 152
From: Norway | | |
|
| @Rose
but since "Gunnar" is the in the apollo team I guess he knows what he is talking about. |
|
Status: Offline |
|
|
wawa
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 22:27:34
| | [ #83 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @tangoone
Quote:
tangoone wrote: @Rose
but since "Gunnar" is the in the apollo team I guess he knows what he is talking about. |
its a matter of credibility. funny enough i remember when gunnars and igors opponents were proven wrong, maybe because i also read german forums. among these there were also some names of amiga hardware department, who (initially) claimed his achievements impossible or faked. maybe therefore there was much mistrust or even open hostility towards him and projects he was involved with. while he might not always openly admit that he changed his mind on something (like usage of coldfire or the matter of implementing bitfields in hardware) he is actually able to reconsider. so along the time i have observed the matter he gained some trust on my part. others opinion my differ, even if i dont know exactly what are they baising it on, maybe some other cathegories as me.Last edited by wawa on 06-Dec-2015 at 10:28 PM.
|
|
Status: Offline |
|
|
Rose
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 23:09:54
| | [ #84 ] |
|
|
 |
Cult Member  |
Joined: 5-Nov-2009 Posts: 982
From: Unknown | | |
|
| @wawa
I don't think that many people have doubted Igor since he's been so open about his development. If you check Gunnar's posts from Natami era you might notice quite quickly why people take his claims with mountain of salt... |
|
Status: Offline |
|
|
meet.mrnrg
 |  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 23:22:30
| | [ #85 ] |
|
|
 |
Super Member  |
Joined: 5-Feb-2007 Posts: 1919
From: UK, AUS, US | | |
|
| @ALL
Taking the long way around to achieve faster and sometimes newer features for the classic Amiga design might just do the trick.
I like what I'm seeing - so when I have it in my hands to play with Ill be laughing my head off.
FPGA future no one really knows what will be available (sky's the limit) for HIGH END FPGA's in the future, and also as other CPU architectures are showing signs of slowing down only ARM is a real contender for future AMIGA NEXT GEN and maybe even a contender for classic systems.
ALL IN ALL - Amiga FPGA teams are succeeding and innovating and improving year on year - all hail FPGA
_________________ Quote:
Easy Pocket Money, Freelancers & Experts Online | MiniMig FPGA, Sam440 Flex 733Mhz PPC, Amiga OS 4.1 Update 2, MorphOS 2.4, Other - AmiKit + Cloanto Amiga Forever 2008 + E-UAE, AmigaSYS |
|
Status: Offline |
|
|
SKOLMAN_MWS
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 23:46:37
| | [ #86 ] |
|
|
 |
Member  |
Joined: 19-May-2006 Posts: 77
From: Poland | | |
|
| |
Status: Offline |
|
|
OlafS25
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 23:50:13
| | [ #87 ] |
|
|
 |
Elite Member  |
Joined: 12-May-2010 Posts: 6472
From: Unknown | | |
|
| @Rose
when the Vampire 2 boards are available everyone can test and decide |
|
Status: Offline |
|
|
SKOLMAN_MWS
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 23:55:17
| | [ #88 ] |
|
|
 |
Member  |
Joined: 19-May-2006 Posts: 77
From: Poland | | |
|
| Next
 _________________ A1200 BlizzardPPC BVision FastATA Eth 3Com, EFIKA5K2
|
|
Status: Offline |
|
|
wawa
|  |
Re: News about Vampire and Apollo Posted on 6-Dec-2015 23:59:30
| | [ #89 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @Rose
Quote:
Rose wrote: @wawa
I don't think that many people have doubted Igor since he's been so open about his development. If you check Gunnar's posts from Natami era you might notice quite quickly why people take his claims with mountain of salt... |
im well aware of gunnars posting history on different forums, or do you want to point me to something particular? |
|
Status: Offline |
|
|
iggy
|  |
Re: News about Vampire and Apollo Posted on 7-Dec-2015 0:33:57
| | [ #90 ] |
|
|
 |
Super Member  |
Joined: 20-Oct-2010 Posts: 1175
From: Bear, Delaware USA | | |
|
| @SKOLMAN_MWS
Uh, that is a bare board.
And while I don't doubt that Majsta has produced boards, even finished products, I still don't see the relevance to whether the new core actually works or not.
I produce boards for projects myself. Like a V9958 based video card (designed by Canadian Luis Antoniosi).
My facebook page has a picture of those.
Latest batch of ten fresh from China. And I know these actually work. Last edited by iggy on 07-Dec-2015 at 12:38 AM. Last edited by iggy on 07-Dec-2015 at 12:37 AM. Last edited by iggy on 07-Dec-2015 at 12:35 AM.
|
|
Status: Offline |
|
|
wawa
|  |
Re: News about Vampire and Apollo Posted on 7-Dec-2015 14:09:31
| | [ #91 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| |
Status: Offline |
|
|
pavlor
|  |
Re: News about Vampire and Apollo Posted on 7-Dec-2015 16:29:17
| | [ #92 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @wawa
Quote:
some dnetc benchmark became available: |
Thanks!
I see on source picture (beware, hosting site is weird, virsus/spyware alert etc.!), OGR-NG test failed. OGR-NG scales better with performance, but RC5-72 result is also comparable. Close to 68060 80 MHz integer performance is safe lower boundary for current Vampire/Apollo speed.
dnet site for comparison: http://cgi.distributed.net/speed/
Note: dnetc benchmark results aren´t directly comparable between different CPU ISAs (eg. comparing 68040 to 68060 or 440EP to G3 is OK, but one must be cautious when comparing to Pentium or ARM).
What about lame benchmarks (for general FPU performance)? 
Edit: You may configure dnetc to use 68000 binary - this worked for me under OS4.1FE as OGR-NG/68020 similarly didn´t like 680x0 emulation.Last edited by pavlor on 07-Dec-2015 at 07:49 PM.
|
|
Status: Offline |
|
|
WOSPUPOS4
|  |
Re: News about Vampire and Apollo Posted on 7-Dec-2015 17:21:47
| | [ #93 ] |
|
|
 |
Regular Member  |
Joined: 24-Feb-2006 Posts: 220
From: Unknown | | |
|
| @iggy
Quote:
iggy wrote: @SKOLMAN_MWS
Uh, that is a bare board.
And while I don't doubt that Majsta has produced boards, even finished products, I still don't see the relevance to whether the new core actually works or not.
I produce boards for projects myself. Like a V9958 based video card (designed by Canadian Luis Antoniosi).
My facebook page has a picture of those.
Latest batch of ten fresh from China. And I know these actually work. |
Sigh! Same board running in an A500. Good enough ?
https://www.youtube.com/watch?v=9ncJs87hDkg
Edit: And before you point out the obvious. YES its NOT running Apollo core.Last edited by WOSPUPOS4 on 07-Dec-2015 at 05:22 PM.
|
|
Status: Offline |
|
|
cdimauro
|  |
Re: News about Vampire and Apollo Posted on 7-Dec-2015 22:20:12
| | [ #94 ] |
|
|
 |
Elite Member  |
Joined: 29-Oct-2012 Posts: 4127
From: Germany | | |
|
| @OlafS25
Quote:
OlafS25 wrote: @cdimauro
the person only said he might do something. Might do and really do are different beasts
|
Sure, but before that there was not even this possibility. Anyway, it was just the last comment. Other developers expressed similar opinions when WinUAE was able to run OS4. Quote:
VS 2015 for any of the NG platforms? Wow 
|
I don't how much work is needed, but here is an example of how to remotely debug a C++ application running on Linux, using Visual Studio 2015. Quote:
if not the comparation makes no sense.
|
Sorry, I don't understand what are you talking about here. Quote:
The only project I know of where is someone really try to port a modern tool is the guy behind Free Pascal Aros trying to port Lazarus to Aros. That would help a lot. But he is one guy so it needs time. No other projects of this kind I am aware of.
|
I think you should know Odyssey, right? Quote:
I do not spread nonsense and make such comparations. I compare 68k to "NG" that is Aros X86, AmigaOS and MorphOS and not to Windows or Linux.
|
Well, the non-sense is comparing 68K, which a retro platform, to the post-Amiga o.ses, which are active (so, no retro).
Apples and oranges. Quote:
What have the NG platforms proven? That they get lots of software.
|
Exactly: they GET software. Because they aren't dead platforms.
Can you tell me any reason why this will suddenly and immediately stop if any of such platform gets "modernized"? Quote:
I already mentioned the differences in codebase to you. In some cases the NG versions are a little newer than 68k, in some rare cases there is no 68k version.
|
It'll be interesting to compare your latest AROS Vision distribution with the latest Icaros.
@Overflow
Quote:
Overflow wrote: @cdimauro
Not really sure why this is even an discussion.
|
Because this's what a forum was made for? Quote:
If there are enough people around to purchase/support a product so the dev(s) can justify to themselves the production, who cares if its less efficient than socalled mainstream products?
|
Who stops them to make such choices? Quote:
There is no argument that "Amiga" spinoffs are less efficient than Windows, IOS and Linux, but luckily we diverse tastes/opinions about what is entertaining and sufficent.
|
Indeed. In fact there are peoples which make difference between an Amiga 500 with a 68K accelerator card, and another Amiga 500 which has a PC's Micro-ITX board inside, albeit they touch the same case, see the same graphic, and hear the same sound.
It's a psychological matter... Quote:
I do agree that all Amiga platforms will remain a curiosity, but that doesnt bother me.
|
It's just an hobby, so I don't see any reason for bothering...
@iggy
Quote:
iggy wrote: @cdimauro
Quote:
68K doom was already decided. |
YES!
I went from MC6800 to MC6809(and HD6309) to MC680XX. LOVED it. What a great ride. Motorola made some great stuff, but you are NEVER going to challenge the performance of a newer ASIC based processor with the FPGA re-implementation of an older CPU. Its not freaking happening (and no, I don't care what GVB has been telling you, its only application is legacy code).
|
There are high-end FPGAs too, but they are quite expensive. And developing an high-end softcore with a features set comparable to a modern processor will take a lot more time too. Quote:
Funny, you guys are still arguing about PPCs, without realizing that they were at one point a very valid direction to go (and for now they still adequately support our NG specifications).
The question is, how does this continue? I think its probably in the creation of a few last PPC boards (motherboards AND plug in cards) and a migration to X64.
This will be sorted out over the next few years, but I challenge anyone to prove me wrong five years from now.
|
In five years and without proper decisions made, the actual niche might become the hangout of pensioners...
@wawa
Quote:
wawa wrote: @cdimauro
Quote:
What's the problem here? JIT is a very good feature of WinUAE (and I think that there's room for improvements, both for general purpose code and FPU), like the new and bigger caches of the Apollo core. |
i have absolutely nothing against jit in uae, and it surely gives you enormous speed gain, supposedly never to be achieved on actual hardware like the fpga systems.
|
See above regarding FPGAs.
But for (FS|Win)UAE JIT there is the possibility to produce better and faster code.
However pay attention that such JIT is always limited by the fact that the same single process (and core) has to emulate the whole chipset at same time. So the performance are always castrated by such fact.
Trance and Petunia don't have the same problem, and that's why the latter can have better performance in WinUAE on the same machine, compared to WinUAE using its 68K JIT.
To fully exploit 68K -> x64 performance a new multicore emulator is required, which uses one core for the chipset and another one for the CPU (with a proper synchronization protocol, to avoid wasting precious CPU time and, therefore, performance). But it's a completely new project/design, so it's very unlikely that it can happen.
A more affordable solution can be something like Vamos or Emumiga, but with a proper JIT on top for 68K code. Quote:
however there are cases when you need to turn jit off, like when mmu needs to be involved or generally while testing and debugging. which is the resaon i run winuae mostly without jit these days. and its why i accidentaly noticed that apollo aibb benchmarks were better.
|
MMU wasn't used on Amiga, except for debugging reasons, so supporting JIT on a 68K + MMU isn't that urgent.
OS4 uses the MMU, and it works even using QEMU's PowerPC -> x86/x64 JIT. So, it can be possible to do the same with 68K, but it takes time, as usual.
However I think that a simpler and more efficient solution might be possible using a special Enforcer with some changes to WinUAE to partially emulate the behavior of a MMU, without emulating one of the 68K family. At the end, you only need to protect the 68K environment and catch illegal operations. |
|
Status: Offline |
|
|
wawa
|  |
Re: News about Vampire and Apollo Posted on 7-Dec-2015 22:57:20
| | [ #95 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @cdimauro
im not sure mmu is only used for debugging on 68k, there are some cases when its involved with hardware drivers, like warp3, but there are some cases with zorro bus memory management or something thor could better elaborate on.
sure its not essential especially for virtual system like uae.. |
|
Status: Offline |
|
|
kolla
|  |
Re: News about Vampire and Apollo Posted on 8-Dec-2015 5:04:05
| | [ #96 ] |
|
|
 |
Elite Member  |
Joined: 20-Aug-2003 Posts: 3359
From: Trondheim, Norway | | |
|
| @SKOLMAN_MWS
What does it mean that it says "overclock ready"? Can the FPGA be overclocked? _________________ B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC |
|
Status: Offline |
|
|
wawa
|  |
Re: News about Vampire and Apollo Posted on 8-Dec-2015 16:24:48
| | [ #97 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @pavlor
Quote:
What about lame benchmarks (for general FPU performance)? |
benchmarks are slowly coming up. keep an eye on their site and forum. |
|
Status: Offline |
|
|
iggy
|  |
Re: News about Vampire and Apollo Posted on 8-Dec-2015 17:36:58
| | [ #98 ] |
|
|
 |
Super Member  |
Joined: 20-Oct-2010 Posts: 1175
From: Bear, Delaware USA | | |
|
| @WOSPUPOS4
Quote:
Same board running in an A500. Good enough ? |
Of course not! That is a design for an 8 bit board. Although, there is one really nice fpga based video board that comes pretty close to Amiga specs that I intend to buy for my Atari 130XE.
For the Amiga, heck they can handle convention expansion cards. So why would you reinvent the wheel? BTW, good enough for an A500? Weren't we beyond that with RTG a long time ago? The point is, the board I mentioned is done and we are moving on. How long has Gunnar been working on CPUs?
|
|
Status: Offline |
|
|
tangoone
|  |
Re: News about Vampire and Apollo Posted on 8-Dec-2015 18:14:08
| | [ #99 ] |
|
|
 |
Regular Member  |
Joined: 2-Jul-2014 Posts: 152
From: Norway | | |
|
| |
Status: Offline |
|
|
Birbo
|  |
Re: News about Vampire and Apollo Posted on 8-Dec-2015 18:50:47
| | [ #100 ] |
|
|
 |
Cult Member  |
Joined: 5-Apr-2007 Posts: 602
From: Zurich, Switzerland | | |
|
| @tangoone
Awesome! _________________ Sometimes we give people a lot of credit just because they're writing nice sentences even if it isn't adding up to much. |
|
Status: Offline |
|
|