![]() |
![]() |
#41 |
Banned
Join Date: Jan 2010
Location: Kansas
Posts: 1,284
|
Dhrystone measures the compiler performance more than the CPU performance. Turning the results into a ratio will not help. CoreMark (embedded) is a better integer benchmark with low requirements but is still dependent on compiler performance which is a handicap for the 68k.
The 68k Mac OS system calls used A-line traps to the OS operating in Supervisor mode which are very slow but potentially give better security than the AmigaOS. The AmigaOS not only stays in user mode but also passes arguments in registers instead of using the more common 68k AT&T ABI which passes them on the stack. Most compilers (and therefore benchmarks) will use the ancient stack based AT&T ABI also giving the 68k a further handicap which the AmigaOS creators were able to partially avoid. |
![]() |
![]() |
#42 | |
son of 68k
Join Date: Nov 2007
Location: Lyon / France
Age: 50
Posts: 5,204
|
Quote:
![]() Oddly enough the Atari ST is the most secure of the three ![]() |
|
![]() |
![]() |
#43 | |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,446
|
Quote:
![]() SysSpeed test results -> http://amiga.resource.cx/perf/sysspeed.html Blizzard 1220 (020/28, 3.0, Maxtor 2585AT, DC-FFS, 512) = 9.33 MIPS M-Tec 68020i (020/14, 882/14, ECS, 3.1, Quantum CTS120S, FFS, 512, Nexus SCSI) = 4.56 MIPS Blizzard 1230 IV (030/50, 3.0, Quantum Lightning 730A, I-FFS, 512) = 16.40 MIPS Blizzard 4030 (030/50, 882/50, 3.1 in RAM, Quantum LP52S, FFS, 512) = 16.50 MIPS EDIT: In the Description of tests is said the following: "Note: There is no real method to test MIPS/MFLOPS, SysSpeed uses its own method and its results cannot be compared to those of other benchmark programs." So try to measure your system with SySSpeed to get similar results! Last edited by amilo3438; 06 January 2017 at 21:33. |
|
![]() |
![]() |
#44 | |
It's all in the reflexes!
Join Date: Nov 2009
Location: Wingkong warehouses
Posts: 206
|
Quote:
|
|
![]() |
![]() |
#45 | |||
Banned
Join Date: Dec 2016
Location: Nottingham, UK
Posts: 481
|
Quote:
Amstrad Action, Amstrad Action, You're Not Fit To Wipe my arse... You're Not Fit to wipe my arse." ![]() (Unofficial, but if you didn't have a theme song of sorts, you weren't a real Future magazine) Quote:
Quote:
![]() If you play from floppy, some games like Civilization are dogs on an AGA Amiga with no fast RAM in AGA, even from hard disk. It was designed for hard disk, but more importantly, some fast RAM too. Sid Meir didn't know any different, that the Amiga version was very very chuggy and clunky at loading from 1MB or more of only Chip RAM. As far as Microprose knew, all future 1MB or more Amigas would have some fast RAM. I don't think ANYBODY ever played the Amiga version from floppy on such a chip mem only machine until long after it was released. Could be wrong on that. The guy I knew who reviewed it was concerned at issues from playing from floppy, but conceded that from hard drive it was truly deep and rewarding. He played it on an 512KB chip RAM 1MB A500, almost certainly mostly from hard drive with extra fast RAM too, but no accelerator. Also, if you are playing ECS game, display in ECS at original res (NTSC for a game like Amiga Civ that came from America) for best fidelity. That helps some on its own. On topic, I guess the way you really have to rate the whole comparison is bus sizes for expansions, processor, chipset, memory, drive possibilities, graphic possibilities. You can do that for individual Amiga types of board and daughter expansion boards, and PC types of mainboard and expansions. Some of it is comparable to the 286-386, but the principle difference advantage on the Amiga side is potential total bus address (RAM, ROM, expansions, everything possible). The A1000 had limitations here, the other Amigas less so. But, the processor socket of even an A1000 can still be used for a bigger 32 bit processing module, which wasn't really an option on 286 and 386 machines, that I am aware of. Where the PCs maybe had the advantage in design was network ability. Every single PC ever, had a whole array of network options. Machines like the A1000, A500, and even the big box Amigas to a certain extent were kind of limited in that respect.Plus, you could do 256 colour graphics and 16 bit sound a lot easier than you could do it any Amiga without spending a lot. Such things worked a whole lot better on a 486, and that's maybe where the Amiga came unstuck - it didn't usually come in a standardized box for all types, that you could just swap the boards, in and out. You could take your old cards from 286 and 386 with you. Price difference and availability of hardware options is the real difference, I guess. You should get better results in ways using an Amiga, but you wouldn't get it cheaper, you'd have to pay more to find out if it worked better. Offputting to people, an unsure question to answer. Last edited by Pat the Cat; 07 January 2017 at 22:10. |
|||
![]() |
![]() |
#46 |
Registered User
Join Date: Sep 2007
Location: Stockholm
Posts: 4,259
|
|
![]() |
![]() |
#47 |
Banned
Join Date: Dec 2016
Location: Nottingham, UK
Posts: 481
|
Shouldn't be slow on any Amiga loaded from ram disk assigned properly, not the original ECS version... I guess it really also depends on whether you are playing the AGA version or the original version, which had an intro that you could not skip. I never played the AGA version. Much nicer looking, apparently.
|
![]() |
![]() |
#48 |
Registered User
Join Date: May 2015
Location: Somewhere in Time
Posts: 444
|
I tried the whdload version on a winuae a1200 setup. It totally sucked without fast mem but with 8mb fastmem it was ok. I remember it running A LOT better on a 486DX33 with 4mb RAM however.
|
![]() |
![]() |
#49 | |
-
Join Date: Jul 2003
Location: Helsinki / Finland
Age: 43
Posts: 9,687
|
Quote:
Hehe, yep, also it seems Apple guys were always so proud to save HW costs by doing stuff in SW, but this was a time where the machines couldn't really pull it off in SW yet. :-) Last edited by Jope; 08 January 2017 at 09:09. |
|
![]() |
![]() |
#50 |
Registered User
Join Date: Oct 2009
Location: Germany
Posts: 3,250
|
Civilization is slow (not HD loading times) on my A1200 040/40 32MB. On Shapeshifter emulations Civ it is much better. So I guess game code isn`t the best.
|
![]() |
![]() |
#51 |
Inviyya Dude!
Join Date: Sep 2016
Location: Amiga Island
Posts: 2,617
|
What would a Vampire V2 compare to in the x86 world?
|
![]() |
![]() |
#52 |
Total Chaos forever!
Join Date: Aug 2007
Location: Waterville, MN, USA
Age: 48
Posts: 2,103
|
|
![]() |
![]() |
#53 | |
Banned
Join Date: Dec 2016
Location: Nottingham, UK
Posts: 481
|
Quote:
The game engine should be very similar, if not identical, between the 2 versions. Last edited by Pat the Cat; 08 January 2017 at 22:27. |
|
![]() |
![]() |
#54 |
Inviyya Dude!
Join Date: Sep 2016
Location: Amiga Island
Posts: 2,617
|
|
![]() |
![]() |
#55 |
Registered User
Join Date: Oct 2009
Location: Germany
Posts: 3,250
|
@Pat the Cat:
Civilization AGA I`m talking about but as you mentioned ECS might be similar (out of memory). If I remember right Civ on A500 + slowram was already (very) slow. I built an own startup-sequence with several patches for max speed. Its much better but you still can see (16x16?) gfx blitting when moving on map. Unit moving and menu navigation could be faster too. All on a small Lowres screen. With mac emulation (shapeshifter) on 640x480 8bit it is much better. The opposite should be the case. |
![]() |
![]() |
#56 |
Registered User
Join Date: Jan 2015
Location: Melbourne, Australia
Posts: 548
|
I tried the OCS/ECS version of Civ in the current version of WinUAE on an emulated stock A1200, from the disks (not installed to HDD) and it was just like I remembered it, in that trying to use the mouse for anything required holding the left button down for a moment, simply clicking it didn't register a mouse click. I didn't play it long enough to have a decent size civilisation to manage, but it was still quite sluggish, and I'm not talking about when it needed to access the disk, which wasn't what made it unbearable in my experience, it was the sluggishness of the game itself when it wasn't accessing the disk. I honestly don't remember if I had the AGA version or not back in the day, probably not.
|
![]() |
![]() |
#57 | |
Registered User
Join Date: Jun 2021
Location: United Kingdom
Posts: 7
|
Quote:
What? It ran fine on an A500, the only problem being disk loading time. There wasn't huge waiting for the CPU to do things. |
|
![]() |
![]() |
#58 | |
Registered User
Join Date: Jun 2021
Location: United Kingdom
Posts: 7
|
Quote:
Here's a six years late quibble, and almost nobody is likely to read it, but here goes ![]() Last edited by Amiga1991; 19 September 2023 at 18:22. |
|
![]() |
![]() |
#59 | |
Registered User
Join Date: Aug 2013
Location: Marseille / France
Posts: 953
|
Quote:
You need to remember that your 2023 demande are not on par with 1993 ones. Many (many) peoples played countless hours on Civ on the Amiga, both ECS and AGA versions, and personnally I'm still playing it nowadays on a configuration lesser than an A4000. The Amiga version was very playable (and still is). Last edited by sokolovic; 19 September 2023 at 00:33. |
|
![]() |
![]() |
#60 |
Zone Friend
Join Date: Mar 2004
Location: Middle Earth
Age: 39
Posts: 2,068
|
maybe a good cpu benchmark is compression and decompression as it doesn't rely on what graphics mode the display has compared to 3d engines.
|
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Fake86 (68k SDL Port) - 8086 PC emulator | NovaCoder | project.Amiga Game Factory | 5 | 08 November 2018 11:13 |
Checking for genuine CPUs? | Mr.Flibble | support.Hardware | 0 | 03 July 2015 21:08 |
The origin of the terms chipmusic/chiptune | goto80 | Amiga scene | 28 | 20 March 2011 23:33 |
Motorola CPUs for sale... | amigamaniac | MarketPlace | 4 | 04 December 2009 23:48 |
Amiga specific hardware terms.... | Paul_s | Nostalgia & memories | 36 | 14 October 2009 11:12 |
|
|