English Amiga Board


Go Back   English Amiga Board > Coders > Coders. Asm / Hardware

 
 
Thread Tools
Old 23 December 2014, 11:47   #21
meynaf
son of 68k
meynaf's Avatar
 
Join Date: Nov 2007
Location: Lyon / France
Age: 44
Posts: 2,459
Quote:
Originally Posted by phx View Post
Same reboot on the Minimig as before. So you might want to send me the source of this small program, and I see what I can find out.
Yup, seen that too late. Big source sent instead.


Quote:
Originally Posted by phx View Post
EDIT: Found the reason for the Minimig reboot.
You are jumping to $e80000 and probably expect that an exception will happen sooner or later (bus error, illegal instruction, etc.). This doesn't happen here. The Minimig has $00 bytes from $e80000 to $f80000. So it executes ORI.W #0,D0 until it reaches the ROM reboot code.
On my A1200 it's full of $FF bytes, so i get a nice LINE-F.
So it worked by accident. Same when the area is protected with an MMU.

Damned myself there. This test is supposed to run in TRACE mode, but didn't activate the trace !

So normally you get a trace exception after the JMP and the area is never touched. But uae gives a BERR instead

It's fixed now.


Release 1.03 is online. But i guess it still crashes for some, huh ?

Btw. Those who have a problem under uae should try with the "more compatible" option (at least winuae has this one).
meynaf is offline  
AdSense AdSense  
Old 23 December 2014, 15:44   #22
phx
Natteravn

phx's Avatar
 
Join Date: Nov 2009
Location: Herford / Germany
Posts: 1,019
Quote:
Originally Posted by meynaf View Post
Release 1.03 is online.
Thanks. That's better. Here comes the Minimig:
Code:
Mey's benchmark suite v1.03             CPU : 68000
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  29      (46 fps)
        Sqr64 calc                      Fail
        Instructions overview           Fail
        Nbr64 factorizing               Fail
        Control scan                    37      (4647 ms)
        Copy fast                       Fail
        Homm2 c2p                       14      (19857 /100 ms)
        System performance              67      (1680 ts/sec)
        Amiga2Date                      19      (16176 calc/sec)
        Mpega imdct_l                   29      (1739 calc/sec)
        Flac decode                     20      (5620 ms)
        Script interpretation           17      (280 iter/sec)
And I also switched on my A1200 with ACA1231 (40MHz):
Code:
Mey's benchmark suite v1.03             CPU : 68030
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  96      (149 fps)
        Sqr64 calc                      82      (16726 calc/sec)
        Instructions overview           97      (12232 iter/sec)
        Nbr64 factorizing               82      (7487 ms)
        Control scan                    106     (1621 ms)
        Copy fast                       112     (14966 kb/s)
        Homm2 c2p                       88      (3286 /100 ms)
        System performance              89      (2242 ts/sec)
        Amiga2Date                      88      (73169 calc/sec)
        Mpega imdct_l                   90      (5270 calc/sec)
        Flac decode                     84      (1360 ms)
        Script interpretation           82      (1365 iter/sec)
Quote:
But i guess it still crashes for some, huh ?
Indeed. Running the latest version on my A2500, with 68020, 68851 and 68881 (OS3.1) makes the system freeze.

With your debug-version I found out that it happens while executing move16. And I could reproduce the same effect with an assembler program just executing a single move16.

That's quite strange. The 68851 has PMMU instructions in the $f0xx area, while the 68881 FPU uses $f2xx. The move16 ($f6xx) should have caused an illegal instruction exception. Maybe something else is executed instead, freezing the system...
phx is offline  
Old 23 December 2014, 18:19   #23
Michael
A1260T/PPC/BV/SCSI/NET

Michael's Avatar
 
Join Date: Jan 2013
Location: Moscow / Russia
Posts: 616
It finally works fine on my miggy!

Normal system
Code:
Mey's benchmark suite v1.03             CPU : 68060
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  126     (195 fps)
        Sqr64 calc                      48      (9870 calc/sec)
        Instructions overview           143     (18008 iter/sec)
        Nbr64 factorizing               33      (18529 ms)
        Control scan                    Fail
        Copy fast                       134     (17904 kb/s)
        Homm2 c2p                       187     (1543 /100 ms)
        System performance              235     (5882 ts/sec)
        Amiga2Date                      353     (292294 calc/sec)
        Mpega imdct_l                   1324    (76793 calc/sec)
        Flac decode                     618     (186 ms)
        Script interpretation           264     (4349 iter/sec)
cpu nocache
System: 68060 68882 (INST: NoCache NoBurst) (DATA: NoCache NoBurst)
very very slow for Nbr64 test

Code:
Mey's benchmark suite v1.03             CPU : 68060
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  72      (112 fps)
        Sqr64 calc                      6       (1229 calc/sec)
        Instructions overview           21      (2701 iter/sec)
        Nbr64 factorizing               4       (149382 ms)
        Control scan                    83      (2049 ms)  <====== WORKS!!!!!
        Copy fast                       67      (8910 kb/s)
        Homm2 c2p                       53      (5440 /100 ms)
        System performance              92      (2309 ts/sec)
        Amiga2Date                      86      (71215 calc/sec)
        Mpega imdct_l                   167     (9710 calc/sec)
        Flac decode                     88      (1295 ms)
        Script interpretation           48      (806 iter/sec)
OXYPatcher!

Code:
Mey's benchmark suite v1.03             CPU : 68060
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  127     (196 fps)
        Sqr64 calc                      86      (17386 calc/sec) <== TURBO MODE!
        Instructions overview           40      (5050 iter/sec) <=== SLOW ...
        Nbr64 factorizing               42      (14589 ms)
        Control scan                    Fail
        Copy fast                       143     (19010 kb/s)
        Homm2 c2p                       177     (1632 /100 ms)
        System performance              219     (5494 ts/sec)
        Amiga2Date                      352     (291878 calc/sec)
        Mpega imdct_l                   1323    (76775 calc/sec)
        Flac decode                     614     (187 ms)
        Script interpretation           260     (4281 iter/sec)
Interesting to see how some things effect performance of different code for better or worth .
Michael is offline  
Old 23 December 2014, 23:36   #24
Don_Adan
Registered User
 
Join Date: Jan 2008
Location: Warsaw/Poland
Age: 49
Posts: 918
"CPU nodatacache" is enough for working Control Scan test for 68040/68060.
Don_Adan is offline  
Old 24 December 2014, 00:12   #25
Lonewolf10
AMOS Extensions Developer
Lonewolf10's Avatar
 
Join Date: Jun 2007
Location: near Cambridge, UK
Age: 38
Posts: 1,917
Just did some tests with the latest versions:


mbench12d - resets Amiga (seems to be same reason as given for Minimig, OR's then CMP's then hits reset code) when using 68000 CPU.

mbench - works with 68000 and 68030 CPU's (see below)



Amiga 600 config (in WinUAE 2.0.1), with 68000 CPU:

Code:
Mey's benchmark suite v1.03		CPU : unknown (id=$5C1010)
Results :
	Test				%vs030	(raw result)
	ST screen conv			14	(22 fps)
	Sqr64 calc			Fail
	Instructions overview		Fail
	Nbr64 factorizing		Fail
	Control scan			13	(13203 ms)
	Copy fast			7	(965 kb/s)
	Homm2 c2p			3	(73647 /100 ms)
	System performance		19	(482 ts/sec)
	Amiga2Date			5	(4588 calc/sec)
	Mpega imdct_l			6	(391 calc/sec)
	Flac decode			4	(23439 ms)
	Script interpretation		6	(109 iter/sec)

Amiga 600 config (in WinUAE 2.0.1), with 68030 CPU:

Code:
Mey's benchmark suite v1.03		CPU : unknown (id=$5C1577)
Results :
	Test				%vs030	(raw result)
	ST screen conv			44	(68 fps)
	Sqr64 calc			18	(3808 calc/sec)
	Instructions overview		Fail
	Nbr64 factorizing		15	(39082 ms)
	Control scan			30	(5571 ms)
	Copy fast			22	(2996 kb/s)
	Homm2 c2p			21	(13543 /100 ms)
	System performance		43	(1081 ts/sec)
	Amiga2Date			22	(18768 calc/sec)
	Mpega imdct_l			19	(1156 calc/sec)
	Flac decode			20	(5693 ms)
	Script interpretation		35	(576 iter/sec)

Last edited by Lonewolf10; 24 December 2014 at 00:20. Reason: added correct data for 68030
Lonewolf10 is offline  
Old 26 December 2014, 10:50   #26
meynaf
son of 68k
meynaf's Avatar
 
Join Date: Nov 2007
Location: Lyon / France
Age: 44
Posts: 2,459
Quote:
Originally Posted by phx View Post
Indeed. Running the latest version on my A2500, with 68020, 68851 and 68881 (OS3.1) makes the system freeze.

With your debug-version I found out that it happens while executing move16. And I could reproduce the same effect with an assembler program just executing a single move16.

That's quite strange. The 68851 has PMMU instructions in the $f0xx area, while the 68881 FPU uses $f2xx. The move16 ($f6xx) should have caused an illegal instruction exception. Maybe something else is executed instead, freezing the system...
Normally MOVE16 should cause a line-F... Then the test sees that and simply knows MOVE16 isn't available. Any other exception will do just as fine.

Indeed what you get is quite strange. But if you could isolate it with a single instruction, it's not my program that's in fault
meynaf is offline  
Old 26 December 2014, 20:55   #27
phx
Natteravn

phx's Avatar
 
Join Date: Nov 2009
Location: Herford / Germany
Posts: 1,019
Quote:
Originally Posted by meynaf View Post
Normally MOVE16 should cause a line-F...
Exactly.

Quote:
Indeed what you get is quite strange. But if you could isolate it with a single instruction, it's not my program that's in fault
A crashing program is always the coder's fault. But maybe you want to find a workaround for these configurations?

I made some more tests today, and it becomes even stranger:
- Illegal $f0xx/$f1xx PMMU instructions cause a Line-F.
- Illegal $f2xx/$f3xx FPU instructions cause a Line-F.
- $f5xx/$f7xx/$f9xx/$fbxx/$fdxx/$ffxx instructions cause a Line-F.
- $f4xx/$f6xx/$f8xx/$faxx/$fcxx/$fexx instructions freeze the system!

Maybe this is something in the 68020 coprocessor interface? As a workaround you could check the 68040-flag in AttnFlags or try to run an 68040 PFLUSH, which has $f5xx and should cause a Line-F here.
phx is offline  
Old 26 December 2014, 21:24   #28
Lonewolf10
AMOS Extensions Developer
Lonewolf10's Avatar
 
Join Date: Jun 2007
Location: near Cambridge, UK
Age: 38
Posts: 1,917
Quote:
Originally Posted by phx View Post
- Illegal $f0xx/$f1xx PMMU instructions cause a Line-F.
Err, can you please enlighten me on what a 'Line-F' is?
Lonewolf10 is offline  
Old 27 December 2014, 14:06   #29
phx
Natteravn

phx's Avatar
 
Join Date: Nov 2009
Location: Herford / Germany
Posts: 1,019
Quote:
Originally Posted by Lonewolf10 View Post
Err, can you please enlighten me on what a 'Line-F' is?
Sure.
Line-F is an instruction code between $f000 and $ffff, which is reserved for co-processor instructions (FPU, MMU). When you have no coprocessor, or executing an opcode from $fxxx which is not handled by any present coprocessor, then the system will cause a Line-F Exception. Which means the exception vector 11 (VBR+$44) is called. AmigaOS will show a Guru #0000000B.

There is also Line-A, for instruction codes between $a000 and $afff, but those were never used by any chip from the 68k family. ColdFire uses this range and the Atari TOS used it to emulate a graphics coprocessor (line drawing, etc.).

All other illegal instructions cause an illegal instruction exception. That's vector 4, or VBR+$10.
phx is offline  
Old 28 December 2014, 12:47   #30
Toni Wilen
WinUAE developer
 
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 42
Posts: 19,939
Quote:
Originally Posted by phx View Post
EDIT: Found the reason for the Minimig reboot.
You are jumping to $e80000 and probably expect that an exception will happen sooner or later (bus error, illegal instruction, etc.). This doesn't happen here. The Minimig has $00 bytes from $e80000 to $f80000. So it executes ORI.W #0,D0 until it reaches the ROM reboot code.
No real 68000 based Amiga has zeros (or all ones) at $e80000-$f7ffff. Reading it (or any other unmapped memory space) returns whatever 68000 has prefetched last. Which means most commonly it is MOVE opcode used to read the data
Toni Wilen is offline  
Old 28 December 2014, 13:56   #31
phx
Natteravn

phx's Avatar
 
Join Date: Nov 2009
Location: Herford / Germany
Posts: 1,019
Quote:
Originally Posted by Toni Wilen View Post
No real 68000 based Amiga has zeros (or all ones) at $e80000-$f7ffff.
Oh, that means we can use that to detect Minimig hardware?
phx is offline  
Old 28 December 2014, 17:44   #32
Thorham
Computer Nerd

Thorham's Avatar
 
Join Date: Sep 2007
Location: Rotterdam/Netherlands
Age: 41
Posts: 2,972
Quote:
Originally Posted by phx View Post
Oh, that means we can use that to detect Minimig hardware?
Until the Minimig core gets updated
Thorham is offline  
Old 28 December 2014, 20:53   #33
Lonewolf10
AMOS Extensions Developer
Lonewolf10's Avatar
 
Join Date: Jun 2007
Location: near Cambridge, UK
Age: 38
Posts: 1,917
Quote:
Originally Posted by phx View Post
Sure.
Line-F is an instruction code between $f000 and $ffff, which is reserved for co-processor instructions (FPU, MMU). When you have no coprocessor, or executing an opcode from $fxxx which is not handled by any present coprocessor, then the system will cause a Line-F Exception. Which means the exception vector 11 (VBR+$44) is called. AmigaOS will show a Guru #0000000B.

There is also Line-A, for instruction codes between $a000 and $afff, but those were never used by any chip from the 68k family. ColdFire uses this range and the Atari TOS used it to emulate a graphics coprocessor (line drawing, etc.).

All other illegal instructions cause an illegal instruction exception. That's vector 4, or VBR+$10.

Thanks for enlightening me
Lonewolf10 is offline  
Old 29 December 2014, 09:57   #34
meynaf
son of 68k
meynaf's Avatar
 
Join Date: Nov 2007
Location: Lyon / France
Age: 44
Posts: 2,459
Quote:
Originally Posted by phx View Post
A crashing program is always the coder's fault.
Not necessarily, there can be hardware faults as well.


Quote:
Originally Posted by phx View Post
Maybe this is something in the 68020 coprocessor interface?
Yes, this could be wrongly handled by the hardware and cause a freeze

The coprocessor interface shouldn't freeze on anything.

Only "strange" thing on 020/030 is that some (normally unused) opcodes trigger a privilege violation instead of a line-F when executed in user mode.


Anyway, thanks to all you for the test results.

I hope my benchmark can be of any use ; i attempted to test all areas, not only raw stupid number crunching.
68060 users can see the nice effect of instruction emulation here ; i hope softcore builders will drop that idea !
Now, missing is the uae jit test on a very fast machine. Is it really as powerful as it looks ?

Btw. Does anyone know of a cpu test able to differentiate the minimig from true 68000 ?
meynaf is offline  
Old 29 December 2014, 15:44   #35
phx
Natteravn

phx's Avatar
 
Join Date: Nov 2009
Location: Herford / Germany
Posts: 1,019
Quote:
Originally Posted by meynaf View Post
Btw. Does anyone know of a cpu test able to differentiate the minimig from true 68000 ?
The Minimig has a true 68000. There is a real Freescale 68SEC000 CPU on the board.
Some days ago I already looked for differences between 68SEC000 and 68000, which are visible in software, but found nothing.
phx is offline  
Old 29 December 2014, 17:53   #36
Michael Sykes
WinUAE end user

Michael Sykes's Avatar
 
Join Date: Jan 2014
Location: Bremen
Age: 38
Posts: 627
Tested on WinUAE, fast 040 config and a realy bloated scalos install.

Code:
18.Ram Disk:> mbench

Mey's benchmark suite v1.03             CPU : unknown (id=$05953E)
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  25255   (38894 fps)
        Sqr64 calc                      4357    (880224 calc/sec)
        Instructions overview           Fail
        Nbr64 factorizing               3751    (165 ms)
        Control scan                    160     (1069 ms)
        Copy fast                       12696   (1687424 kb/s)
        Homm2 c2p                       16111   (18 /100 ms)
        System performance              314     (7874 ts/sec)
        Amiga2Date                      4844    (4009205 calc/sec)
        Mpega imdct_l                   18806   (1090803 calc/sec)
        Flac decode                     7666    (15 ms)
        Script interpretation           5224    (85949 iter/sec)

18.Ram Disk:> mbench

Mey's benchmark suite v1.03             CPU : unknown (id=$05953E)
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  25040   (38562 fps)
        Sqr64 calc                      4310    (870772 calc/sec)
        Instructions overview           Fail
        Nbr64 factorizing               3706    (167 ms)
        Control scan                    161     (1063 ms)
        Copy fast                       10679   (1419257 kb/s)
        Homm2 c2p                       14500   (20 /100 ms)
        System performance              314     (7874 ts/sec)
        Amiga2Date                      4819    (3987983 calc/sec)
        Mpega imdct_l                   18846   (1093080 calc/sec)
        Flac decode                     6764    (17 ms)
        Script interpretation           5204    (85611 iter/sec)

18.Ram Disk:> mbench

Mey's benchmark suite v1.03             CPU : unknown (id=$05953E)
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  25090   (38640 fps)
        Sqr64 calc                      4331    (874910 calc/sec)
        Instructions overview           Fail
        Nbr64 factorizing               3706    (167 ms)
        Control scan                    158     (1086 ms)
        Copy fast                       12943   (1720129 kb/s)
        Homm2 c2p                       14500   (20 /100 ms)
        System performance              317     (7936 ts/sec)
        Amiga2Date                      4818    (3987443 calc/sec)
        Mpega imdct_l                   18634   (1080817 calc/sec)
        Flac decode                     6764    (17 ms)
        Script interpretation           5229    (86031 iter/sec)
Michael Sykes is offline  
Old 30 December 2014, 16:58   #37
modrobert
old bearded fool

modrobert's Avatar
 
Join Date: Jan 2010
Location: Bangkok
Age: 50
Posts: 425
UAE JIT test on a laptop with 2.30GHz AMD A10-4600M APU.

Code:
E-UAE 0.8.29-WIP4
Build date: Mar 27 2007 21:53:04
...
Building CPU table for configuration: 68020/881
1866 CPU functions
Building CPU function table, 45954 opcodes (3 0 0).
Max CPUID level=13 Processor is AuthenticAMD [x86-64]
JIT: Building compiler function table.
JIT: Supposedly 42346 compileable opcodes.
JIT: Enabled direct memory access.
JIT: Allocated 16384 KB translation cache.
Tested in AmiKit 1.7.0...

Code:
Mey's benchmark suite v1.03     CPU : unknown (id=$1D1176)
Results :
    Test                        %vs030  (raw result)
    ST screen conv              8448    (13011 fps)
    Sqr64 calc                  2663    (537942 calc/sec)
    Instructions overview       Fail
    Nbr64 factorizing           2726    (227 ms)
    Control scan                332     (518 ms)
    Copy fast                   2761    (366981 kb/s)
    Homm2 c2p                   5576    (52 /100 ms)
    System performance          2       (50 ts/sec)
    Amiga2Date                  3981    (3294452 calc/sec)
    Mpega imdct_l               10784   (625509 calc/sec)
    Flac decode                 3382    (34 ms)
    Script interpretation       3839    (63162 iter/sec)
modrobert is online now  
Old 31 December 2014, 03:16   #38
roomeo
Registered User

 
Join Date: Aug 2014
Location: Telemark
Posts: 203
Tested my A600 and A4000 with this benchmark tool:
Here are my results.. let me know if some other configuration of these setups needs to be tested..

Cheers!
Aksel


Code:
Stock A600 - SetPatch
Kickstart version 37.300, Workbench version 38.35

Mey's benchmark suite v1.03             CPU : 68000
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  14      (22 fps)
        Sqr64 calc                      Fail
        Instructions overview           Fail
        Nbr64 factorizing               Fail
        Control scan                    12      (13599 ms)
        Copy fast                       Fail
        Homm2 c2p                       3       (73804 /100 ms)
        System performance              19      (478 ts/sec)
        Amiga2Date                      5       (4647 calc/sec)
        Mpega imdct_l                   6       (391 calc/sec)
        Flac decode                     4       (23495 ms)
        Script interpretation           6       (109 iter/sec)
Code:
Stock ACA620/A604n - Setpatch
Kickstart 40.63, Workbench 40.42
System: 68020 (INST: Cache)

Mey's benchmark suite v1.03             CPU : 68020
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  38      (60 fps)
        Sqr64 calc                      33      (6666 calc/sec)
        Instructions overview           36      (4614 iter/sec)
        Nbr64 factorizing               32      (19173 ms)
        Control scan                    42      (4064 ms)
        Copy fast                       49      (6576 kb/s)
        Homm2 c2p                       35      (8120 /100 ms)
        System performance              65      (1642 ts/sec)
        Amiga2Date                      35      (29258 calc/sec)
        Mpega imdct_l                   35      (2058 calc/sec)
        Flac decode                     31      (3617 ms)
        Script interpretation           29      (489 iter/sec)
Code:
68020 (INST: NoCache)
Kickstart 40.63, Workbench 40.42

Mey's benchmark suite v1.03             CPU : 68020
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  35      (55 fps)
        Sqr64 calc                      29      (6038 calc/sec)
        Instructions overview           35      (4406 iter/sec)
        Nbr64 factorizing               31      (19681 ms)
        Control scan                    41      (4149 ms)
        Copy fast                       40      (5343 kb/s)
        Homm2 c2p                       24      (12008 /100 ms)
        System performance              65      (1633 ts/sec)
        Amiga2Date                      30      (24873 calc/sec)
        Mpega imdct_l                   35      (2057 calc/sec)
        Flac decode                     28      (4102 ms)
        Script interpretation           23      (391 iter/sec)
Code:
Stock A4000-a3640 - Setpatch - NoCache
System: 68040 68882 (INST: NoCache NoBurst) (DATA: NoCache NoBurst)
Kickstart 40.70, Workbench 40.42

Mey's benchmark suite v1.03             CPU : 68040
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  36      (56 fps)
        Sqr64 calc                      40      (8171 calc/sec)
        Instructions overview           31      (3888 iter/sec)
        Nbr64 factorizing               77      (7963 ms)
        Control scan                    33      (5116 ms)
        Copy fast                       20      (2685 kb/s)
        Homm2 c2p                       23      (12174 /100 ms)
        System performance              47      (1193 ts/sec)
        Amiga2Date                      31      (26062 calc/sec)
        Mpega imdct_l                   76      (4422 calc/sec)
        Flac decode                     30      (3733 ms)
        Script interpretation           16      (278 iter/sec)
Code:
Stock A4000-a3640 - Setpatch
System: 68040 68882 (INST: Cache Burst) (DATA: Cache CopyBack)
Kickstart 40.70, Workbench 40.42

Mey's benchmark suite v1.03             CPU : 68040
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  70      (108 fps)
        Sqr64 calc                      101     (20539 calc/sec)
        Instructions overview           98      (12324 iter/sec)
        Nbr64 factorizing               93      (6638 ms)
        Control scan                    Fail
        Copy fast                       28      (3732 kb/s)
        Homm2 c2p                       86      (3350 /100 ms)
        System performance              182     (4566 ts/sec)
        Amiga2Date                      112     (93256 calc/sec)
        Mpega imdct_l                   141     (8194 calc/sec)
        Flac decode                     117     (975 ms)
        Script interpretation           91      (1504 iter/sec)
Code:
Stock A4000-a3640 - No Setpatch
System: 68040 (INST: Cache Burst) (DATA: Cache Burst)
Kickstart 40.70, Workbench 40.42

Mey's benchmark suite v1.03             CPU : 68040
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  61      (95 fps)
        Sqr64 calc                      98      (19967 calc/sec)
        Instructions overview           92      (11615 iter/sec)
        Nbr64 factorizing               92      (6722 ms)
        Control scan                    26      (6415 ms)
        Copy fast                       39      (5220 kb/s)
        Homm2 c2p                       78      (3677 /100 ms)
        System performance              125     (3125 ts/sec)
        Amiga2Date                      104     (86144 calc/sec)
        Mpega imdct_l                   136     (7941 calc/sec)
        Flac decode                     115     (998 ms)
        Script interpretation           90      (1490 iter/sec)

Last edited by roomeo; 31 December 2014 at 03:18. Reason: do i need one?
roomeo is offline  
Old 01 January 2015, 12:23   #39
meynaf
son of 68k
meynaf's Avatar
 
Join Date: Nov 2007
Location: Lyon / France
Age: 44
Posts: 2,459
Thanks again for all results.

Here are some others i could get :


Unexpanded A1200
(got by disabling my B1230)
Code:
        ST screen conv                  47      (73 fps)
        Sqr64 calc                      27      (5454 calc/sec)
        Instructions overview           16      (2084 iter/sec)
        Nbr64 factorizing               25      (24494 ms)
        Control scan                    19      (8469 ms)
        Copy fast                       20      (2773 kb/s)
        Homm2 c2p                       23      (12094 /100 ms)
        System performance              37      (938 ts/sec)
        Amiga2Date                      23      (19654 calc/sec)
        Mpega imdct_l                   19      (1141 calc/sec)
        Flac decode                     20      (5557 ms)
        Script interpretation           19      (324 iter/sec)

A4000D Cyberstorm II 68060 75MHz
(early result, not from a machine i own)
Code:
        ST screen conv                  131     (202 fps)
        Sqr64 calc                      58      (11727 calc/sec)
        Instructions overview           176     (22079 iter/sec)
        Nbr64 factorizing               39      (15654 ms)
        Control scan                    Fail
        Copy fast                       145     (19339 kb/s)
        Homm2 c2p                       196     (1479 /100 ms)
        System performance              258     (6451 ts/sec)
        Amiga2Date                      441     (365672 calc/sec)
        Mpega imdct_l                   1664    (96532 calc/sec)
        Flac decode                     766     (152 ms)
        Script interpretation           310     (5105 iter/sec)

apollo_mini_1809_66 - CPU : unknown (id=$DC3116)
(this is a test from apollo-core.com softcore)
Code:
        ST screen conv                  45      (70 fps)
        Sqr64 calc                      Fail
        Instructions overview           Fail
        Nbr64 factorizing               Fail
        Control scan                    21      (7853 ms)
        Copy fast                       180     (23979 kb/s)
        Homm2 c2p                       64      (4498 /100 ms)
        System performance              28      (720 ts/sec)
        Amiga2Date                      232     (191986 calc/sec)
        Mpega imdct_l                   723     (41975 calc/sec)
        Flac decode                     307     (374 ms)
        Script interpretation           267     (4405 iter/sec)
meynaf is offline  
Old 18 January 2015, 15:43   #40
pawelek
Registered User

 
Join Date: Aug 2014
Location: Poland
Posts: 38
A1200 + Blizzard 1260/64 Mhz + exec 45.20 + MMULib (68060.library 43.6) + MuFastZero MOVESSP FASTEXEC ON

Code:
Mey's benchmark suite v1.03             CPU : 68060
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  116     (180 fps)
        Sqr64 calc                      39      (7971 calc/sec)
        Instructions overview           63      (7940 iter/sec)
        Nbr64 factorizing               35      (17455 ms)
        Control scan                    Fail
        Copy fast                       141     (18808 kb/s)
        Homm2 c2p                       178     (1628 /100 ms)
        System performance              236     (5917 ts/sec)
        Amiga2Date                      376     (311657 calc/sec)
        Mpega imdct_l                   1413    (81973 calc/sec)
        Flac decode                     653     (176 ms)
        Script interpretation           277     (4570 iter/sec)

the same + MuRedox (Oxypatcher replacement)

Code:
Mey's benchmark suite v1.03             CPU : 68060
Results :
        Test                            %vs030  (raw result)
        ST screen conv                  116     (180 fps)
        Sqr64 calc                      84      (17001 calc/sec)
        Instructions overview           44      (5590 iter/sec)
        Nbr64 factorizing               48      (12814 ms)
        Control scan                    Fail
        Copy fast                       141     (18777 kb/s)
        Homm2 c2p                       178     (1628 /100 ms)
        System performance              229     (5747 ts/sec)
        Amiga2Date                      376     (311613 calc/sec)
        Mpega imdct_l                   1415    (82071 calc/sec)
        Flac decode                     653     (176 ms)
        Script interpretation           277     (4569 iter/sec)

Last edited by pawelek; 18 January 2015 at 15:51.
pawelek is offline  
AdSense AdSense  
 


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

Similar Threads
Thread Thread Starter Forum Replies Last Post
Benchmark Code marduk_kurios Coders. Asm / Hardware 3 20 February 2014 19:14
functions benchmark wawa Coders. System 2 15 April 2013 19:55
External windows program communicating with program running inside WinUAE xxxxx support.WinUAE 10 19 February 2013 10:27
68k benchmark suggestions fishyfish Amiga scene 5 20 May 2012 17:32
Benchmark software. Johnyt request.Apps 3 21 February 2005 03:10

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 04:52.


Powered by vBulletin® Version 3.8.8 Beta 1
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
Page generated in 0.40861 seconds with 12 queries