• Please review our updated Terms and Rules here

Looking for volunteers to help test a new benchmark

16 MHz comparative

16 MHz comparative

These are the results of my four 16 MHz machines. I can't understand them...

IBM PS/2 8555 (Model 55SX) 80386SX 16 MHz Microchannel
Code: 
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97b
;This file contains fingerprinting information about your computer.  Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UIDFF2B107F9D]
MemoryTest=480
OpcodeTest=274
VidramTest=1253
MemEATest=266
3DGameTest=196
Score=21
CPU=Intel 80386SX
CPUspeed=16 MHz
BIOSinfo=        COPR. IBM 1981, 1988 (11/02/88, rev. 0)
MachineModel=0CF8
BIOSdate=19881102
BIOSCRC16=FF2B
VideoSystem=VGA
VideoAdapter=VGA, unknown Chipset, 256kb Video Memory (BIOS)
Machine=PS/2 Model 55SX 16MHz 386SX

IBM PS/1 Model 2011 80286 overclocked to 16,6 MHz
Code: 
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97b
;This file contains fingerprinting information about your computer.  Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UID9C90133282]
MemoryTest=434
OpcodeTest=186
VidramTest=1230
MemEATest=241
3DGameTest=183
Score=23
CPU=Intel 80286
CPUspeed=17 MHz
BIOSinfo=        COPR. IBM 1981, 1989 (02/16/90, rev. 0)
MachineModel=0BFC
BIOSdate=19900216
BIOSCRC16=9C90
VideoSystem=VGA
VideoAdapter=VGA, unknown Chipset, 256kb Video Memory (BIOS)
Machine=PS/1 Model 2011 286-10

IBM PS/1 Model 2121 80386SX 16 MHz
Code: 
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97b
;This file contains fingerprinting information about your computer.  Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UIDE4AFC7E7C]
MemoryTest=519
OpcodeTest=199
VidramTest=319
MemEATest=259
3DGameTest=193
Score=33
CPU=Intel 80386SX
CPUspeed=16 MHz
BIOSinfo=        COPR. IBM 1981, 1991 (06/25/91, rev. 0)
MachineModel=30F8
BIOSdate=19910625
BIOSCRC16=E4AF
VideoSystem=VGA
VideoAdapter=VGA, unknown Chipset, 256kb Video Memory (BIOS)
Machine=PS/1 Model 2121 16MHz 386SX

Acer M1207 AT Clone with 80286 16 MHz
Code: 
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97b
;This file contains fingerprinting information about your computer.  Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UID9C41130701]
MemoryTest=393
OpcodeTest=204
VidramTest=317
MemEATest=266
3DGameTest=195
Score=36
CPU=Intel 80286
CPUspeed=16 MHz
BIOSinfo=R(C)1985-1990, American Megatrends Inc.,All Rights Reserved.1346 Oakbrook Drive, Suite-120, Norcross, GA-30093. Phone-(404)-263-8181. (06/13/90, rev. 0)
MachineModel=01FC
BIOSdate=19900613
BIOSCRC16=9C41
VideoSystem=VGA
VideoAdapter=VGA, Trident TR8900CL or D, VESA, 768kb Video Memory, 512kb Vide
Machine=AT clone

And here are my questio
-I can't understand why a 386 with MCA bus performs so bad, I thought it would be the fastest machine of the tested but it is the slowest.
-How can the 286 AT clone perform so good? It wins the comparative.
-I find very strange the VidramTest results: 1253 and 1230 in slower machines, 319 and 317 on faster machines. I think the four VGAs are similar, maybe the trident on the AT clone should perform a little better than the rest...
 
I think the differences can be explained by the fact that the 386 is actually slower than the 286 at the same speed. The 386 SX internally has a 32 data bus and an external 16 bit data bus. There is probably a loss of efficiency due to this. Also some instructions may take another clock cycle to execute on the 386. Therefore the 286 may be slightly faster at the same speed.

The VGA chip set found in each system has been shown to have a substantial of fact on the overall score. Early IBM VGA chip sets have been recognized be among the slowest ever made. Your results seem to indicate this may be true even if IBM did not manufacture the chip set itself. I would not be surprised if IBM used eight bit VGA chip sets in the first two systems whose benchmark results you reported.
 
Great Hierophant said:
Early IBM VGA chip sets have been recognized be among the slowest ever made. Your results seem to indicate this may be true even if IBM did not manufacture the chip set itself. I would not be surprised if IBM used eight bit VGA chip sets in the first two systems whose benchmark results you reported.
Click to expand...

That makes sense, both machines have IBM 90X8941 VGA chipset.
Thanks for the explanation!
 
Last edited:
486 DX4-100 Benhcmark

486 DX4-100 Benhcmark

CAT Computers (486 DX4-100 Whitebox PC)

[UIDF9DDBEC30]
MemoryTest=23586
OpcodeTest=23592
VidramTest=30462
MemEATest=33779
3DGameTest=40760
Score=186
CPU=Intel i486DX4
CPUspeed=100 MHz
BIOSinfo=,d Software Inc. Awar,Copyright (C) 1984-94, Award Software, Inc. (10/17/95, rev. 0)
MachineModel=01FC
BIOSdate=19951017
BIOSCRC16=F9DD
VideoSystem=VGA
VideoAdapter=VGA, S3 unknown, VESA, 1024kb Video Memory, 1024kb Video Memory
Machine=AT clone

Will do the other's I have later
 
nestor said:
I find very strange the VidramTest results: 1253 and 1230 in slower machines, 319 and 317 on faster machines. I think the four VGAs are similar, maybe the trident on the AT clone should perform a little better than the rest...
Click to expand...

Smaller numbers are better for the microsecond timing lines, so 319 is a faster result than 1253.

Also, because the CPU must inevitably perform the tests, CPU speed can affect the vidramtest line. For example, if you put a VGA card in an 8088 and run the test, then you pull it out and put it a 386 and run the test, the vidramtest will have smaller numbers on the 386 even though it is the same card. The vidramtest is a test of how quickly the CPU can move system RAM to video RAM.

Larger numbers are better for the Score= line.
 
Hello Trixter!
I also used your program on a rare machine that is called KAT from a Greek company(that no longer exists) called Gigatronics.You can find some info about it here: http://www.applefritter.com/node/205 and my presentation in a Greek forun here: http://translate.google.com/translate?sl=el&tl=en&js=n&prev=_t&hl=el&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.retromaniax.gr%2Fvb%2Fshowthread.php%3F3807-Gigatronics-%28KAT-computer%29%2Fpage7%26highlight%3Dgigatronics&act=url and here: http://translate.google.com/translate?sl=el&tl=en&js=n&prev=_t&hl=el&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.retromaniax.gr%2Fvb%2Fshowthread.php%3F3807-Gigatronics-%28KAT-computer%29%2Fpage9%26highlight%3Dgigatronics&act=url (you must be registered to see the photos).It is hybrid system that has a 8088 (so it is an IBM compatible) but it also has a 65c816 that makes it compatible with Apple II :grin:. When I'll get the apple dos disks I'll present it to our forum.
The results from your program are:
Code: 
[UID54132F60]
MemoryTest=3810
OpcodeTest=1822
VidramTest=2487
MemEATest=2020
3DGameTest=1920
Score=4
CPU=Intel 8088
CPUspeed=4.77 MHz
BIOSinfo=unknown
MachineModel=0000
BIOSdate=20001001
BIOSCRC16=5413
VideoSystem=CGA
VideoAdapter=CGA
Machine=unknown, ID : 1449

Your program has also another interesting "side" effect:after running it I can play the classic game "Alley cat" .Without running your program first if I try to run the game the system restarts :( Any explanation?
 
Last edited:
Cimonvg said:
hello trixter
got my "rev to 486" today. Have to read the manual - but another day, I will replace the original 386 with a "486" ;)
/cimonvg
Click to expand...

Neat However, it still won't speed it up that much. Your memory bandwidth is what is really hurting the score. Installing the 486 may actually finally get you to 386 speeds
__________________
download 5000 backlinks
 
I'll just stick my Pentium MMX computer here.

[UIDCB8A16E4D1]
MemoryTest=28
OpcodeTest=15
VidramTest=155
MemEATest=20
3DGameTest=20
Score=295
CPU=Intel Pentium with MMX
CPUspeed=200 MHz
BIOSinfo=(C)1985-1996, American Megatrends Inc.,All Rights Reserved,6145F Northbelt Parkway,GA-30071,USA.(770)-263-8181. (07/15/95, rev. 0)
MachineModel=01FC
BIOSdate=19950715
BIOSCRC16=CB8A
VideoSystem=VGA
VideoAdapter=VGA, S3 86c868/86c968 PCI, VESA, 256kb Video Memory (BIOS)
Machine=AT clone

I typed this whole thing out because I'm too lazy to turn on my other computer that has a floppy drive and usb support at the moment, so excuse me for any errors, but I'm pretty sure that its all right.
 
Hello
yes - still old pc , but feels quick. On page 9 (in this tread) you can se I got from score=23 to score=42 :)
The Inboard uses RAM from the Inboard so it has 32 bit access from CPU->RAM . (the 256kb motherboard RAM is only for the bootup).
...and I did try windows 95 on the PC , but after install it lock's / halts the pc :( ... But it is my hobby - so I try again some other day.
/cimonvg

kritiana0 said:
Neat However, it still won't speed it up that much. Your memory bandwidth is what is really hurting the score. Installing the 486 may actually finally get you to 386 speeds
__________________
download 5000 backlinks
Click to expand...
 
angel_grig said:
Your program has also another interesting "side" effect:after running it I can play the classic game "Alley cat" .Without running your program first if I try to run the game the system restarts :( Any explanation?
Click to expand...

Wow, you're saying running my benchmark actually increases compatibility for some software? That's crazy. The only explanation I can think of is that I reprogram the timer from its normal rate of 54.9ms to 50ms for performing the tests, then I set it back to normal. Maybe this corrects something odd your BIOS does? Thanks for the numbers, btw -- you have a nice stock 4.77Mhz 8088 under the hood.

BTW, to everyone here, I'm really sorry I haven't finished the final tool yet; lots of stuff getting in the way. I have thought about just forgetting a nice CUI using Turbo Vision and just coding up the interface myself, and not worrying about editing entries from within the interface (would have to do it by editing a text file). The result would be pretty fugly but at least it would get done...
 
This may be a dumb question, but is there an option to display the results immediately instead of writing them to a file? It cuts down on the steps, which can be helpful when testing many different slowdown options.
 
Great Hierophant said:
is there an option to display the results immediately instead of writing them to a file?
Click to expand...
The reason he made the program write the results to a file is so that then the tester could then post the results here by just copying and pasting the text file contents. Much easier.

Maybe an option that you'd like better is after each test completes, the benchmark asks the user whether it should write the results to a file instead of automatically doing it. Then, you'd only write the wanted results to the disk.
 
hello
found my old (and still untested) "Number smasher rev.3.0" card. And in thiese days of benchmarking it could bee funny if the smasher really gives double performance. The number smasher is a 8086+640Kb 16 bit RAM add in card for the 8088 PC....
...hope it still works!
/cimonvg
 
Unless it runs at 8MHz or faster, it won't give truly "double" performance. If it runs at 7.16MHz (likely) it will give roughly 1.75x performance. In any case, feel free to post a score.
 
Well, I finally tried this on two of my older computers. Here's what I got on my IBM 5150 with a V20:

[UID3B8510F6]
MemoryTest=2370
OpcodeTest=1420
VidramTest=2070
MemEATest=1729
3DGameTest=1318
Score=6
CPU=NEC V20
CPUspeed=4.77 MHz
BIOSinfo=5700945 COPR. IBM 1981 (05/02/82, rev. 0)
MachineModel=0000
BIOSdate=19820502
BIOSCRC16=3DCD
VideoSystem=EGA
VideoAdapter=EGA
Machine=PC

I have a VGA Wonder 16, but it is in EGA mode so maybe that's why it didn't ID that. I don't really understand what these numbers even mean; lower is better? What's 3D game test? Is there an explanation of this somewhere that I'm missing?

The benchmark seemed to just do nothing on my Packard Bell 550 when it got to "Checking CPU Speed." The computer wasn't locked up because I could toggle my CAPS LOCK, but I left it for half an hour and it never got any further. It didn't save anything to the log either. It was a 90 MHz Pentium, so not like it really matters as this is designed to benchmark older systems. Yes, I did try DSTUB too.

I'll have to see what I get on some other systems sometime. Very interesting; I do hope you finish it soon: It will be very useful and was much needed.

-Isaac
 
Ok, I actually read all 22 pages of this thread and that answered most of my questions: I see these numbers are how many milliseconds the operation took so obviously lower is better. Sill not sure on what some of them mean like MemEATest and 3DGameTest.

I have a theory about the Packard Bell 560 hangup – I tried this benchmark on one other system and got the same result – CAPSLOCK can be toggled but the program appears to be not responding. Both systems have a few things in common, but I think only one is the real issue as the “CPU speed detect” is the hang-point. It seems that there is a problem detecting newer “generation” processors that are underclocked – It is probably particular to the Pentium line as it was so long lived and there were numerous changes and die shrinks, etc.

Let me explain what I mean better: I have a Pentium Socket 5 motherboard that can be set to 75, 90 and 100 MHz processor frequencies. The bus speed can *only* be 50 or 60 MHz, so the 100 MHz setting is NOT 66 MHz x 1.5 (as is more common) but 50 MHz x2. Due to the fact than newer 133 MHz and faster Pentiums run cooler (newer generation die shrink but still compatible), I use a 133 MHz Pentium run underclocked at 90 MHz/100MHz because it runs cool enough that I need no fan on my CPU. It seems like this “arrangement” is causing a detection problem because the other system is also run this way (with a 133 MHz Pentium underclocked to100 MHz as a 50 MHz bus x2). I surmise that because a Socket 7 Processor is being used in a Socket 5 motherboard and is underclocked that there must be some detection problem with its frequency. I am only surmising here, I have no idea how this program measures speed, but after reading other posts in this thread I recall reading that if the CPU speed cannot be detected thru the “primary method” that it resorts to using an “inaccurate” secondary detection method. I don’t really see how this could cause a hangup but its really the only thing that is related in common in these two systems.

I could understand if this was a Pentium MMX run a 3.3v in a Socket 5 Motherbaord – that is something it is not designed for but will sometimes work if the MMX can tolerate the higher voltage, but that is not the case in this instance. I would have to switch out the 133 MHz Pentium and put in a real 90 MHz/100 MHz part and try it then. If it worked, that would lend at least some credence to the theory.

Sorry I’m rambling here – I just don’t see anything else that is unique about these two systems that could cause identical detection failure - No one else seems to have reported a similar problem either so I’m left guessing.

-Isaac
 
You must log in or register to reply here.
Back
Top