• Please review our updated Terms and Rules here

Looking for volunteers to help test a new benchmark

Uh-oh...just got this on one of the boxes here:
Code: 
For best results avoiding jitter, run this stub from a freshly-booted system 
without any TSRs loaded.  Contact trixter@oldskool.org in case of trouble.

Waiting for floppy drive spindown... done.
Benchmarking your system:  Memory speed... CPU opcode exercises... 
video memory speed... effective addressing and instruction fetch... 
3-D game opcode distribution... Calculating your system's TOPBENCH score... 
Done.  Identifying your system:  CPU... CPU speed... BIOS... 
Machine Model... Date... CRC... Video subsystem... Video adapter... Runtime error 201 at 0342:033A.

OUTPUT.INI reads thus:
Code: 
[UID9B2216A1A5]
MemoryTest=8
OpcodeTest=16
VidramTest=97
MemEATest=18
3DGameTest=17
Score=444
CPU=VIA Cyrix 6x86MX-S (clock x3.0 mode)
CPUspeed=200 MHz
BIOSinfo=war,Copyright (C) 1984-97, Award Software, Inc.

The video card is an STB Lightspeed 128 with 4MB installed. It's an ET6000 card.
 
As previously mentioned I had a lot of issues getting this to do anything but crash on my tandy 1k... from DOS 3.3 it does indeed appear to run in FAST mode...

Tandy 1000 HX, FAST (7.16mhz), V20
[UID7214448]
MemoryTest=2033
OpcodeTest=1232
VidramTest=1262
MemEATest=1599
3DGameTest=1138
Score=7
CPU=NEC V20
CPUspeed=7.16 MHz
BIOSinfo=Copyright (C) 1984,1985,1986,1987 (06/01/87, rev. 0)
MachineModel=0000
BIOSdate=19870601
BIOSCRC16=7214
VideoSystem=CGA
VideoAdapter=Tandy 1000
Machine=Tandy 1000

When dropped to slow (4.77mhz) that same machine gets this far:
[UID7214743]
MemoryTest=2584
OpcodeTest=1595
VidramTest=1521
MemEATest=1977
3DGameTes

and then bombs with RUNTIME ERROR 200 at 0354:02A2

switched back and forth between speeds multiple times, it hates 'slow' on a V20 equipped HX... so I dragged out my 8088-2 equipped HX, and DOS 2.11 or Dos 3.3 doesn't matter, as soon as I start the .exe the screen goes blank, the speaker comes on beeping full blast, and the only thing I can do is kill the power to the machine.

For laughs I ran it on my "midling" K6/2-450 with the Voodoo 5 in it.
[UID1F5F156822]
MemoryTest=4
OpcodeTest=8
VidramTest=80
MemEATest=5
3DGameTest=5
Score=547
CPU=AMD K6-2
CPUspeed=450 MHz
BIOSinfo=war,Copyright (C) 1984-99, Award Software, Inc. (01/03/00, rev. 0)
MachineModel=01FC
BIOSdate=20000103
BIOSCRC16=1F5F
VideoSystem=VGA
VideoAdapter=VGA, unknown Chipset, VESA, 256kb Video Memory (BIOS)
Machine=AT clone

If I have time tomorrow I might grab one of the 1000 EX I have down in the garage and see what it does.... I'll also test my sharp PC-7000.

Runtime Error 200... overlay files... so this is in Turbo Pascal? Interesting it's getting a division by zero error; just what are you doing?

Hell, i'm wondering what on earth you need a ovl file for... the overhead of the overlay unit, overlay itself and code ties subtracted from the .exe ends up smaller than 64k... are you up to sharing the code, maybe we could cut your size down a bit.
 
Last edited:
Compaq Deskpro 286 said:
[UID22DD479C3]
MemoryTest=623
OpcodeTest=268
VidramTest=800
MemEATest=346
3DGameTest=264
Score=22
CPU=Intel 80286
CPUspeed=12 MHz
BIOSinfo=\JSF'PNA'CHL'5CRJ (C)Copyright COMPAQ Computer Corporation 1982,83,84,85 (01/28/88, rev. 0)
MachineModel=01FC
BIOSdate=19880128
BIOSCRC16=22DD
VideoSystem=CGA
VideoAdapter=CGA
Machine=AT clone
Click to expand...
The Compaq ate the 3½" floppy after the test was done. Twice. Switched to the 5¼" drive, and it was fine. Might just be a cheap floppy drive going bad, though.
Canon Notejet III said:
[UIDA1582F57]
MemoryTest=72
OpcodeTest=37
VidramTest=195
MemEATest=30
3DGameTest=25
Score=203
CPU=Intel Pentium
CPUspeed=185 MHz
BIOSinfo=Copyright (C) 1985-1994 Phoenix Technologies Ltd. (04/19/90, rev. 0)
MachineModel=01FC
BIOSdate=19900419
BIOSCRC16=A158
VideoSystem=VGA
VideoAdapter=VGA, VESA, 256kb Video Memory (BIOS)
Machine=AT clone
Click to expand...
Home-built K6-2 said:
[UIDEC1C31343]
MemoryTest=12
OpcodeTest=13
VidramTest=64
MemEATest=11
3DGameTest=6
Score=673
CPU=AMD K6-2
CPUspeed=350 MHz
BIOSinfo=war,Copyright (C) 1984-98, Award Software, Inc. (03/24/99, rev. 0)
MachineModel=01FC
BIOSdate=19990324
BIOSCRC16=EC1C
VideoSystem=VGA
VideoAdapter=VGA, Tseng ET , DRAM, unknown Chipset, VESA, 256kb Video Memory
Machine=AT clone
Click to expand...
 
Chuck(G) said:
Uh-oh...just got this on one of the boxes here:
Code: 
Runtime error 201 at 0342:033A.
Click to expand...

Thanks for the detailed info, which allowed me to reproduce this. I'll debug the code tonight and put out a new version.

For Jason, I'll also include a version that doesn't use overlays and contains copious debug output, to see if we can isolate why it doesn't work on his HX.
 
Toshiba Satellite 105CS laptop:

Code: 
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.95
;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.

[UID35EE6CE19]
MemoryTest=62
OpcodeTest=22
VidramTest=216
MemEATest=11
3DGameTest=16
Score=169
CPU=Intel Pentium
CPUspeed=75 MHz
BIOSinfo=(c) COPYRIGHT 1990,95 Toshiba Corporation.ALL RIGHTS RESERVED.   (12/18Á95, rev. 0)
MachineModel=01FC
BIOSdate=19951218
BIOSCRC16=35EE
VideoSystem=VGA
VideoAdapter=VGA, VESA, 256kb Video Memory (BIOS)
Machine=AT clone
 
deathshadow said:
Hell, i'm wondering what on earth you need a ovl file for... the overhead of the overlay unit, overlay itself and code ties subtracted from the .exe ends up smaller than 64k... are you up to sharing the code, maybe we could cut your size down a bit.
Click to expand...

The program requirements without an overlay are 88KB memory; with overlay, 54KB. The goal of the overlay was to be able to run on a stock PCjr, which, after subtracting 16K for video ram (text mode) and another 20K for DOS 2.1, leaves about 92K free. The largest chunk of code is 19K which is the CPU detection library (99% of which is not my code) and the SVGA card detection library (also not my code) and since they run only once it didn't make sense to force them both to take up memory at the same time.

And, surprise, those libraries are where the errors are, so I'll have to dive into them more tonight. All other code (non-VGA card detection, machine detection, etc.) is mine and -- not to break my arm patting myself on the back or anything -- works perfectly.

I will put out another stub package tonight, containing two versions (one using overlays and another that does not and also contains copious debug info).
 
Tandy 1000 EX said:
[UID989024CA]
MemoryTest=3323
OpcodeTest=1844
VidramTest=1863
MemEATest=2196
3DGameTest=1844
Score=4
CPU=Intel 8088
CPUspeed=7.16 MHz
BIOSinfo=Copyright (C) 1984,1985,1986 (07/14/86, rev. 0)
MachineModel=0000
BIOSdate=19860714
BIOSCRC16=9890
VideoSystem=CGA
VideoAdapter=Tandy 1000
Machine=Tandy 1000
Click to expand...
Unisys CWD said:
[UID61F51A854]
MemoryTest=76
OpcodeTest=56
VidramTest=125
MemEATest=81
3DGameTest=42
Score=215
CPU=AMD Enhanced Am486DX4
CPUspeed=100 MHz
BIOSinfo=,d Software Inc. Awar,Copyright (C) 1984-94, Award Software, Inc. (08/29/95, rev. 0)
MachineModel=01FC
BIOSdate=19950829
BIOSCRC16=61F5
VideoSystem=VGA
VideoAdapter=VGA, Cirrus CL-GD5424, VESA, 256kb Video Memory (BIOS)
Machine=AT clone
Click to expand...
IBM Thinkpad 755CD said:
[UIDACB316242A]
MemoryTest=70
OpcodeTest=44
VidramTest=171
MemEATest=66
3DGameTest=30
Score=187
CPU=Intel i486DX4
CPUspeed=100 MHz
BIOSinfo= COPR. IBM 1981, 1996 (11/07/96, rev. 0)
MachineModel=01FC
BIOSdate=19961107
BIOSCRC16=ACB3
VideoSystem=VGA
VideoAdapter=VGA, Paradise PVGA1A, 16-Bit-Video, 8-Bit-ROMfixed-sync, VESA, 7
Machine=AT clone
Click to expand...
Micron Trek 2 AGP said:
[UIDC2D4C3C]
MemoryTest=15
OpcodeTest=8
VidramTest=229
MemEATest=8
3DGameTest=9
Score=234
CPU=Intel Pentium II
CPUspeed=266 MHz
BIOSinfo=unknown
MachineModel=01FC
BIOSdate=19981223
BIOSCRC16=C2D4
VideoSystem=VGA
VideoAdapter=VGA, Trident GUI9420, VESA, 768kb Video Memory, 4096kb Video Mem
Machine=AT clone
Click to expand...

Recommendation: don't put the software in zip files. They're difficult to open on older operating systems.
 
Some requests :

IBM PC Model 5150 w/8088 & V20
IBM PC/XT or Portable Models 5155 or 5160 w/8088 & V20
IBM PCjr. Model 4860 w/8088
IBM PC AT Model 5170 @ 6MHz and 8MHz
IBM PC/XT 286 Model 5162
IBM PC Convertible Model 5140
Tandy 1000 SX or EX @ 4.77MHz 8088
Tandy 1000 TX/TL/TL/2/TL/3
Tandy 1000 SL/SL/2 w/ 8088 & V30
Tandy 1000 RL w/8088 & V30
IBM PS/2 Model 25 or 30 w/8086, V30 and 80286
 
Great Hierophant said:
Some requests :

...
IBM PC Convertible Model 5140
...
IBM PS/2 Model 25 or 30 w/8086, V30 and 80286
Click to expand...

I'm still trying to generate a 720Kb diskette for the Convertible (with a 80C88 CPU)...

The 8086-based Model 25 / Model 30 will be interesting, because I have Intel 8086-8MHz CPUs, with a 1978 copyright at latest. Intel 8088 CPUs with 1978 copyrights have a published bug that was later fixed with a release also marked with a 1981 copyright. That would seem to mean that the same bug was not present on the Intel 8086.
 
New version of the stub is now available at http://dosbenchmark.wordpress.com/downloads/ . It contains the regular stub that runs in 53K RAM, and also contains a "debug" version of the stub called "dstub" that produces more output during dodgy sections and doesn't attempt to overlay code. It requires 90K free RAM, so 128K systems might be challenged. Either stub produces identical output.ini results, so you can run whichever one you want. However, uncompressing everything no longer fits on a 360K disk, so it might be best to skip extracting "dstub.map" from the .zip file unless you really need it. (The full TOPBENCH tool *will* fit onto a 360K disk including its database with room for more entries, so no worries there.)
  • Chuck: I found the bug that was affecting you and fixed it. The code was in a library (not mine!) that attempted to write a word value to a byte-sized port. Please re-run if you could, although I have reproduced both the bug and the fix on my end.
  • Jason: I have no idea what is going on with your systems, especially the HX where it freaks the hell out, but try running dstub and see if that helps. There was another Tandy 1000 HX that tested fine, so I'm curious what yours does with the new code.
For everyone who contributed results in the last 36 hours -- thank you! It will be nice to ship the first version of the benchmark with so many systems already logged in the database.

For the assembler-inclined: The debug stub also comes with the full linker map file in case it craps out anyway and you'd like to see exactly what died and where; symbolic info is also attached to the .exe so that variable/procedure/symbol names show up in Turbo Debugger if you are so curious.
 
Tiberian Fiend said:
Tracking it down and getting it on my computer for one-time use is the problem, especially for files that aren't large enough or numerous enough to warrant being zipped.
Click to expand...
Huh, I've always kind of thought of an unzip utility as a vital utility for DOS use, myself...I hardly have a machine I don't have it installed on.
 
IBMMuseum said:
I'm still trying to generate a 720Kb diskette for the Convertible (with a 80C88 CPU)...

The 8086-based Model 25 / Model 30 will be interesting, because I have Intel 8086-8MHz CPUs, with a 1978 copyright at latest. Intel 8088 CPUs with 1978 copyrights have a published bug that was later fixed with a release also marked with a 1981 copyright. That would seem to mean that the same bug was not present on the Intel 8086.
Click to expand...

I agree that will be interesting, as my I-thought-it-was-correct-but-now-maybe-its-not 80C88 detection code might flag on it. If it gets it wrong, let me know.
 
That's not a QFP package, but a PLCC. It is possible to unsolder the old CPU and solder a PLCC socket in its place, but you should be skilled at SMT work. The V30 did come in a PLCC also, but you may have trouble locating one.
 
Okay, first off here's the Cyrix 686 box that had problems before:

Code: 
[UID9B22106C66]
MemoryTest=9
OpcodeTest=9
VidramTest=95
MemEATest=7
3DGameTest=5
Score=450
CPU=VIA Cyrix 6x86MX-S (clock x3.0 mode)
CPUspeed=225 MHz
BIOSinfo=war,Copyright (C) 1984-97, Award Software, Inc. (05/05/98, rev. 0)
MachineModel=01FC
BIOSdate=19980505
BIOSCRC16=9B22
VideoSystem=VGA
VideoAdapter=VGA, Tseng ET , DRAM, unknown Chipset, VESA, 256kb Video Memory 
Machine=AT clone

Here's a Supermicro server running dual 750MHz P3s:

Code: 
[UIDFCDC111A1B]
MemoryTest=3
OpcodeTest=5
VidramTest=120
MemEATest=4
3DGameTest=4
Score=399
CPU=Intel Pentium III E
CPUspeed=750 MHz
BIOSinfo=(C)1985-1996,American Megatrends Inc.,All Rights Reserved,6145F Northbelt Parkway,GA-30071,USA.(770)-246-8600. (01/15/99, rev. 0)
MachineModel=01FC
BIOSdate=19990115
BIOSCRC16=FCDC
VideoSystem=VGA
VideoAdapter=VGA, VESA, 256kb Video Memory (BIOS)
Machine=AT clone

And finally, here's a Compaq Deskpro ENP originally equipped with a 600MHz P3, but now running a 1.3GHz Celeron in a "slocket":

Code: 
[UIDB4B717D48B]
MemoryTest=1
OpcodeTest=5
VidramTest=110
MemEATest=6
3DGameTest=2
Score=432
CPU=Intel Pentium Pro (undistinguished)
CPUspeed=1295 MHz
BIOSinfo=\PJBMAPEMRPLGDRJAA1MARRJVRACC NP T    (C)Copyright COMPAQ Computer Corporation 1982,2000** (08/22/99, rev. 0)
MachineModel=01FC
BIOSdate=19990822
BIOSCRC16=B4B7
VideoSystem=VGA
VideoAdapter=VGA, VESA, 256kb Video Memory (BIOS)
Machine=AT clone

It would make testing a bit easier if you'd open OUTPUT.INI for "append" rather than "create". If I could add a one-line comment to each run, that would let me test more systems quickly.
 
You must log in or register to reply here.
Back
Top