021005-2 - Cyrix and Lacuna Part 2
Episode 2: The Empire, not intel but "AMD", Strikes Back
I forgot to mention that I found et9603.exe
contained cache driver for OS/2 after I wrote comment to George Albus. George
wrote me it didn't improve DISK I/O marks. I had a good result with Cyrix with
DOS and Win9x. So I wanted to know if Cyrix is really not good for Warp...
Pulled out CD-ROM-II and disassembled it. Connected DC power and tested it.
Caddy went down inward to the drive. Next, connected SCSI cable and powered it
up. No go. Looked into SCSI pins and found one pin bent down. I don't know if
it was the reason of the trouble but after I straighten the pin the CD ROM-II
became alive.
Time to install Warp... but I completely forgot install procedure for
coexistence of DOS and OS/2. Well it was not the point. May be setting for boot
manager was wrong but installation completed somehow. All I wanted is running
SYSBENCH on OS/2 for Cyrix and DX4.
Installation completed and WARP 3 came up. Set resolution to 800x600 65536C.
Installed SysBench 0.9.4g.
System Specs
9577-BTG 33MHz planar
Hard Drive: EIDE Quantum Fireball ST, 2160AT
CD-ROM: Toshiba XM3401BMA 2.1x speed off from IBM SCSI II (OEM'ed FD MCS-700)
RAM: 8MB x4
Resolution: 800x600, 65K color depth
OS: OS/2 Warp 3 Connect, NO FIX Pack applied yet.
File system: FAT
Results
Test result of SysBench 0.9.4g.
CPU |
CPU Integer |
CPU Float |
DIVE |
PM |
Disk I/O |
CD I/O |
Intel 486DX4-100 |
44.159 |
4.809 |
10.225 |
21.411 |
15.999 |
17.304 |
Cyrix 5x86 3x *1 |
45.262 |
6.088 |
10.136 |
14.230 |
9.056 |
17.284 |
Cyrix 5x86 3x ET *2 |
45.159 |
6.066 |
10.116 |
14.232 |
8.778 |
17.285 |
IBM 5x86 3x *1 |
45.138 |
6.060 |
10.127 |
14.231 |
9.440 |
17.285 |
IBM 5x86 3x ET *3 |
49.007 |
5.816 |
10.094 |
13.093 |
6.341 |
17.268 |
Cyrix 5x86 4x ET *3 |
51.821 |
7.756 |
10.140 |
14.270 |
9.355 |
17.291 |
Cyrix 5c86 4x ET *4 |
62.924 |
7.377 |
10.120 |
13.168 |
6.478 |
17.269 |
Cyrix 5x86 4x ET *5 |
62.986 |
7.471 |
10.111 |
13.154 |
6.456 |
17.270 |
AMD 5x86 4x QFP *6 |
50.331 |
6.085 |
10.212 |
21.648 |
15.300 |
17.308 |
AMD 5x86 4x PL586 *7 |
65.945 |
5.492 |
10.249 |
21.834 |
17.263 |
17.312 |
Sorry, I can't locate appropriate 3x interposers for Lacuna socket except
Evergreen 586 at this moment.
ET means cache driver et586OS2.sys from
et9603.exe is used.
EverGreen 586 Interposer is used for *1, *2, *3, *4, *5. CB/on.
PowerLeap PL586 Interposer is used for *7. WB/on.
*1 Cache driver not used
*2 Cache driver settings: default (/CCR4=1D)
*3 Cache driver settings: default (/CCR4=1D)
*4 Cache driver settings: /CCR4=1D /WBE
*5 Cache driver settings: /CCR2=56 /CCR4=1D /WBE (worked fine on 5551-R and Win98)
*6 EverGreen amd5x86/QFP Upgrade, WB/ON
*7 AMD 5x86/PGA on PowerLeap PL586 on an 486 socket interposer (like the one for PODP)
Cyrix 5x86 used for *3 through *4 is 133MHz unit. Can't find my favorite x4 100MHz beast.
Cyrix 5x86 *2 is x3 100GP.
HAS-33T in the
Interposer page does not fit to the socket. Interposer's PCB hits the Power
connector from the PSU.
PowerLeap PL586
also does not fit to the socket. SW block under the PCB hits capacitors inside
the socket. It needs another socket.
PGA DX4 ZIF
lever will hit overhang of the interposer for voltage regulator!
Impression
No, It shouldn't be... But Fact is fact. I have not yet tested other SW
settings, but so far it seems that Cyrix 5x86 is not good at DISK I/O and Video
test on OS/2 (and on Lacuna). CPU integer and Floating point hit really a good
score but PM and Disk I/O are really slower than DX4-100. AMD is better than
Cyrix on Video and Disk I/O.
I shall seek and find other register setting... not now. but in SOMEDAY (as
a matter of fact, I should have to write "tomorrow" in stead of "someday").
Monologue
I still can't believe the result! Wondering if Sysbench was programed to use
special instructions (or function?) of Intel CPU. Actual performance of Cyrix
does not seem so bad... Next time I'll try with x4 capable 100MHz unit.
|