Inconsistant Athlon result's; please read!

[Greebe]

Here's the biggest complaint I have... I've done alot of testing and have found inconsistant results by simply rebooting the machine!

Currently; MS-6167 mb, Athlon 500 Eclipse'd to 850 1:3 cache latency, 128 meg PC100 SDRAM set for cas2, 10/100 pci nic and an Ultra fast G400 Max!

example: using CPUmark 99
test1
800 L2 1:2 each test run twice
reboot1
66.5
reboot2
65.4
reboot3
70.4

test2
850 1:3
reboot1
70.4
reboot2
72

So as you can see there's obiviously a problem here! I'll add that I see this effect regardless of what benchmark progam or application I'm running!

------------------
Designer of the first Athlon mod. Currently marketing the "Eclipse" through AZZO Computers.
Anyone up for an Eclipse'd Athlon 850? Email me for more information

2836 mips, 1.107 gflops @ 850 mHz 1:3 latency
CPUmark99 71.6

[Jorden]

Okay Mike, tell me, since when does any PC run exactly the same benchmarks everytime??

Jord.

[Maggi]

Hey Greebe ...

must be your lame system ...



sorry dude, I'm just jealous ...



------------------
Cheers,
Maggi
________________________

Working Rig:
Asus P2B-DS @ 103MHz FSB
Double Pentium III-450 @ 464 MHz
4 x 128MB CAS2 SDRAM
Matrox Millennium G400 32MB DualHead
Eye-Q 777 (22" with 127kHz) on primary VGA
Nokia 445Xi (21") on secondary VGA


Home Rig:
Asus P2B-S Bios 1010 @ 100MHz FSB
Celeron 333A @ 500MHz
2 x 128MB CAS2 SDRAM
Matrox Millennium G400 32MB DualHead @ 150/200MHz
CTX VL710T (17")
and a brand new Pioneer 303S SCSI-DVD

[Greebe]

Well Jord when you execute one (for example CPUmark 99) and it consistantly reports the same proformance everytime it's run (possible exception being the first "run").
Then simply rebooting and immediatly rerunning the test (excluding the first run) and it scores repeatedly the same thing although higher or lower than previous reboots then there's a problem.

[Jorden]

Yes Mike, but that could be due to anything.
Heat, a single component in the PC, the videocard, the CPU, the PSU giving irregular current, you name it.

It could also be a bug in the software. And you give me an example of software without a bug?

Jorden.

[This message has been edited by Jorden (edited 14 December 1999).]

[MXR4LIFE]

I believe thats one of the reasons ZD upgraded to rev1.1 for CPU Mark99. If you run 1.0 it takes about 1min, 1.1 takes about 3min

[MXR4LIFE]

I just ran mine for another ref
K7M rev1.04 1007 BIOS
K7-500 w/ Eclipse set:
x7=705Mhz
1.7v
1/2 Cache
3x64Mb PC100
RAM timings by SPD (Default cas3)

ZD WinBench 99 rev 1.1
CPU Mark99

After 1 1/2 Hrs of surfing on net: 59.7
After fresh Boot #1 : 62.8
Reboot #2 : 63.2
Rerun after Reboot#2 : 63.0
Reboot #3 : 63.0
Rerun after Reboot#3 : 63.5

Complete range 3.8
Toss out first 0.7
Toss out H&L .2

Not too bad

I'll check mine at higher clocking later

[This message has been edited by MXR4LIFE (edited 14 December 1999).]

[Greebe]

MXR, CPUmark 99 ver 1.0 is supplied with WinBench 99 ver. 1.1!

[Greebe]

Gord'n, why do you think I started this thread in the first place?

[Jorden]

THE TRUTH IS OUT THERE

[MXR4LIFE]

I guess your right Mike, your adding the 1.0 to the CPU99 threw me. I cant find a ver number on any of my ZD WinBench2000 CD's or my WinBench99 copies. They call it CPUmark99 and it was updated during the WinBench99 ver1.1 from CPUmark32 that came with ver1.0
But since it is the first ver it must be 1.0

Heres a link from ZDBop

http://www.zdnet.com/zdbop/stories/m...387580,00.html

And it takes about 6 1/4 mins to run rather than the 3 I guessed at.

[This message has been edited by MXR4LIFE (edited 15 December 1999).]

[Scout255]

What programs are starting up when you turn it on?

[Greebe]

Nothing, even set it to standard vga... same effect!