4-5 crashes in 24 hours normal??

[aaroncgi]

I was reading Anandtech's recent KT133 board roundup article today, and was shocked to see this. He said that now the KT133 boards are better than they were three months ago, when 6-7 crashes was normal?? And this from a chipset that is supposed to be better than KX133, particularly in regards to stability?

I just found this rather odd. While I never have my machine up for 24 hours (so far maybe 12-16 max), it doesn't ever crash, and certainly not 4-5 or 6-7 times in 24 hours! If it did, I'd consider it unstable and unacceptable, I thought not crashing was the norm?

However, I am pretty sure Anand (and other online hardware review sites) run their memory at PC133, CAS2, while I run mine at PC100 and CAS3 (but in Turbo mode now). Granted, if I ran my memory bus at 133, I got lockups. I tried running at CAS2 (even at 100), and get an error before getting into Windows, and on bootup, the system suggests I use CAS3.

If 4-5 crashes/24 hours are what you gain by getting the last drop of performance out of your machine, doesn't seem worth it to me! Same can be said for memory, I'm not willing to spend 50-100% more for a 3-5% increase in speed, and by the looks of it, a decrease in stability.

Just some rambling thoughts as usual..

<A HREF="http://www.aaroncgilbert.com/computer/html">My non-crashing KX133 system</A>

Cheers,

Aaron


[This message has been edited by aaroncgi (edited 30 November 2000).]

[This message has been edited by aaroncgi (edited 30 November 2000).]

Ok, fixed the link, thanks Spazm_1999 !

[This message has been edited by aaroncgi (edited 30 November 2000).]

[spazm_1999]

There's an error in your link... you ramplaced the last . by a /
see... http://www.aaroncgilbert.com/computer/html

corected one

<A HREF="http://www.aaroncgilbert.com/computer.html">His non-crashing KX133 system</A>

Spazm

[The PIT]

Mmm that comment puts me off Via chipsets again. I would say if a machine crashes 6 or 7 times a day thats theres a hardware failure off somekind. Perhaps he's still using an engineering sample.

[Brian R.]

I'll bet the computers that have this problem were not set up correctly. It seems likely that these sites don't have the time to investigate the best combinations of HW/drivers/PCI slot configurations that many of us are willing to spend. Many people run these MBs without all these reboots.

IMHO

[aaroncgi]

I too would suspect the machines are not set up correctly if they are crashing this much. I've heard this about Anandtech before, and thus don't always put as much faith in his reviews compared to others (though I think his are quite detailed and generally very informative).

I just have to believe that a guy who runs a hardware site and tests dozens, maybe even hundreds of systems, has got to be better at setting them up than me! My KX133 was the first system I built from scratch, and sure it took me a few days to get it running right, thanks to ACPI and Win98, but now it's the most stable system I have owned to date.

I also remember, when I was still undecided about my video card, reading some of the problems experienced with Matrox G400 and VIA chipsets. Other than the ACPI problem during installation/setup, which was eventually solved, I have no complaints. I get 4X speed on the card, no lockups, looks great, and so far runs all the games I've thrown at it. Though that isn't many compared to most of you, probably.

Cheers,

Aaron

[Kosh Naranek]

Maybe it's not the computer that crashes but a particular piece of software that does the crashing every time ?

[SteveC]

Maybe it's Anand...?

------------------
Cheers,
Steve

"Life is what we make of it, yet most of us just fake"

[SteveC]

Oh, and yeah - comments like that re-affirm my choices when I always buy intel!

Kosh Naranek - yeah - Windows 98!

------------------
Cheers,
Steve

"Life is what we make of it, yet most of us just fake"

[Indiana]

My system is now running since more than 8 days without a single shutdown and heavy Web-browsing every day, often with both, IE and Netscape6, some DVD watching, some MJPEG capturing from TV with my Hauppauge and nearly uninterrupted WinAMP - and I did not get any crashes.
Granted this is on Win2K, with Win98 you can easily get more than 5 crashes a day, especially when using Netscape 4.7, or trying OGL demos with Matrox' ICD...


Ah, my system does have a VIA KT133 (ABit KT7) chipset and it's oc'ed as well (Duron800@1026), I do have the RAM set to 133 MHZ (144MHz to be accurate...) with CAS2.
So maybe Anand just isn't able to set-up a mainboard/OS correctly?

EDIT: I forgot: It's up "only" since 8 days because I had to reboot into Win98 for getting digital-out with a DVD and not because of a crash.
When will Creative FINALLY fix their Win2k drivers??!? GRRR

[This message has been edited by Indiana (edited 30 November 2000).]

[aaroncgi]

I'm certainly open to the idea that it could be Anand's problem. I do consider myself a novice when it comes to building computer systems, have really only totally built this one.

And after all, I/we do have one advantage that Anand probably doesn't use - MURC! Without it, I would have had a much harder time getting mine working right..

Cheers,

Aaron

[Gurm]

This is not the first time Anand has utterly failed to find the correct driver revisions.

But it's also partly Via's fault.

The 4-in-1 drivers go through worse cycles than nvidia's detonators.

The 4.23 4-in-1 was the buggiest POS I've used in years. The current 4.24 is nice and stable. What do they keep changing? I don't know. But if Anand ran his test on a day when a bad 4-in-1 build was the "most current", I could see why it would crash a lot.

I have faith that Via can, if they want to, make quality stuff. They just seem to not want to some days.

*shrug* I'm still using Intel, but I've built a few Via machines lately. The 133a and the 694 are good chipsets, AFAIK.

- Gurm

------------------
Listen up, you primitive screwheads! See this? This is my BOOMSTICK! Etc. etc.

[DentyCracker]

The 25a seem stabler than 24, which were the first to allow me to burn without fear with DMA enabled on the Plextor

------------------
1° System
Inwin Q500 Case
Abit KT7 Motherboard
AMD Duron 700@850 MHz.
384MB PC100 Cas 2 PC100 SDRAM.
Maxtor 20 GB 5400 RPM UATA 66 HD
Quantum Fireball 4.3 GB UATA 33 HD
Plextor PX-W8432 burner
Toshiba SD-M1402 DVD ROM drive
Creative Labs SoundBlaster Live! X-Gamer 5.1
Matrox G400 DH 32MB
Mag Innovision DX17T
Klipsch ProMedia v2.400 Speakers
ATI TV Wonder*sigh*
Realtek 8039AS based 10Mbps NIC
Running WinME, Win2K SP1, BeOS 5.03

2° System
Generic Finger Slicer AT Case
FIC VA503+
IBM Cyrix PR 266 *brrr* (long story)
128 MB Generic PC100 SDRAM
Maxtor 10 GB 7200 PRM UDMA33 HD
Creative Labs Ensonic AudioPCI64 (es1371)
Matrox Millennium G200 AGP
Realtek 8019 ISA NIC
Win98 2222A
Win2k Pro SP1

3° System

PCChips M571
Intel Pentium MMX 233
128 MB EDO DRAM (SIMMS)
IBM 10 GB HD
Diamond Stealth S3 trio 64
Realtek 8019 ISA NIC

[paulcs]

My favorite thing about the 4in1's in general is the way VIA renames the busmastering drivers with nearly every release.

The "typo" which misidentified the 4.03e AGP driver as ver. 4.04e was entertaining. And now, like Intel chipsets, all hard drives (since 4in1 4.23) are identified as Generic IDE Type 47 instead of by name.

I just have this feeling that Wen Chi Chen's halfwitted son-in-law is running that part of the show, because they seem to be doing most things well lately.

Paul
paulcs@flashcom.net

[paulcs]

Aaron, there's nothing "normal" about Anand & Co. getting 4 or 5 crashes in 24 hours. This is a "use vs. abuse" situation. These crashes happen during the course of their extended stability tests, which are designed to make most boards crash (including BX boards).

I'm under the impression that this has little to do with VIA or Intel or anything most of us would do to our machines. It tends to be somewhat independant of chipset, and variances are typically from manufacturer to manufacturer. Boards from AOpen, Asus, the newer Abits, and MSI seem to do best with Anand's stability test and the reviewer claimed several of the boards were extremely stable.

Paul
paulcs@flashcom.net

[aaroncgi]

Paulcs,

Thanks for the tips regarding Anand's stress testing. I guess it makes sense, and I suppose a much higher percentage of people who read articles at Anandtech, or even know what the heck it is, are into overclocking and tweaking. I get the impression, especially with his motherboard reviews, that if a board doesn't support overclocking or some kind of tweaking, it doesn't matter what else it does right. Certainly a valid viewpoint, but without reading between the lines, one could get the idea that none of the boards or systems he puts together with them are that good.. And yet as you say, he claims they are extremely stable.

Anyhow! I've also noticed the oddities with the VIA chipset, including the generic HDs and such. Also had a problem installing the AGP portion of the 4-in-1 4.24. I'd run the installation program and everything would go by the book from all outward appearances. Then when I'd boot up again, Windows would report that viagart.vxd was missing, thus my system would only use AGP 1X. Was running my system for two months before I figured out where viagart.vxd was supposed to be, and simply copied the darn thing there, now I get 4X rock solid. Since I haven't had experience with any other motherboard chipsets, I can't say if my troubles are any better or worse than others. My dad's put together quite a few systems, however, as far as I know exclusively with K6-2s, which if I'm not mistaken will use either the ALI or VIA chipsets, and he's never complained..

Then again, he's retired, so he has all day to work on these things..

Cheers!

Aaron