Sign in with
Sign up | Sign in
Your question

Every chip has glitches.

Last response: in CPUs
Share
October 5, 2001 1:48:52 PM

http://www.theinquirer.net/04100109.htm

The next time someone says "intel is 100% stable" I hope everyone realizes that NO SYSTEM is 100% stable, and small rare glitches cannot be used to degrade or defame a chip,chipset, or operating system. Unless you admit that every os has glitches and nothing is immune, not intel, not amd, not linux, not windows. Nothing.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~

More about : chip glitches

October 5, 2001 1:55:32 PM

why not?
My little p60 firewall/router has been running 24/7 solid for a few months now, I'd call that stable.


Next time you wave - use all your fingers
October 5, 2001 1:57:41 PM

This was more for the people who claim AMD sucks because it is not stable, when intel has issues as well.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
Related resources
October 5, 2001 4:52:22 PM

You know burger, you really have a post count racket going with those clickable posts heh.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 5, 2001 5:38:02 PM

Not really. I get maybe one a day or so. That's just a drop in the bucket for me. I'd be more worried about you replying to every post in your threads :) 

<font color=green>I post so you don't have to!
9/11 - RIP</font color=green>
October 5, 2001 6:16:48 PM

and glitches have VIA chipsets

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 5, 2001 6:24:37 PM

Heh, I so hope that the kt266a is a better chip.

PS: Burger, I have never replied to my own post, what are you talking about?

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 5, 2001 6:39:44 PM

it <i>still</i> has a VIA southbridge! just a newer link

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 5, 2001 6:40:16 PM

Quote:
replying to every post in your threads


<font color=green>I post so you don't have to!
9/11 - RIP</font color=green>
October 5, 2001 7:40:50 PM

That is a software driver glitch. It has nothing to do with any hardware. Additionally, it is fixed in the latest version of said software.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
a b à CPUs
October 5, 2001 8:11:13 PM

The only stability problems I have had with an AMD platform that have not occured on Intel platforms have all been related to VIA chipsets. I don't own any VIA/Intel platforms and refuse to buy them for my customers. I do about 1 computer every two weeks for a local shop that can't contain certain VIA problems. While most of these are VIA/AMD systems, the last one I did was a VIA/Intel system. That says a lot about VIA.

Back to you Tom...
October 5, 2001 8:20:47 PM

that's a driver/software issue, stupid! learn to read, monkey!

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
October 6, 2001 12:35:33 AM

My point ignorant, was that glitches can occur no matter what the cpu running the system, and blaming amd for software glitches is as ignorant as blaming p4 for this glitch.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 6, 2001 3:40:11 AM

Opps already brought up in another thread...sorry.

Video editing?? Ha, I don't even own a camera!<P ID="edit"><FONT SIZE=-1><EM>Edited by Ncogneto on 10/05/01 11:45 PM.</EM></FONT></P>
October 6, 2001 4:52:30 AM

yes, its a software issue.
software is written to patch up the glitches in the hardware.

btw do you have a VIA southbridge that posts more than 15 MB/sec on BM-IDE??? I tried every combination of BIOS and VIA 4in1 drivers on a fresh windows install and I couldnt get past even 15. the same hard disk posts as much as 88 on the Intel board, without setting up anything else!

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 6, 2001 8:36:09 AM

Girish, no harddrive can transfer data at 88Mb, so the intel board has to be wrong.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 6, 2001 10:41:04 AM

Even I am surprised at that, because the drive was a old Seagate U5 series 8.4 GB UATA/66!!! I used BCM Diagnostics to test the disk performance. The VIA system was installed with the latest version of 4in1 drivers and its IDE update while the Intel system was D815EEA2 with Intel UltraATA storage driver.

Actually most ATA/66 drives are capable of 100 MB/s as some report said, and that 8.4 GB was one of them I guess. The VIA system dint go past 15 MB/s, while the Intel system instantly got it upto 88 MB/s average, top being 91 something and least being 84.

This speed difference was apparent in boot-up. Alongwith Intel's RapidBoot, it booted (fresh install) in 27 seconds flat, while Windows took about 22 of those. The VIA system took 48 seconds in which Windows took about 37! this 15 seconds difference shows up the difference in Hard disk speed. I marked the moment that showed "Starting Windows 98.." for Windows bootup time.

The VIA system was a Duron 600 MHz on a Luckystar K7VAT KT133 board while the D815EEA2 had a P-III 600E MHz, almost comparable give or take 10%. We all know that OS bootup process is largely disk-intensive so there is little room for the processor speed to take part, yet, a difference of 15 seconds is huge.

I guess I should again test the systems with newer veriosn of the software of some other software like SiSoft Sandra.

Will keep you posted guys, no gals in this forum yet I guess.

girish

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 6, 2001 11:35:48 AM

88MB/sec is impossible for any single drive, it is deffinatly a misreading.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 6, 2001 12:55:25 PM

yes, maybe thats why i decided to retest.

but then, Intel boards boot better than VIA boards...for what reason? may not be 88, I had expected it to be around 45~50, 88 was too high to believe. But a bootup time of 27 seconds made me believe that what i was seeing was right!

I know no drive can perform even closer to its rated speed, so why was ATA/100 invented over ATA/66? intel board had ATA/100 controller in its ICH2 so it did perform at ATA/100 giving more than 66 MB/s.

girish

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 6, 2001 10:23:17 PM

Just because a drive is ata100 does not mean it will use any where near that bandwidth, go to www.storagereview.com for an explination on how drives work.

For your older drive 15mb/sec actually sounds spot on where it should be. 88mb/sec is WRONG. The fastest ide drives available right now in the world only transmit at BARELY 40mb/sec, but closer to 35mb/sec average.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
Anonymous
a b à CPUs
October 7, 2001 5:08:57 AM

Maxtor and Fujitsu claim media transfer speeds in excess of 50MB/s. As with all claims those need to be verified in the real world first.

Now anything, including a bad cable may hamper performance on those new ATA66/100 drives. Let's not forget IDE is not a terminated bus so the lenth of the cable IS an issue here. With most outlets selling 80 conductor IDE cables 18 inches or longer, poor performance is to be expected. The IDE standard calls to maximum lenghts of only 16 inches if I recall this exactly (someone please correct me if I'm mistaken).

Just my two cents...

Chers!

Bogdan

***Life is Beautiful***
October 7, 2001 7:45:33 AM

www.storagreview.com

the fastest drive in the world just broke 42mb/sec if I recall(ide not scsi) They can claim what they want, but it is theoretical performance.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 8, 2001 5:40:11 AM

well if the same software is giving vastly different results on different systems what do you make out?

it was a buffered read, which can be deceptive. but if it does reflect in performance, i take it.

i recently clocked a ATA/66 drive on an intel motherboard with SiSoft Sandra 2001te free edition, and it gave a speed of 63 MB read and 61 MB write. I am testing a couple of ATA/100 controllers (connected to ATA/66 drives) and maybe a VIA platform today. will keep you posted.

girish

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 8, 2001 11:52:20 AM

Girish - I'm with Matisaro on this one, it is just not physically possible to get that much read, let alone write, performance from a drive - your measurement tool - whatever it is, is wrong. You can verify this with NT or 2000 and use the physical disk stats in perfmon, which will tell you the same thing that we are.

The ATA interface specifications allow for faster devices to transmit together and maximum transfer rate of cache. Since disk cache for IDE is pretty much never over 2Mb, you cannot get an extended read or write of more than approx. 40Mb/s - this is near the platter limit for a disk. The SCSI drives are faster, but again, they are in everything aren't they?

A typical IDE drive, well optimised, might transfer 30-35 Mb/s.

-* This Space For Rent *-
email for application details
Anonymous
a b à CPUs
October 8, 2001 1:56:14 PM

He must be looking at the buffered read. The true speed of your HD is the unbufered speeds.

-Mike
October 8, 2001 5:37:24 PM

Quote:
it was a buffered read, which can be deceptive.


I agree with 1200mike. There's no way you can use a buffered read speed as your results. It has to be SUSTAINED, which means more than the .005 seconds it's taking to use up the 512k cache (guess) on your 8.4gig drive.

<font color=green>I post so you don't have to!
9/11 - RIP</font color=green>
October 8, 2001 6:32:58 PM

well, I had already said it was a buffered read speed. but then, if it is giving results then why not take it?

my 8.4 G drive does have exactly 512k cache, and newer drives now come equipped with caches as large as 2MB! if use of such caches improves the performance of the drives then there is really no need of knowing the raw speed of the disk. it might be a MB to 10 MB or even lesser.

This drive which was a ATA/66 compliant posts 88 MB when used on a ATA/100 controller, 61 on a ATA/66 controller, of course in buffered mode sustained read, <i>AND</i> it reflects in the results that it boots much faster and the machine runs well. Well, the performance might drop to half of it in buffered random read, that might reflect real world performance.

All use the L1 and L2 cache to improve the performance and advertise it as a performance enhancing feature. If you dont consider the disk cache, then you dont need to consider the CPU cache. all tests should be done with both L1 and L2 cache disabled. That will give the *raw* speed of the processor! So one processor wont have an advantage over the other because it has larger L1 and/or L2 cache.

Output is what really matters, no matter how you get it!

girish

<font color=red>No system is fool-proof. Fools are Ingenious!</font color=red>
October 8, 2001 6:50:26 PM

Becuase speed for .005 seconds doesn't matter in real life. Sustained speed makes much more of a difference. Unless all you do is read the same 512k file over and over and over and over and over again. Then you would actually be getting 88MB/s.
Wait...is that 88 megaBYTES, or 88 megaBITS?
If you're measuring in bits, that would solve the whole problem.

<font color=green>I post so you don't have to!
9/11 - RIP</font color=green>
October 9, 2001 6:20:59 AM

How many programs do you know which are smaller than 512k?

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
!