What's a better stress-test program for memory than PRIME95?

[BonzaiDuck]

In other threads today -- even over the past week wearing my heart on my sleeve and pooping out a trail of tears over three forums in unbecoming panic -- I've finally discovered that a BSOD was just caused by an insufficient CPU_VTT (FSB) voltage setting. I apparently only had to kick it up a couple notches to "get what I want."

But if I continue to stress-test RAM with PRIME95 Blend-Test as I try and tweak a few more Mhz out of this sucker, I'll BSOD again, maybe again . . . . and again . . . .

I think that's bad. There must be a memory stress-test program that offers a margin of safety.

Does OCCT -- if that's what it's called -- does it do that?

Is it a good RAM-stress-test program?

Where can I get it? Thought I'd ask, but I'm going to start surfing tonight on my own. Maybe somebody here knows where I can find it . . . .

 

[Quiksilver]

This Here

 

[BonzaiDuck]

Thanks, Quiksilver.

Can you -- or anyone out there using a more recent nVidia chipset -- tell me if the IntelBurnTest ("Linpack"??) is compatible with the subject chipset family?

I remember TAT -- and the fact that it wouldn't run on an nVidia board.

Meanwhile, on the surface of things . . . . DA!! Love those Russian programmers, Comrades!! I have OCCT running. I got Kaspersky to "limit" it -- it popped up a message about "potentially dangerous program." Russians . . . . warning me about . . . . Russians.

Help me pra-teck mah precious bodily fluids!!

Anyone have experience with OCCT? EES EET SAFE??? EES EET SAFE???
IS IT SEFFF??!!! [Perhaps you'd like a little oil of cloves, for that tooth?] :laugh:

 

[error8]

Originally posted by: Quiksilver
This Here

Jesus, I'm topping 84 C with this program. It scared me and I stopped it before the test ended. What the hell is he doing?

 

[LOUISSSSS]

occt is a damn good stress test. it's turned up errors in 20 mins that would normally take me 12 hours in P95. if it passes its 2 hour ram/cpu test, it is also (to my experience) stable for 24 hours in any of p95's 3 tests.

 

[BonzaiDuck]

I ran OCCT for about 42 minutes before stopping it.

The "splash" screen -- with the Soviet-era workers with the hammers, which says "Loading . . . ." was still on there. I suspect this is because I told Kaspersky to "Limit" the program's operations.

That's why I am ask-ingk: EES EET SAFE?? Is it safe?

The designer is an artist, of sorts, and he has a sense of humor. I DO think it's a pretty good program. [And . . . . Nooo-ooo!! I'm "not all Georgians . . ." either . . . ]

Anyone know if "Serge" -- the Russky designer of that SnM stress-test program -- ever changed his mind and decided to update it for current-gen hardware? In a lot of ways, that program was "SchweeeeT."

EDIT: But . . . BUT! . . . What's the skinny on IntelBurnTest's compatibility with nVidia chipsets? Will it work with my 680i board?

 

[Cutthroat]

OCCT is safe, and a great app. When stress testing I run OCCT for 2 hours, it's never lasted longer than 40 minutes on an unstable system. And then I run Folding@Home for 24 hours to be sure.

 

[Cutthroat]

Ooops, double post.

 

[BonzaiDuck]

OK, then. I'll tell Comrade Kaspersky to let Comrade OCCT in the door, so to speak.

Still wondering if IntelBurnTest will even run on an nVidia 680i chipset.

My biggest gripe about PRIME95 Blend-Test is the BSODs I was getting for an insufficient CPU_VTT voltage setting. I only bumped it up to a reported value of 1.39V, I've left the Corsair DHX RAM settings all stock except for pushing the FSB up 20 Mhz and therefore putting the DDR = 810 Mhz.

Takes a licking, keeps on ticking. I'm almost "out of the woods" on my over-clocking ordeal. I think the termination voltage setting might let me go higher with this old 680i Striker board . . . .

Meanwhile, I see someone started a "post your results" thread for the E0 steppings -- E8600 and Q9650, I think. Stunning. IF . . . . they're reporting well-tested settings. A VCORE at 1.32V or less? . . . with an OC speed result of 4.1 Ghz?

Gotta save some ducats. . . . . I gotta have it!

 

[Cutthroat]

Originally posted by: BonzaiDuck
Still wondering if IntelBurnTest will even run on an nVidia 680i chipset.

You could try it and let us know...

Personally, it scares me. It make my cores go over 80C, but I pass.

 

[myocardia]

Use MemTest HCI for testing your RAM, and OCCT or the Intel Burn Test (which is just Intel's LinPack, with a GUI) to test your CPU/northbridge, etc.

 

[jmmtn4aj]

Originally posted by: Cutthroat
OCCT is safe, and a great app. When stress testing I run OCCT for 2 hours, it's never lasted longer than 40 minutes on an unstable system. And then I run Folding@Home for 24 hours to be sure.

I've had F@H go for days one settings that wouldn't make several minutes of OCCT :S

 

[betasub]

Duck, you've been watching too many classic movies.

Lots of ppl use OCCT - you've seen the xtremesystems thread that Quiksilver linked. Now perhaps you can ignore Kaspersky and tell us EES EET SAFE? :light:

 

[Extelleron]

The only test that matters to me anymore is Linpack 64-bit (Intel Burn Test is linpack with a GUI). Make sure you run the 64-bit version.

The linpack code generates more heat than any other app (10C+ over Prime95) and produces stability errors much faster. I've had Linpack show errors after 2 iterations (~6-7 minutes depending on the amount of memory you have) when Prime95 took ~8 hours to detect instability. If you're running a benchmarking system and want the highest frequency you can get, then it is not your app. But if you want a 100% stable system..... then Linpack 64-bit for hours is by far the best way to go.

 

[nerp]

I like to use Memtest86+ by booting a CD and letting it run overnight to rule out memory errors.

 

[BonzaiDuck]

Last time I ran it, the "OCCT Loading" screen remained running on the desktop after I stopped and closed the application. I thought maybe this was a result of choices I made in Kaspersky's warning dialog.

I'm running it again, this time for the two-hour RAM testing. Should the splash screen with the two "Reds" hammering the processors remain after I close it? Is this a programming flaw? Or should it disappear when I exit the red program screen?

 

[Quiksilver]

Originally posted by: error8

Originally posted by: Quiksilver
This Here

Jesus, I'm topping 84 C with this program. It scared me and I stopped it before the test ended. What the hell is he doing?

It's Linpack (Intel's stress testing program) with a GUI. It hammers the your cores way better than Prime95 or OCCT. On top of that it's way faster and bring your temp's through the roof; especially if you have a quad.

It's basically use it if you want an absolute stable overclock; even if you already pass prime95/occt

----
BonzaiDuck, as far as I know their are no problems using this an a nForce chipset.

 

[Idontcare]

Originally posted by: BonzaiDuck
Last time I ran it, the "OCCT Loading" screen remained running on the desktop after I stopped and closed the application. I thought maybe this was a result of choices I made in Kaspersky's warning dialog.

I'm running it again, this time for the two-hour RAM testing. Should the splash screen with the two "Reds" hammering the processors remain after I close it? Is this a programming flaw? Or should it disappear when I exit the red program screen?

Too late dude...your computer are belong to the mother russia and is currently crunching away as a botnet hacking into Georgia's three ATM machines and has already coerced them to spit out the equivalent of 6 USD.

The feds already have your MAC and IP address from the cyberattacks and are preparing to issue you with a strongly worded warning that you better get out of Georgia or else they won't let you play in the sandbox come Monday. (they just used the same memo GWB sent Putin and Medvedev, saving taxpayer dollars and all)

 

[BonzaiDuck]

[Snicker.] I was 5-years-old in front of the TV watching MacArthur give his "old soldiers never die" speech to congress. From Catholic grade school "newsletters" published by Cardinal Spellman, I feel brainwashed by the hype and hysteria. After I spent two months and 8 hours a day pouring over NARA-College Park and Library of Congress documents, catching up on my reading based on Clinton-era declassifications, I've been red with anger every since (no pun intended, but there's the heart, and the head, and the heart points toward the spectrum of . . . )

I guess this belongs in the political forum, though . . .

EDIT: but back to weird science. OCCT just popped up "Test Completed -- Stable!" I especially like those test-measurement graphs, and it reminds me of Serj's "SnM." Too bad -- he wrote me early last year that he wasn't going to update the program anymore.

You have to admire these people. There was a big spread in the Washington Post in the late '90s, with stories about these little-nobody people acting on their own and collecting scrap from junkyards to build things. One guy built a submarine he'd take out weekly to cruise the Volga; another guy, who'd lived most of his life on a collective farm, built his own airplane -- surviving a test-run crash and living to rebuild it.

EDIT: BACK TO WEIRD SCIENCE:

As long as I still have the last post of the moment, I'll just "edit."

The temperatures reported under OCCT are slightly higher than those reported under RealTemp with PRIME95 (although the room-ambient has increased a tad.) If only "slightly" higher, they must be accurate. I feel more comfortable running my AV suite while this thing runs -- for positive reasons: it seems to remain stable.

The 2-hour RAM test completed with flying colors, and I"m running the CPU-intensive 2-hour test now. I've already done a 9-hour small-FFTs on this rig at a range of settings that include the current ones. I'm just wondering how good a 2-hour OCCT stress-test might be, as opposed to a marathon of "PRIME95-ing."

[I'd still rather somebody else test IntelBurnTest/LinPack with an nVidia chipset, though. I feel your fear -- for those who said the intel program scared them. I can dig it -- really!

 

[error8]

Originally posted by: Quiksilver

Originally posted by: error8

Originally posted by: Quiksilver
This Here

Jesus, I'm topping 84 C with this program. It scared me and I stopped it before the test ended. What the hell is he doing?

It's Linpack (Intel's stress testing program) with a GUI. It hammers the your cores way better than Prime95 or OCCT. On top of that it's way faster and bring your temp's through the roof; especially if you have a quad.

It's basically use it if you want an absolute stable overclock; even if you already pass prime95/occt

Yeah, you're so right. This program is amazing! It detected instability at 4,01 ghz, that Prime95 used to find after couple of hours of stress test. Linpack found it in under 10 minutes!!!! But it got to 86 C, some 14 degrees until T junction. Dangerous stuff...

The thing is that I don't think there will be a real situation when my chip will ever touch those temperatures. But this is by far the fastest error scanner for an overclocked cpu.

Why isn't this program used and advertised all around the web instead of Prime, Orthos and OCCT?

 

[BonzaiDuck]

Originally posted by: error8

Originally posted by: Quiksilver

Originally posted by: error8

Originally posted by: Quiksilver
This Here

Jesus, I'm topping 84 C with this program. It scared me and I stopped it before the test ended. What the hell is he doing?

It's Linpack (Intel's stress testing program) with a GUI. It hammers the your cores way better than Prime95 or OCCT. On top of that it's way faster and bring your temp's through the roof; especially if you have a quad.

It's basically use it if you want an absolute stable overclock; even if you already pass prime95/occt

Yeah, you're so right. This program is amazing! It detected instability at 4,01 ghz, that Prime95 used to find after couple of hours of stress test. Linpack found it in under 10 minutes!!!! But it got to 86 C, some 14 degrees until T junction. Dangerous stuff...

The thing is that I don't think there will be a real situation when my chip will ever touch those temperatures. But this is by far the fastest error scanner for an overclocked cpu.

Why isn't this program used and advertised all around the web instead of Prime, Orthos and OCCT?

I've been an Intel CPU and chipset user for a good part of my journey with this technology, but I'm currently using nVidia chipsets. So, knowing that I can't run TAT on those chipsets, either someone has tried it, succeeded or failed, or I'm just not "go there."

The other thing -- I may have forgotten when I downloaded it. Does OCCT work reliably with 64-bit VISTA? or is there a special version of it for 64-bit? I think I forgot to distinguish, but I don't remember two versions posted.

 

[BonzaiDuck]

I sent the OCCT program's author an e-mail, asking about 64-bit support. It was VISTA compatible already in an earlier version, but it didn't specify anything about 64-bit OS's.

 

[GarfieldtheCat]

I'm running it on Vista 64-bit without problem.

 

[BonzaiDuck]

It runs as a 32-bit program in VISTA-64. This becomes obvious if you open Task Manager and look at actively running processes -- you'll see it with "*32" following the process name. I'm awaiting our Russian programmer friend's response on this matter.

I cannot see, on the surface of it, why a 32-bit program wouldn't stress the system every bit as much as a 64-bit version. And as I said, the reported temperatures are a tad higher than what I get in PRIME95's 64-bit version.

I suppose I might try IntelBurnTest on this nVidia system, but -- as some say -- those temperatures are pretty scary.

 

[myocardia]

Originally posted by: error8
Why isn't this program used and advertised all around the web instead of Prime, Orthos and OCCT?

1) It only works with Intel CPU's. 2) While LinPack has been out for a couple of years, it had no GUI, and most people couldn't figure out how to use it. The app that was linked by Quicksilver is LinPack with a GUI, but it's brand new, so people are just finding out about it.