Numerous RAM Errors

[Psykoikonov]

I think you just got a couple bad sticks of Ram, all indications lead to that conclusion. I have Gskill's, work perfect, Corsair same deal. As for heat spreaders they're all show.

 

[MisterEd]

I don't think it's the CPU. The fact the error count went down as the voltage went up proves that the memory was marginal, and you really need to try some known good non-overclocked memory chips and quit messing around with junk from G.Skill, Corsair, or any company that sells heatsinked modules with house branded chips on them.

Just wondering, but why did you go from G.Skill to Corsair instead of to known good memory?

Politically correct RAM!!!

Why can't it be the CPU? Thinking back now this 3-year old CPU always ran hotter than it should. It took water cooling to cool it to acceptable temperatures. Now it can't handle DDR3 RAM faster than 1066MHz. A year ago it could handle RAM even up to 1600MHz. It probably was degraded even more a year ago when the computer it was in took a direct lightning strike. That computer lost the motherboard, 1 stick of RAM, and its TV tuner.
---------------------------------------------------------------------

After reading your condescending post I took apart my two computers and switched CPUs. I reran Memtest86+ on each computer.

Original setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD Phenom II X4 955 3.2GHz | GSkill RipjawsX (2x2GB) DDR3 1333
Computer 2: ASUS M4N82 Deluxe | AMD Phenom II X4 960T 3.0 GHz | Corsair XMS2(2x2GB) DDR2 800

New setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD Phenom II X4 960T 3.0 GHz | GSkill RipjawsX (2x2GB) DDR3 1333
Computer 2: ASUS M4N82 Deluxe | AMD Phenom II X4 955 3.2GHz | Corsair XMS2(2x2GB) DDR2 800

Each computer ran over 2 hours without any errors. I know that is not enough time for a definitive test. Later I will run the test for at least 12 hours.

Of course I realize that the AMD Phenom II X4 955 is now using its DDR2 controller that alone does not prove anything. I do know that it can no longer handle DDR3 RAM faster than 1066MHz.

---------------------------------------------------------------------

To get a better comparison I bumped up the CPU and RAM speeds for Computer #1. The AMD Phenom II X4 960T is now running 3.2GHz and the RAM is running 1600MHz. It ran flawlessly with Memtest86+ for 2 hours. It even ran Prime95 for over an hour before I stopped it.

Current setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD Phenom II X4 960T 3.2 GHz | GSkill RipjawsX (2x2GB) DDR3 1600
Computer 2: ASUS M4N82 Deluxe | AMD Phenom II X4 955 3.2GHz | Corsair XMS2(2x2GB) DDR2 800

The CPU temps are also much better for Computer #1.
AMD Phenom II X4 955 3.2GHz: 30C idle, 45C load
AMD Phenom II X4 960T 3.2GHz: 20C idle, 28C load

Also Hardware Thermal Control is now turned on on computer #1. Cool-and-Quiet varies the CPU speed like it should now from 800MHz to 3200MHz. Before with the Phenom II X4 955 Hardware Thermal Control had to be disabled because it always kept the CPU speed at 804MHz. That was even with water cooling keeping its temperature in the 30-40C range.

Note I am back to using the GSkill RAM. With the AMD Phenom II X4 955 at default settings I would get thousand of errors. With the AMD Phenom II X4 960T I am getting absolutely no errors even when the CPU is bumped up from 3.0 to 3.2GHz and the GSkill RAM is bumped up from 1333 to 1600MHz.

 

[Psykoikonov]

Should have known the Deneb C2 had weak IMC, good to see you got it working.

 

[larrymoencurly]

I don't think it's the CPU. The fact the error count went down as the voltage went up proves that the memory was marginal, and you really need to try some known good non-overclocked memory chips and quit messing around with junk from G.Skill, Corsair, or any company that sells heatsinked modules with house branded chips on them.

Just wondering, but why did you go from G.Skill to Corsair instead of to known good memory?

Politically correct RAM!!!

Why can't it be the CPU?

It's a lot less likely to be the CPU because CPUs are all major brand products tested to very high standards with very expensive equipment, except black market items, which are rare. OTOH most retail memory modules are usually tested only with PCs and to much lower standards (one company even said 2 errors was OK), and a lot of the heatsinked ones are made with overclocked chips.

The politically correct thing to say is that heatsinked major brand memory is the best and that you can trust reviews based on samples hand-picked by the manufacturers.

Thinking back now this 3-year old CPU always ran hotter than it should. It took water cooling to cool it to acceptable temperatures. Now it can't handle DDR3 RAM faster than 1066MHz. A year ago it could handle RAM even up to 1600MHz. It probably was degraded even more a year ago when the computer it was in took a direct lightning strike. That computer lost the motherboard, 1 stick of RAM, and its TV tuner.
---------------------------------------------------------------------

After reading your condescending post I took apart my two computers and switched CPUs. I reran Memtest86+ on each computer.

What was condesending about any of my posts? I said if a computer doesn't run right, get rid of any questionable components, and let's face it: retail memory with unidentifiable chips has a high defect rate.

How do you know your 3-year-old CPU ran hotter than it should have, unless you compared it to a known-good identical CPU under the same conditions? Temperatures are acceptable if they're within the manufacturer's specifications, let the computer run 100% reliably, and don't shorten the lifespans of other components (like wet capacitors). A CPU at 60C - 70C may not be anybody's idea of cool, but it should be fine for anything not overclocked.

 

[MisterEd]

It's a lot less likely to be the CPU because CPUs are all major brand products tested to very high standards with very expensive equipment, except black market items, which are rare. OTOH most retail memory modules are usually tested only with PCs and to much lower standards (one company even said 2 errors was OK), and a lot of the heatsinked ones are made with overclocked chips.

The politically correct thing to say is that heatsinked major brand memory is the best and that you can trust reviews based on samples hand-picked by the manufacturers.


What was condesending about any of my posts? I said if a computer doesn't run right, get rid of any questionable components, and let's face it: retail memory with unidentifiable chips has a high defect rate.

How do you know your 3-year-old CPU ran hotter than it should have, unless you compared it to a known-good identical CPU under the same conditions? Temperatures are acceptable if they're within the manufacturer's specifications, let the computer run 100% reliably, and don't shorten the lifespans of other components (like wet capacitors). A CPU at 60C - 70C may not be anybody's idea of cool, but it should be fine for anything not overclocked.

Sorry I did not realize you were such an expert.

 

[hokiealumnus]

larrymoencurley, we get it. You think only Samsung DDR3-1600 chips are acceptable and everything from everyone else is crap. We. Get. It. Stop posting in people's threads harping on the point. If you must, create your own thread in the memory section and extol the virtues of Samsung memory, occasionally dropping by to complain about everyone else. That way you don't have to pollute every thread with memory issues just to make your claims.

As far as your thread MisterEd, it looks like you got it sorted; glad it worked out for you!

 

[larrymoencurly]

larrymoencurley, we get it. You think only Samsung DDR3-1600 chips are acceptable and everything from everyone else is crap. We. Get. It. Stop posting in people's threads harping on the point.

I don't mind accusations as long as they're true, but that's a pretty gross misrepresentation of my words, and I've never told anybody to buy ONLY Samsung chips or modules, have I? If I'm wrong, please provide a quote.

I just said when bad memory is suspected, substitute known good modules to prove or disprove this. That's no different than telling someone who suspects a power problem to try a good power supply instead of a 600W $19.95 special. I can understand how repeating the same advice over and over can seem like nagging, but it doesn't make it invalid.

 

[hokiealumnus]

Sorry MisterEd, I'll try not to derail your thread longer after this.

I don't mind accusations as long as they're true, but that's a pretty gross misrepresentation of my words, and I've never told anybody to buy ONLY Samsung chips or modules, have I? If I'm wrong, please provide a quote.

Excuse me, Samsung or Crucial, not just Samsung; basically, anything without heatspreaders. The anecdotal evidence is pretty strong...

From a practical standpoint, buy Samsung or no-heatsink Crucial modules.

In practical terms, considering what kind of memory available on the maket, your only choices are:

A. Good manual overclocking (run Samsung 1600 MHz @ 2133 MHz)

B. Bad automatic overclocking (G. Skill 1333 MHz @ 2133 MHz)

C. Manually under-overclocking the automatically overclocked memory (G. Skill 1333 MHz @ 1333 MHz)

If I had to overclock, I'd choose A first, no questions, especially because it has such a good track record, and 1600 MHz chips can probably run faster than 1333 MHz chips. And with those Samsung modules you'll get something that's rock-solid at 1600 MHz, no matter which sample you receive.

Why should G.Skill 2133 MHz modules made with 1333 MHz chips be more likely to run reliably at 2133 MHz than Samsung chips made with 1600 MHz chips? It's not as if G.Skill does super-strict testing because their testers use nothing but PCs.

... buy those Samsungs and run them at their stock 1600 MHz.

...you really need to try some known good non-overclocked memory chips and quit messing around with junk from G.Skill, Corsair, or any company that sells heatsinked modules with house branded chips on them.

Just wondering, but why did you go from G.Skill to Corsair instead of to known good memory?

... let's face it: retail memory with unidentifiable chips has a high defect rate.

I'm not accusing you of anything, I'm just categorizing what is in your posting history. It seems pretty simple really. According to you, anything with a heatspreader is bad and anything without is good, because you can see the ICs. So, I'm saying start your own thread to extol the virtues of heatspreader-less RAM. Lambast the ones with heatspreaders if you feel like it. Go right ahead....just stop polluting other people's threads with your opinion.

 

[MisterEd]

I don't mind accusations as long as they're true, but that's a pretty gross misrepresentation of my words, and I've never told anybody to buy ONLY Samsung chips or modules, have I? If I'm wrong, please provide a quote.

I just said when bad memory is suspected, substitute known good modules to prove or disprove this. That's no different than telling someone who suspects a power problem to try a good power supply instead of a 600W $19.95 special. I can understand how repeating the same advice over and over can seem like nagging, but it doesn't make it invalid.

I guess you missed post #17 by me. I tried new ram. It did not solve the problem like I said. Only switching out the CPU with another computer solved the problem.

BTW, the original Phenom II X4 955 that is now in the other computer is on its last legs now. Instead of 3200MHz it now is only stable at 800MHz. Anything higher causes the computer to crash during boot. If the multiplier is manually set to anything higher that 4 ASUS's Expressgate forces the multiplier back to 4 (4x200=800MHz).

 

[Psykoikonov]

I guess you missed post #17 by me. I tried new ram. It did not solve the problem like I said. Only switching out the CPU with another computer solved the problem.

BTW, the original Phenom II X4 955 that is now in the other computer is on its last legs now. Instead of 3200MHz it now is only stable at 800MHz. Anything higher causes the computer to crash during boot. If the multiplier is manually set to anything higher that 4 ASUS's Expressgate forces the multiplier back to 4 (4x200=800MHz).

If your looking to replace it there is a 955BE for $80 and an 1100T for $135 in the classifieds.

 

[Neuromancer]

He got a Zosma, it has the thuban IMC so he should be in good shape now

If you are going to run at 1600 MHZ Effective RAM speed you should increase your CPUNB speed to match (IE 1600mem=2400Cpunb minimum) my 1090T has no issues running up to 2700 24/7 so 2400 should be easy peasy.

 

[Psykoikonov]

He got a Zosma, it has the thuban IMC so he should be in good shape now

If you are going to run at 1600 MHZ Effective RAM speed you should increase your CPUNB speed to match (IE 1600mem=2400Cpunb minimum) my 1090T has no issues running up to 2700 24/7 so 2400 should be easy peasy.

Original setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD Phenom II X4 955 3.2GHz | GSkill RipjawsX (2x2GB) DDR3 1333
Computer 2: ASUS M4N82 Deluxe | AMD Phenom II X4 960T 3.0 GHz | Corsair XMS2(2x2GB) DDR2 800

New setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD Phenom II X4 960T 3.0 GHz | GSkill RipjawsX (2x2GB) DDR3 1333
Computer 2: ASUS M4N82 Deluxe | AMD Phenom II X4 955 3.2GHz | Corsair XMS2(2x2GB) DDR2 800

He has 2 computers, just giving him a cost effective option to keep them both going strong if he wants too.

 

[Neuromancer]

Ahh... missed that post saying it was dying.

 

[larrymoencurly]

Excuse me, Samsung or Crucial, not just Samsung; basically, anything without heatspreaders. The anecdotal evidence is pretty strong...

I'm not accusing you of anything, I'm just categorizing what is in your posting history. It seems pretty simple really. According to you, anything with a heatspreader is bad and anything without is good, because you can see the ICs. So, I'm saying start your own thread to extol the virtues of heatspreader-less RAM. Lambast the ones with heatspreaders if you feel like it. Go right ahead....just stop polluting other people's threads with your opinion.

I said anything rated for a higher than normal operating voltage or with chips that can't be identified by actual manufacturer should have to prove its innocence through very thoroughly testing. While that includes anything with heatspreaders, it doesn't exclude all modules without them.

If the anecdotal evidence is pretty strong, why are you disagreeing with me?

 

[larrymoencurly]

I guess you missed post #17 by me. I tried new ram. It did not solve the problem like I said. Only switching out the CPU with another computer solved the problem.

That's good evidence of a defective CPU, but also could mean that the new RAM was also marginal, and considering the track record I've recently had with Corsair (which never let me down in the old days), I can't rule that out.

BTW, the original Phenom II X4 955 that is now in the other computer is on its last legs now. Instead of 3200MHz it now is only stable at 800MHz. Anything higher causes the computer to crash during boot. If the multiplier is manually set to anything higher that 4 ASUS's Expressgate forces the multiplier back to 4 (4x200=800MHz).

I didn't realize it was THAT bad. That changes everything.

 

[hokiealumnus]

The anecdotal evidence of your war against anything with heat spreaders and in favor of Samsung (and Crucial). I do not agree with your opinions.

 

[MisterEd]

Computer #2 with Phenom II X4 955 no longer posts so I am assuming that the CPU finally died. If I really want to use it I could always throw in an old Athlon 64 X2 3800+. Not the best but better than nothing.

Computer #1 with the Phenom II X4 960T is still using the GSkill RAM which I was unsure about before. Remember this same RAM was very unstable with the old Phenom II X4 955 CPU at anything higher than 1066MHz. Now it is very stable now even at 1600MHz. Since I will be taking online classes for the next year I need to keep at least this computer going for now. Next month during the term break I will decide whether to replace the CPU then or wait for the AMD Piledriver which is due out in the next few months.

 

[Psykoikonov]

Computer #2 with Phenom II X4 955 no longer posts so I am assuming that the CPU finally died. If I really want to use it I could always throw in an old Athlon 64 X2 3800+. Not the best but better than nothing.

Computer #1 with the Phenom II X4 960T is still using the GSkill RAM which I was unsure about before. Remember this same RAM was very unstable with the old Phenom II X4 955 CPU at anything higher than 1066MHz. Now it is very stable now even at 1600MHz. Since I will be taking online classes for the next year I need to keep at least this computer going for now. Next month during the term break I will decide whether to replace the CPU then or wait for the AMD Piledriver which is due out in the next few months.

Good to hear it's working and that your assumption that it wasn't the Ram was correct. Unlikely that a processor would go like that but it happens. Can be hard to help from a distance on the forums like this, much easier if it's hands on. Regardless good to see one computer is working and the other one depending on the school work with a 3800+ X2 might work well too.

 

[MisterEd]

I wanted to make this last reply to tie things up. I was finally able to resolve things after the Spring Term ended.

It wasn't quite true that Computer #2 with the Phenom II X4 955 would not boot at all. To recap after I powered it up normally no video would ever show on the monitor. Sounds from the computer and activity from the boot drive indicated that it was going through the normal boot process.

I discovered that after I let the computer go through the normal boot process for ~20 seconds I would hold the power button down to force it to shut down. A few seconds later I would hit the power button again. The computer would then boot with the normal video appearing on the screen. The only problem was that the CPU was only running 800MHz instead of the normal 3.2GHz. Nothing I would do would get the computer to boot at any faster speed or even change the speed after it booted. I put up with this for the last nine months. It worked OK just at a slower speed. I could even play games on it since the GPU still worked at its normal speed.

Last month (April 2013) I bought a new CPU for Computer #1 which allowed me to return the CPU in it to Computer #2. Here is a summary of what changes were made:

Original setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD Phenom II X4 960T @3.0 GHz | GSkill RipjawsX (2x2GB) DDR3 1333 | Windows 7 Ultimate 32-bit
Computer 2: ASUS M4N82 Deluxe | AMD Phenom II X4 955 @800MHz | Corsair XMS2(2x2GB) DDR2 800 | Windows 7 Home Premium 32-bit

New setup
Computer 1: Gigabyte GA-990FXA-UD3 | AMD FX8350 @4.0 GHz | GSkill RipjawsX (2x2GB) DDR3 1600 | Windows 7 Ultimate 32-bit
Computer 2: ASUS M4N82 Deluxe | Phenom II X4 960T @3.0 GHz | Corsair XMS2(2x2GB) DDR2 800 | Windows 7 Home Premium 32-bit

Future setup*
Computer 1: Gigabyte GA-990FXA-UD3 | AMD FX8350 @4.0 GHz | GSkill RipjawsX (2x4GB) DDR3 1600 | Windows 7 Ultimate 64-bit
Computer 2: ASUS M4N82 Deluxe | Phenom II X4 960T @3.0 GHz | Corsair XMS2(2x2GB) DDR2 800 | Windows 7 Home Premium 32-bit

*when I get everything documented and backed up