Sign in with
Sign up | Sign in
Your question
Solved

Installing RAM

Last response: in Memory
Share
January 20, 2010 3:45:57 PM

Hello I need some help installing my desktop computer RAM. Below is a link to my motherboard.

http://www.newegg.com/Product/ImageGallery.aspx?CurImag...

You see, there are 4 memory slots, 2 blue, and 2 white. The installation guide said to put one in the white and one in the blue, but when I did that cpu-z recognized it as single channel mode. When I put them both in the blue, it read them as dual channel mode.

Right now I have them both in the blue, and cpu-z is reading them as dual channel. I also had to go into the bios to manually set my RAM speeds at 1600.

Can anyone tell me if I've done everything properly? if putting them both in the blue and running dual channel is the best performance? Windows experience index only rates them at 5.9 which i find odd because they are dd3 1600...my laptop with dd2 is rated at 6.3, both computers have 64 bit OS and 2x2GB sticks.

thank you!

More about : installing ram

a c 81 } Memory
January 20, 2010 5:53:35 PM

If they are detected in dual channel mode, then they are installed correctly. The experience index is from Vista or Windows 7?
m
0
l
January 20, 2010 6:18:58 PM

The mobo manual is a bit confusing, but on page 16 it shows on the Dual Channel Memory Configuration table that for dual channel operation with two DIMMs you put them either in DDR3_1 and DDR3_2 (which are both blue) or in DD3_3 and DDR3_4 (which are both white).

The problem might not be the slots but the timings the memory is set at.

What memory do you have?
Is it on the memory compatibility list for the mobo?
What are the timings set at?
Have you changed any timings on the memory besides the primary ones - like 8-8-8-24?
After installing system, did you set BIOS to select "load optimized defaults"?

Have you run memtest to check memory stability? If not you can test memory with memtest, a free utility found that you can download with instructions from:

http://www.memtest.org/

You want to run memtest for 8 hours without any errors to prove memory is stable. If ANY errors are reported, you trouble shoot the problem.

First try to run 10 passes of the test with all memory installed to get an intermediate level of assurance of reliability. If it passes, you can either proceed to immediately to try the 8 hour run, or if more convenient do the 8 hour run overnight.

m
0
l
Related resources
a b } Memory
January 20, 2010 6:26:34 PM

Is the desktop running VISTA ?? - If so the windows experience rating maxes out at 5.9 !! they increased the rating system to 7.0 with the release of win7 so if the laptop is running win7 and the desktop is still Vista that would explain the rating difference !!
m
0
l
January 21, 2010 10:34:18 AM

OK I can't believe I installed those RAM sticks wrong. Stupid mistake. I confirmed with my manual and sure enough. I've got em both in blue with dual channel config, so that's good.

Now the windows experience index is only at 5.9 (on windows 7) which seems too low for what i thought was some of the fastest memory available. heres what cpu-z is telling me.

http://yfrog.com/14cpuzmemoryap

If I'm not mistaken, my CAS latency is 2 notches higher than the technical specification. Am I right? can someone tell me how to fix it?

thanks you've all been great so far!
m
0
l
a b } Memory
January 21, 2010 12:48:25 PM

llmercll said:

If I'm not mistaken, my CAS latency is 2 notches higher than the technical specification. Am I right? can someone tell me how to fix it?

thanks you've all been great so far!


YOu are correct the timings are all a couple notches above what they should be - go back into your bios and manually set the timings. I have the same MOBO so what you want to do is :

Enter the Bios - go to the MMb Intelligent Tweaker page - find the DRAM configuration line and press enter -(this will bring up a second window with a bunch of timing settings - on the DDR3 timings items change the first 3 to 9 - the 4th to 25 (Minimum RAS active TIme) and then the 5th line to 2T (Command Timing)- after making those changes save the changes and let the system boot.

If for some reason it does not boot and gives you an error - go back in and check the Voltage settings and ensure that it is set to 1.5V (MB intelligent Tweaker page) - DDR3 Voltage Control (change from normal to 1.5 if needed) and then reset those timings again and retry the boot. If still no Luck the Ram is not performing at the rated speeds and you can either RMA them or try upping the Voltage a small amount (try .05V increases once or twice) and see if that makes them stable. (sometimes it takes one or two small Voltage increases since the actual Voltage is sometimes just a bit lower than the reported reading.)
m
0
l
January 21, 2010 3:11:55 PM

Correcting the timings should give you the performance you expect. After making the changes, you can run memtest to confirm memory stability.

You might also want to the free system test at PCPitStop that provides some basic checks and benchmarks on your system. Saving the benchmarks provides a basis for comparison if later you think your system might not be performing well or if you want to see how performance is affected by upgrades.

http://www.pcpitstop.com/
m
0
l
January 21, 2010 8:00:54 PM

I successfully changed the ram timings. the bios had 2 rows next to what i had selected, SPD and Auto. I just changed everything manually to what SPD was at.9-9-9-24 etc. there were a few other ones too.

CPU-Z read the ram timings correctly, at 9-9-9-24, but windows experience index STILL rates it a 5.9.

wtf is with that. Is it normal than or what?
m
0
l
a b } Memory
January 21, 2010 8:06:55 PM

What are the timings on the DDR2 Ram in the laptop ?? - it's possible that the DDR2 is actually faster if it has better timings (the difference between DDR3 and DDR2 are not like the difference between DDR and DDR2 were !

YOu might also try running your DDR3 1600 at DDR 1333 and lower the latency to something like 7-7-7-20 and see if that is stable and might get better performance - with the newer DDR3 memory it takes a bit of experimenting to see what gives the best throughput on different modules\MOBO combos.
m
0
l
January 21, 2010 8:41:41 PM

You can:

1. Run memtest as recommended above.
2. Go to PCPitStop and run their system test as suggested above.

Both of those are things you should learn to do as a normal course of operations anyway.

3. Go to the Gigabyte official website forum to learn more about the other timings:
......a/ Search based on your mobo and memory and see if any are provided, or
......b/ Post your issue just like you did here but relate all the things done to date

4. Download some benchmarking software and see how it evaluates your memory - might just be a problem with the experience index.
m
0
l
January 21, 2010 9:25:51 PM

OK, I'll get to that rockyjohn. Any idea what that memory should be posting around?
m
0
l
January 22, 2010 7:20:24 PM

I ran pcpit stop. Seems gimmicky to me. Just told me my harddrive was fragmented and recommended I make a bunch of registry fixes to increase speed.

It did test my memory, though i don't know what to make of the results. It came in at 18,000 MB/s

I'll take your advice and run a benchmark. anyone know a good one out there where I can test my memory speeds and compare it to others like mine? Just to make sure I'm getting the most bang for my buck.

thanks a lot everyone
m
0
l
January 23, 2010 10:27:51 AM

bump
m
0
l
a c 81 } Memory
January 23, 2010 11:06:03 AM

What's the performance on memtest86+? Why you're so worried about benchmarks? Is your system really slow?
m
0
l
January 25, 2010 1:46:01 PM

I didn't run memtest, it says invalid file.

I wouldn't say I'm "worried", I just like squeezing the best performance out of my hardware. Nothing wrong with that =p
m
0
l
a c 81 } Memory
January 25, 2010 2:06:27 PM

Are you booting from the CD on which you burned the memtest86+ ISO?
m
0
l
January 26, 2010 2:28:42 AM

Yeah, I didn't troubleshoot it too much though. I should try reburning it in imgburn.
m
0
l
January 31, 2010 2:44:40 AM

Fixed it. I ran memtest 86+ for a little over an hour, 2 passes. There were 0 errors =)

unfortunately it didn't benchmark anything. Back at square one, anyone have a reputable benchmarking software I can use?
m
0
l
January 31, 2010 4:18:28 AM

No memtest does not benchmark - but it tells you if it is stable. As I said above - you need to run for 8 hours to confirm stability. 2 passes is not enough.

I don't know what benchmarking program would be good, but if it passes memtest for 8 hours I doubt there is anything wrong with the memory.
m
0
l
February 1, 2010 2:40:01 AM

I wasn't as concerned with stability as much as performance (I just assume it's stable unless I have any problems) but I'll run the test anyway.

Is it safe to leave it overnight though? I just feel like stressing my computer for that long would be bad for it.

Just read some newegg reviews...

"Even running at 1333 it got a 7.5 rating from Win7."

"7.8 Windows Experience Index on Windows 7."

"Works well with W7 64-bit."

"Scored a 7.5 in windows experience!!!"


WTF why are mine only giving me 5.9???? What else can I fu**** do? Could it have to do with on-board video?
m
0
l

Best solution

February 1, 2010 1:44:30 PM

llmercll said:
I wasn't as concerned with stability as much as performance (I just assume it's stable unless I have any problems) but I'll run the test anyway.

Is it safe to leave it overnight though? I just feel like stressing my computer for that long would be bad for it.


LOL. Stability is basic to any performance - it is the lowest level of "performance". If not stable, if not working correctly, then how does performance matter? What does it matter that your car can go 150 MPH if you can't steer it?

In the worse case, lack of stability will crash computer, and you have no "performance". In the middle, it will slow down performance or you might see glitches in performance, but not know the cause, maybe like memory performance not being what you expect. It is also possible that you won't seen the errors - they be will hidden. The sum on your spread sheet is not correct and you will use the wrong number, or the shot in your game is not a hit.

Do you always not listen after asking for help?

If the PC is not overheating in normal operation, it is safe to run it memtest overnight - otherwise why would I recommend it? First, running memtest does not stress memory, where did you get that idea? It just methodically checks it running operations to ensure it checks all parts of the memory. Second, computers are designed to run in continuous operation under load - youi just need to ensure you have provided the proper cooling for the system. If you have done that right, you can run prime95, which does "stress" the computer - actually just the CPU when running that part of the test - with no problem. It is no different to computer than sitting in front of computer and running full load on a game for an hour. If you don't want to stress your PC you should stop playing games on it. Whether you do it for 8 hours overnight or 8 hours at one hour a day for 8 days makes little difference.

I have been waiting for you to complete the basic stability test before proceeding on to the next level of troubleshooting. You have thus far wasted 11 days by not doing the complete stability test recommended 12 days ago.

If performance is slower, it could easily be timing issues. The best place to find an expert on those is on the Gigabyte official forum at TweakTown.com. If you go there and want to get the most out of their help, I suggest you listen more closely and try what they suggest rather than be still asking questions about something recommended 11 days earlier and only asking because they were telling you for the fourth time. They might not be so patient with you.

http://forums.tweaktown.com/f69/
Share
February 1, 2010 5:14:49 PM

rockyjohn said:
LOL. Stability is basic to any performance - it is the lowest level of "performance". If not stable, if not working correctly, then how does performance matter? What does it matter that your car can go 150 MPH if you can't steer it?

In the worse case, lack of stability will crash computer, and you have no "performance". In the middle, it will slow down performance or you might see glitches in performance, but not know the cause, maybe like memory performance not being what you expect. It is also possible that you won't seen the errors - they be will hidden. The sum on your spread sheet is not correct and you will use the wrong number, or the shot in your game is not a hit.

Do you always not listen after asking for help?

If the PC is not overheating in normal operation, it is safe to run it memtest overnight - otherwise why would I recommend it? First, running memtest does not stress memory, where did you get that idea? It just methodically checks it running operations to ensure it checks all parts of the memory. Second, computers are designed to run in continuous operation under load - youi just need to ensure you have provided the proper cooling for the system. If you have done that right, you can run prime95, which does "stress" the computer - actually just the CPU when running that part of the test - with no problem. It is no different to computer than sitting in front of computer and running full load on a game for an hour. If you don't want to stress your PC you should stop playing games on it. Whether you do it for 8 hours overnight or 8 hours at one hour a day for 8 days makes little difference.

I have been waiting for you to complete the basic stability test before proceeding on to the next level of troubleshooting. You have thus far wasted 11 days by not doing the complete stability test recommended 12 days ago.

If performance is slower, it could easily be timing issues. The best place to find an expert on those is on the Gigabyte official forum at TweakTown.com. If you go there and want to get the most out of their help, I suggest you listen more closely and try what they suggest rather than be still asking questions about something recommended 11 days earlier and only asking because they were telling you for the fourth time. They might not be so patient with you.

http://forums.tweaktown.com/f69/


I'm sorry rockyjohn, It wasn't my intention to frustrate you, but I completely understand your feelings. You've been with me 100% on this issue and I'm sorry for disregarding your advice. I never thought stability could cause a decrease in speed, but it's pretty stupid, and hypocritical of me, to come here asking for help on something I know little to nothing about and then shun the advice.

Anyway, I'll run memtest+ overnight and post the results first thing in the morning. Thanks for the clarification =)
m
0
l
February 2, 2010 4:53:23 PM

OK, I ran the test overnight for a total of 11h:22m. Set it up a bit early so it ran for longer than 8 hours.

Here are the results...

Memory: 4580 MB/s

13 Passes

0 errors (yay)

ECC off (not sure what this is, but figured I'd add it anyway)

I guess this means my RAM is good? so there must be a bad setting somewhere right?
m
0
l
February 3, 2010 12:15:54 AM

Yes, the RAM should be good - and since it is stable the settings are not too bad - not bad enought to make unstable. But you might be able to improve performance by tweaking them.

Now you should go ask the experts at TweakTown for help on that - and you can tell them you have already run memtest for 11 hours with no failures.

They might also be able to recommend some benchmarking software.
m
0
l
a b } Memory
February 9, 2010 10:41:34 PM

Best answer selected by r_manic.
m
0
l
!