Mega-Overclocking generic DDR2!!!

salgado18

Distinguished
Feb 12, 2007
933
376
19,370
Could somebody please tell me how is it possible to have a generic DDR2-800 module, originaly set to 5-5-5-15, work at 4-4-3-5 !!!!!!! Two screenshots from CPU-Z:

imagem1zg6.jpg

imagem2sx8.jpg
 

salgado18

Distinguished
Feb 12, 2007
933
376
19,370
Talk about pure luck... it was set on the bios itself, the first two didn't boot when using 3, but the others are on the motherboard minimum...! I didn't try setting the Command Rate to 1T, because reseting bios is very boring, but it is still impressive!

Does anyone know a way to test memory speed? Something that's affected by latency?
 

rammedstein

Distinguished
Jun 5, 2006
1,071
0
19,280
you see, the thing is, value ram, dispite what people say is quite overclockable, in generally i can get a kingston kvr667 kvr kit from ddr2-667 cl5 to ddr2-800 cl4, all you do is bump up the voltage, sure it may not overclock wildly like corsairs dominators or ocz reapers or whatever but it is a damn side cheaper and works, also since you said your ram is rated at ddr2-800 and your ram is actually running at more ddr2-720~ youwould also be able to lower your latencies a bit...
 

salgado18

Distinguished
Feb 12, 2007
933
376
19,370
Must agree with you, altough it's not *a bit* lower... :)
Now, seriously, is it possible that the RAM is flawed? Like the settings are having no effect? And is it possible to benchmark it?
 

salgado18

Distinguished
Feb 12, 2007
933
376
19,370
Well... try it!
I just benched the memory with MemTach, the whole routine took 167s on stock, and 154s on the oc settings. Definitely worth it.

If you want to try, lower one setting at a time and then reboot everytime. If the system boots, lower again, if it doesn't, return to the last stable number and go to the next. should be like this:

5-5-5-15 (boot)
4-5-5-15 (boot)
3-5-5-15 (no boot, back one)
4-4-4-15 (boot)
4-3-5-15 (no boot, back one)
4-4-4-15 (boot)
4-4-3-15 (boot, minimum for motherboard, next)
4-4-3-14 (boot)
... and so on. AFTER, run a stress test to make sure it's stable.

Know that I'm an amateur on overclocking, so take my advice as my own experience, and not as a guarantee of results.
 

shargrath

Distinguished
May 13, 2007
237
0
18,680


142s for me
I tend to use MemTest for my memory for a few hours the orthos for my cpu when ever i push my oc.
 

weskurtz81

Distinguished
Apr 13, 2006
1,557
0
19,780
slaqado18,

You probably will not see a real world difference between the two settings. You have tightened the memory up a bit, but do some real world tests before you decide it is worth it. The reason is, right now you are testing ONLY your memory performance, but when you do something real world you are testing system performance. Something like a faster CPU, HDD, GPU normally make the biggest real world differences.

The reason I say wait to decide if it is worth it is because your system might become a little unstable. The thing with messing with memory is that it is possible for it to crash very inconsistently. I would leave it alone, the gain probably won't be noticeable. It's cool to know the memory can do it though.
 

salgado18

Distinguished
Feb 12, 2007
933
376
19,370
I agree. I'll use the tighter memory settings until I see the system crashes the first time, then I'll relax them, probably to stock level. If it never crashes, I won't need to "unclock" it.
And I know that CPU and GPU performance are the most important for games for example, but they're not worth overclocking right now. One is very limited by motherboard, and the other is still enough for me. But thanks for the advice!
 

weskurtz81

Distinguished
Apr 13, 2006
1,557
0
19,780
Yeah,

If the motherboard is setting them memory, then it will probably be perfectly stable. It looks like you got good memory at a discount, it happens alot. It is very possible that it will be perfectly stable, just be careful.

I have corrupted my OS messing with memory stuff before. It only happened to me after I crashed the system to many times. So, if it crashes just relax it a bit and you will be fine.
 

cyberjock

Distinguished
Aug 1, 2004
305
0
18,780


So you bought a high performance car last week and the manual says you'd better use 92 octane gas or you'll be sorry. You put 87 in because it's so much cheaper. You fill that tank full and nothing happens, you do it for a few months with no ill effects. Suddenly your nice new car is in the repair shop, and they're telling you "Haha. You didn't put 92 octane gasoline in your car, did you?". So if you are gonna wait till you break something that's BAD logic.

I work in nuclear power. Do you want to know what would happen if I decided to use my infinite amounts of engineering skills and prowess to push the limits on metals because 'it hasn't failed yet'? I'll tell you one thing. If even one person doing my job thought as ignorantly as you, I'm sure there's thousands, if not millions, of lives that will never forget that day, because it just might be their last...

Sure, this is an extreme example, but it really scares me how many people want to push the envelope for little or no benefit, and say "I'll lower it if/when it crashes". I'm not talking just about computers and cars. I'm talking about everything. People writing checks because 'they won't clear until Friday', the 'I'm gonna ignore the manufacturers recommendations for gasoline', and 'I know that the brake system on my car will work just as well with brake fluid as it will with water'(Yes, I know someone that does this, and I'm just waiting for him to die in a crash because his brakes suddenly didn't work). I could write 50 examples right now, but I'm sure this forum gets the point. Unfortunately if you burn out your components, you'll go crying that company x has bad products, you'll get it RMAed and buy new. Who just paid for that new MB you got because you were too ignorant to know better? ME! I'll pay extra at the checkout because YOU didn't follow simple directions like "Use as directed.".

Bottom line: Your logic is terribly flawed, and I hereby will question all motives you have when posting. You, sir, scare the **** out of me and I hope you don't live in my city. Read my signature and eat it!

(BTW.. do you use Microsoft Windows for an OS? Of course you do, look at your screenshots! It WILL crash. I guarantee that. And do you have the expertise to KNOW whether it was your RAM that cause it or another component? NO! I bet there's less than 10 peopleon this whole forum that have the knowledge and skills to take a memory dump from a blue screen and determine what exactly happened.)

If you think this is a flame, it pretty much is. I hope to God you rethink your attitude with pushing the envelope. You do it now and you probably do it work and I don't want to know that you manufacture parts for the military where lives are on the line. My friend might be out there and die because you are a moron. Complacency kills.
 

weskurtz81

Distinguished
Apr 13, 2006
1,557
0
19,780
cyberjock,

all good points, but work and play are two different things. If he is just messing around with his home pc that he doesn't have important data on just to see what he can do, I don't really see the harm.

Also, I would like to think you can treat your work and home environment differently. It really all boils down to good judgment. Knowing when you need to limit your risk and play it safe. I my job we don't cut corners and take risks because we could kill 100's of people. At home, messing around with a pc, I can't kill anyone, and can't really fry memory unless you up the voltage to much. I can make the decisions, and I hope he can as well.

My job that I was referring to was the military. And you are very much correct.... lives were always on the line. Before you get into something like that you look at things differently. But when you know your irresponsibility could kill, your eyes really start to open. When it came to safety..... I really can't put it into words.... just think of you knowing that your siblings life was in your hands.... that is how I always felt.

Rant... I know. I don't think that this guy messing with his memory is a big deal. And I think the whole point of it is, he is using Auto settings and his bios is making the memory run at those latencies, suggesting that his memory is better than what it was advertised as. I don't think he forced the latencies in his pc.... just noticed they were not as advertised.
 

jedimasterben

Distinguished
Sep 22, 2007
1,172
1
19,360
I had some generic DDR1 a while ago that would go down to somewhere around 2-1-2-2-1T, from the stock latencies at 3-4-4-8 or so. Didn't have too much of a performance difference, but it was fun to see a 1 in the settings.
 

salgado18

Distinguished
Feb 12, 2007
933
376
19,370


I'm quoting your message, but I'd like to quote your signature: your RAM has the same settings as mine, except for frequency...
 

zenmaster

Splendid
Feb 21, 2006
3,867
0
22,790
Overclock or UNDERCLOCK?

Looks underclocked to me?

1) It's not running at DDR2-800. So you have slowed the RAM Down.
2) It's running in Single Channel which kills performance but does allow tighter timings.

If you ran it at DDR2-800 in Dual Channel mode, it would perform far better than your "Overclock", if that is what you want to call it.
 

salgado18

Distinguished
Feb 12, 2007
933
376
19,370
Since I'm running an Athlon64, it won't run at 800, only at 723 or 904, which would be a real overclock and the memory does not support. But actually 723MHz is already an overclock since it is rated at 667. And yes, it's single-channel, I'll only find it out when I upgrade to Vista and add another memory on dual-channel...