G-Skill's price went down finally!

[Capt_Caveman]

I thought about going the DFI nf4-Ultra and Mushkin Redline route but reading over at xs, it has even been admitted by DFI that that their boards have fried or killed ram when going over 3.3v. Even though the board and ram would be warranted, I have no desire to have to rma and have no pc for several weeks.

 

[Eldonko]

No offense but there are equally as good products from Mushkin and OCZ at a fraction of the price. Gskill = Over-rated.

I wouldn't say GSkill LA and LE TCCD is over rated. Sure OCZ has better customer service and cheaper prices, but if you look at the top RAM clocks around at the moment (~DDR640), the majority of the sets doing these speeds are GSkills. Whats better than getting 305-310Mhz at 2.5-3-3? Nothing. Thats why you pay the $$, for the few extra Mhz.

 

[deception``]

I thought about going the DFI nf4-Ultra and Mushkin Redline route but reading over at xs, it has even been admitted by DFI that that their boards have fried or killed ram when going over 3.3v. Even though the board and ram would be warranted, I have no desire to have to rma and have no pc for several weeks.

Wrong. To this day, there is no conclusive evidence that the DFI NF4 is at fault for the bad memory. Nonetheless, it goes without question that people familiar to the situation are looking at this closely. Before you make such comments, keep in mind that there are a number of individuals who have yet to experience such problems, myself included.

deception``

 

[rseven]

Wrong. To this day, there is no conclusive evidence that the DFI NF4 is at fault for the bad memory. Nonetheless, it goes without question that people familiar to the situation are looking at this closely. Before you make such comments, keep in mind that there are a number of individuals who have yet to experience such problems, myself included.

deception``

I f that doesn't sound like a 1960's press release from the tobacco industry I don't know what does. I believe the key operative word here is "yet". Furthermore, even if you never have problems the only thing that proves is you are lucky. The simple compelling logic is that high voltage kills ram.

 

[Capt_Caveman]

Wrong. To this day, there is no conclusive evidence that the DFI NF4 is at fault for the bad memory. Nonetheless, it goes without question that people familiar to the situation are looking at this closely. Before you make such comments, keep in mind that there are a number of individuals who have yet to experience such problems, myself included.

deception``

You're not going to have the problem b/c you're not pushing high dimm voltages w/ your ram.

Here's just one link to the problem:
http://www.xtremesystems.org/forums/showthread.php?t=63932

And obviously not every board and person is going to have this problem, but enough folks have reported and rma'd their boards and memory to tell me not to change the ram voltage to +5v on a DFI board. We're talking about about going to extreme's with voltage, which by that fact could lead to issues.

I had ordered a DFI nF4-Ultra and planned to use my TCCD memory but canceled when I could get the silent Asus A8N-Sli Premium and heard folks were getting 300+mhz 1T.

 

[burningrave101]

I f that doesn't sound like a 1960's press release from the tobacco industry I don't know what does. I believe the key operative word here is "yet". Furthermore, even if you never have problems the only thing that proves is you are lucky. The simple compelling logic is that high voltage kills ram.

If your not using a crazy amount of voltage and you have active cooling on the RAM then the chances of voltage killing it before the RAM is long retired isn't that great, especially if its BH style RAM. The majority of sticks that have died were UTT-CH based sticks. Very few people have lost UTT-BH sticks. Just use the 3.3V jumper and if you need a little more then 3.2V then raise the +3.3V rail on your PSU. The OCZ PowerStreams make this a simple process and then your not shooting +5V through your board. You can usually hit 250Mhz 2-2-2-5 with 3.1-3.2V on UTT-BH if its the good stuff like G.Skill GH.

 

[RedDragonXXX]

I believe that is like saying there is no harm in smoking. The effects are not immediate so people thought for years there was no harm done. However, over time the effects are deadly. I don't think UTT is not worth the risks for the supposed gain, but hey, whatever floats your boat.

My point exactly rseven.

 

[crimedog]

Awesome, haven't seen a bh5/utt vs tccd thread in quite a while! Time to beat that horse!

There are situations where either is clearly better. For example on a dfi nf3/4 board you will want bh5/utt. If you want the best, that's it. You must run high vdimm, you must actively cool the ram and the board, but it's the best. If it's not worth it to you than fine, you're just not extreme enough.

If you're running an amd64 board without enough vdimm, or if you're an intel owner you want tccd because you can run high speeds, and intel doesn't care much about timings.

Refusing to run high voltages is personal preferance. Not buying a chip with a high enough multi is your mistake. Not wanting to run with a divider is just silly.

There is no proof that the 5v jumper kills anything. For every person complaining about it there are 5 who have been running the jumper 24/7. The guy who made that thread at XS is just looking for a pat on the head. All that really happened was the tech support realized that this kid should not be messing with the 5v jumper because he didn't know what he was doing. I'm not saying that the dfi nf4 is bulletproof because dfi makes the most tempermental products out there. They overclock like crazy right up to the point where they die. But we're talking about simple voltage adjustments so anyone who blames that gets ignored by me.

 

[RedDragonXXX]

All these posts about UTT's make me want to get some Mushkin Redlines and see how they stack up against my TCC5, even though they are not the best OC'er (290MHz highest). Heck for $145 I might just do it.

 

[Samk]

CPU (buying the best batch)

I read here alot about the 3000 amd64 with the batch
code - LBBLE - . How do you go about buying that coded
chip? On all the venders web sites that is not a option.

Thanks

 

[Eldonko]

I read here alot about the 3000 amd64 with the batch
code - LBBLE - . How do you go about buying that coded
chip? On all the venders web sites that is not a option.

Thanks

You go to the store and if you are lucky they may let you hand pick a chip.

 

[windwithme]

in newegg
NEWEGG

1GBFR DDR600 2T $235.00

1GBGH DDR400 2 2-2-5 $147.01 BH5

GSKILL 1GBFF.LA.FR.LE.LC.LD is SAMSUNG TCCD not TCC5

 

[dumo]

Heres 'balls to the wall' pure speed comparo....
tccd @335 and +100mhz on cpu with 3.1V >< bh-5@271 with 3.6V....

 

[Skeen]

Hmm not sure what to make of the test from a RAM p.o.v. It seems the test with the tccd, in which the cpu is running 100MHz faster, is fractionally faster than the test with the BH5. Pretty much what you would expect from Super Pi.

 

[HousERaT]

I think that pretty much sums up the superiority of low latancy..... 100mhz less on the cpu clock and almost the same time....... You should do a comparison at the same clock speed just to see what the difference is if any......

 

[Skeen]

As I stated about two posts up, I didn't think that this was a significant find (dumo's test, that is) but when I tried it myself, I got about 1 whole second slower on superpi when I tested at 2700MHz and 2600MHz. All RAM timings were the same. So I guess HousERat is correct, low latencies rule (on super pi, on the AMD64 platform).

 

[dumo]

AM3....245 and 313

 

[burningrave101]

AM3....245 and 313

You dont suppose that has anything to do with the fact that the FX-55 with the higher score is overclocked a little higher then the other do you?

At the end of the day lower frequencies with tighter timings are going to give you relatively the same exact performance if not a little better then higher speeds and more lax timings and outside of benchmarks you'll never notice the difference.