Buffalo Firestix PC2-8000 on EVGA 680i

[MakubeX]

These GKX sticks are surprisingly good for the price.

600MHz at just 2.1v !

*Everything was done in BIOS. No memset or anything a like was used.*

SuperPI 32M Stable

Here's 1M with a much tighter trc timing. Still just 2.1v, as you can see.

I would try 32M but it takes too long to do it right now.

Sadly, it seems that 600MHz is its sweet spot. Any higher required a significant jump in voltage. Got it up to 625MHz with 2.5v.

Still, for ~$214, this memory is hard to beat. Running the memory at 550MHz 24/7 with just 2.1v is very sweet. Knowing that is able to handle 600MHz at the same voltage is even better.



.

 

[MakubeX]

I was told that it might be my memory controller giving out after 600MHz. So later, when I get the chance, I'll try 600MHz but with tighter timings and see how tight can they go.

 

[Rob Williams]

DDR2-1200 without the need of a fan is great! Apparently the Tracer 8500 kit I have can do the same, but I haven't tested it out. (cf)Eclipse has hit 618MHz 5-5-4-7 with 2.15v I believe.

I'd be curious to know if it is indeed your memory controller. I haven't had a problem with the 9+ kits I've tested in the eVGA board, but the Crucial went from 100% stable at 609MHz to completely unstable at 610MHz. Makes one wonder.

 

[MakubeX]

I do have some fans over the memory. I'm using the CORSAIR CMXAF1 Fans:

Well it seems it was the memory controller that didn't like going higher than 600MHz. Check out these new results:

I was unable to lower the CAS latency under 5 at 600MHz no matter the voltage (up to 2.5v), but the other latencies were more cooperative :

*As said before, this was all done in BIOS. No changes to memory latency, bandwidth or voltage were made in Windows, which means the memory had to load Windows with all my crap in it (this is not a benching rig, this is my personal - gaming - everyday rig).*


@ 2.1v:

With a little bump to 2.35v:

I'm sure I've could've take the Tras and Trc lower with more voltage, but it wasn't worth it.


I couldn't get the "RAS to CAS" and "RAS Precharge" latencies to go under 3 @ 600MHz even with 2.5v, but still I'm very happy with the results.

For $214 I don't think you'll find better memory (that is from a retailer/ e-tailer ).

 

[Rob Williams]

I didn't mean to make it sounded like you didn't have a fan, I just mean you shouldn't need one at that voltage. They would be warm, but not finger numbing hot.

For $214 I don't think you'll find better memory (that is from a retailer/ e-tailer ).

If I have time today, I'll see what the Tracer can handle at 600MHz. NewEgg sells those for $175 after rebate and Mwave for $199, so there -might- be something out there that can beat them ;-)

As for stability, you should really try running MemTest for Windows as well. Super Pi is a lame stability benchmark, in my opinion. I find it's better for CPU overclocks than memory clocks. 3D Mark 06 is another good one to finish off with.

http://hcidesign.com/memtest/ (two instances, 900mb per)

 

[MakubeX]

I didn't mean to make it sounded like you didn't have a fan, I just mean you shouldn't need one at that voltage. They would be warm, but not finger numbing hot.



If I have time today, I'll see what the Tracer can handle at 600MHz. NewEgg sells those for $175 after rebate and Mwave for $199, so there -might- be something out there that can beat them ;-)

As for stability, you should really try running MemTest for Windows as well. Super Pi is a lame stability benchmark, in my opinion. I find it's better for CPU overclocks than memory clocks. 3D Mark 06 is another good one to finish off with.

http://hcidesign.com/memtest/ (two instances, 900mb per)

Yeah, I know SuperPI doesn't prove stability, let alone at 1M, but I wasn't going for stability. I just wanted to show that it could at least run SuperPI and how fast. It wasn't probably not 100% stable at those clocks and I wasn't about to try and prove it as didn't want to risk corrupting files.

I usually only test overclocks I plan to keep for for 100% stability.

I didn't realize the Tracers were so cheap (not from newegg cuz I don't do MIR, but $205 at Mwave! Sweet!) Then there definitely could be something better for the price, but I won't say there is until I see better results with a retail pair, not one provided by Crucial.

 

[Rob Williams]

I don't believe Crucial provides cherry picked modules, simply because the kit I have probably would have performed even better. I could be wrong though, but it wouldn't surprise me if they were one of the few who didn't send out picked kits. That said, I'd like to see performance results from a retail kit as well.

 

[(cf)Eclipse]

I didn't mean to make it sounded like you didn't have a fan, I just mean you shouldn't need one at that voltage. They would be warm, but not finger numbing hot.

your Tracers must run a lot cooler than mine. at 533mhz with 1.85v, mine are bordering between very warm and almost hot to the touch. Thermal gun puts it at 47-49c after a pass or two of test5 in memtest. 600mhz at 2.1v would definitely be unsafe without a fan IMO


But I agree, I also believe that Crucial does not provide handpicked samples. After all, you get a reciept just like you had purchased memory from the Crucial store

 

[Rob Williams]

Color me wrong then. I should really do more of this testing before I speak my mind.

"But I agree, I also believe that Crucial does not provide handpicked samples. After all, you get a reciept just like you had purchased memory from the Crucial store."

That's a good point, I didn't consider that. They do indeed put it through as a normal order... so I solidly believe they are not cherry picked. Just gotta wait for someone to pick up a kit and verify ;-)

 

[MakubeX]

I really do hope you guys are right about Crucial. Sending random samples to reviewers is the way it should be done.

Yeah, I'm always more comfortable having airflow over my RAM. The amount of voltage you set your RAM to doesn't determine the heat alone. You also have to take into account the other factors like frequency. Having low temps always helps to gain more headroom for overclocking.

 

[Rob Williams]

I do firmly believe Crucial doesn't cherry pick... the main reason is that they are just not as concerned as overclocking as other companies. I know this because I received their Ballistix 6400 kit for review last year, and it clocked way -worse- than retail modules that others had. So... I'm pretty confident.

The amount of voltage you set your RAM to doesn't determine the heat alone.

Oh I'm aware. I believed a lot of it would have to do with the spreader itself, but I had it all wrong. I'm an idiot, I won't sugar coat.

 

[MakubeX]

I do firmly believe Crucial doesn't cherry pick... the main reason is that they are just not as concerned as overclocking as other companies. I know this because I received their Ballistix 6400 kit for review last year, and it clocked way -worse- than retail modules that others had. So... I'm pretty confident.

Yeah, you're probably right. Guess we'll be sure once we see some results from retail modules.

Oh I'm aware. I believed a lot of it would have to do with the spreader itself, but I had it all wrong. I'm an idiot, I won't sugar coat.

lol, don't be so hard on yourself.

 

[(cf)Eclipse]

The amount of voltage you set your RAM to doesn't determine the heat alone. You also have to take into account the other factors like frequency. Having low temps always helps to gain more headroom for overclocking.

yes, but voltage changes definitely dominate the change in heat output for ddr2. don't forget drive strengths too. increasing the drive strengths have made more of a difference for me than mhz has in the past

 

[MakubeX]

yes, but voltage changes definitely dominate the change in heat output for ddr2. don't forget drive strengths too. increasing the drive strengths have made more of a difference for me than mhz has in the past

Yes, of course voltage has more influence, no doubt about that. As for drive strength, I can't change it in my BIOS.

 

[Rob Williams]

As for drive strength, I can't change it in my BIOS.

dat be cuz u 2 weak!

aahahahaha

(No I am not drunk)

 

[(cf)Eclipse]

As for drive strength, I can't change it in my BIOS.

time to get a new board!

 

[MakubeX]

time to get a new board!

Hahaha, naa I like this board.

 

[Rob Williams]

You that board?

No I agree, I love the board as well. I do find it occasionally quirky though. I am overclocking a 4GB kit now, and it's going nowhere. It's stock 5-5-5 2.1v, so I tried 4-4-4 2.4v and the computer would beep right after reboot. The readout on the board would show a C1 error, even though nothing is really wrong there.

Resetting the BIOS usually works, but I don't know whether it's the board or the memory to blame.

 

[MakubeX]

You that board?

Lol, didn't notice the missing verb.