[firebird]

@ stephen

Intel Conroe E6600 @ TBD GHz, ASUS P5W DH, OCZ (667) special ops 2GB RAM , Asus 7600 512- P4 2.6 1gig FX5900 - Fragbox P4 3.2 2gig GTX7800
Rhino, 3DSMax, Lightwave, Softimage XSI
Rhinoman/AIR, Flamingo, Maxwell

This is a very nice set up!

you shouldn´t have any problems to go to 3120 MHz, since we both got the same cpu and mobo!

keep me up to date about your experiences with overclocking!

[numerobis]

I haven´t made it yet to overclock @30% and run the RAM with PC6400!

...I see your RAM is gskill pc6400 cl4 and you are only at 260MHz with cl5 at the moment. I think this chips should be able to run at your 346MHz 1:1 with cl4 - easily. You are now far under its specifications.

[firebird]

any advice to get this solved, running my ram 1:1 with the CPU ???

thx

[numerobis]

so have you tried it allready and it doesn't work? then you should check your vdimm... is the 1,95v in ai booster standart? (auto?) - your ram should run fine at ddr2 400 cl4 with auto settings. if it isn't stable at this level you can up your vdimm 0.1v and test this... but it should also be stable at auto.
what power supply do you have? are there more than one power plugs on the mainboard?

[deoli]

sorry to interrupt.

as I'm planning to get an ASUS P5W DH with an E6600 pretty soon, I was following this topic with interest. but I just can't get my head round this ram malacky.

am I correct in assuming that I'm better off getting a PC2-6400? or does this depend wether I want to overclock or not?

[firebird]

so have you tried it allready and it doesn't work? then you should check your vdimm... is the 1,95v in ai booster standart? (auto?) - your ram should run fine at ddr2 400 cl4 with auto settings. if it isn't stable at this level you can up your vdimm 0.1v and test this... but it should also be stable at auto.
what power supply do you have? are there more than one power plugs on the mainboard?

No, the 1.95 V it set manually in the bios, and then switch back to the prests, where everything is on auto, hence you can´t change that value anymore, but the bios still remebers the 1.95 V !?! On the internet it said, the my ram runs fine with 1.9 to 2.0 V!

@ deoli

I don´t know what big difference is between pc 6400 to pc5300, but the pricings are almost the same, except you go for the state of the art!

I will give it further tests tonight, since I am here at work stuck eith my G5 imac! But OSX is a pretty cool OS!

cheers

[-Adrian]

sorry to interrupt.

as I'm planning to get an ASUS P5W DH with an E6600 pretty soon, I was following this topic with interest. but I just can't get my head round this ram malacky.

am I correct in assuming that I'm better off getting a PC2-6400? or does this depend wether I want to overclock or not?

It depends on whether you want to overclock. Any headroom which you don't make use of is wasted as PC5300 and e.g PC8000 share the same architecure and perform equal at equal speeds. PC6400 is supposed to top out at 400MHz which gives you 3600MHz 1:1 on an E6600. If you want you can compensate for that you may run with a lower multi (9*400 -> 8,7,6*400) or a memory divider, but highend ddr2 is very expensive and you sure won't feel the bandwidth increase much in maxwell.

My recommendation is to go with PC6400 value ram ignorning good latencies and instead spend it well on the other components or another 2GB, or icecream.

I didn't see any solutions to the thread question. I know that the 6600 is HT capable, so why is there no difference between 2 and 4 threads?

I'm pretty sure it's not, they planned on including it, at least in the top range, but dropped it pretty soon. So no HT for any Core2Duo chips. Even if it worked, i was so pissed with Intel's netburst campaign that i'm glad it's gone

[deoli]

Thanks for your quick answer. I'll see how much money I've got left at the end of the month and then get it, or not

[-Adrian]

same here btw, need to get a new system sometime

[firebird]

well, i just set the RAM voltage to 2.2V, and it runs now on pc6400, maybe i set it down to 2.15! depends on how stable this system runs, but everything is fine right now!

[numerobis]

...now this is more than pc6400 (433MHz at FSB:DRAM=4:5) - so it's already slightly overclocked.

...G.Skill 2GB RAM (@ 867 MHz, CL ???)

CL means "CAS Latency" - your CPU-Z screen shows that you are runnning at CL5. Don't know why your RAM can't do 346MHz 1:1 with cl4 at 1.95v... but I think 433MHz at cl5 should be better.
Now I would run a prime95 (sp2004) test, to see if it is really stable - "blend test" for testing the RAM. Check your chipset- and RAM-temps, 2.2v could be critical without cooling the RAM.
ITE Smartguardian is a good tool to watch the temps/voltages and set the fanspeeds. But I don't know if it works with the ASUS P5W.

[firebird]

yep, I know what CL means, but I don´t know on which CL it is running currently, propably 5 as you said! yep I will set the Voltage a bit lower and see how it works then!

the RAM speed rose with rising Voltages! there was a difference between 1.95 and 2.05, but no difference anymore to 2.2V, and by the way, I set voltages in the bios to 2.15, and the AI booster showed 2.2 ??? weird!


what do you mean by RAM cooling, just these coloured metalsheets around the ram or a heavier stuff! ???

prime95??? I am running on xp64, I give these programms all a try and will post the results later on!

[firebird]

an 2.4 GHz opteron is about 800 Euros, an Conroe 2.4 GHz is about 300 Euros!

any questions?

my conroe (2cores) @ 3120 MHz runs the speedtest in 1h17m...
the best result of an dual stack opteron 275 (4cores) runs the speed test in 53m..

[tom]

So I did another speedtest with V 1.0

the results are

completed in 1h37m11s
BM 44.14

so, why, if V1.1 is faster than V1.0, does my machine render the speedtest scene in 1h 37m with V 1.0 and 3h44m with V 1.1!

any clues?

thx

Hi firebird,

Because SLx of 1.1 is not equal to SLx of 1.0. Comparing with same SL between different versions doesn't make sense. The only benchmarking critera between different versions is the time. You should run both for the same time and compare the output. In your example, the quality of SL12 on 1.1 in fact corresponds to a higher SL with 1.0. That's why it took lots of time than you've expected.

Best regards,
Tom

[ludi]

I don't see a (significant) difference between your renderings. So, may be you could post a comparison with the same render time, as tom said.