>>> CAS Clocking <<<

[nick1985]

good thing mushkin is 2-2-2-7 stock... what other brands are like that

 

[Technonut]

Originally posted by: nick1985
good thing mushkin is 2-2-2-7 stock... what other brands are like that

That is for their PC2700..... Mushkin specs their High Performance PC3200 DDR @ 2.5-4-4, and that is only available in 256MB sticks. Their 512MB sticks of "regular" PC3200 is speced at the same timings. (2.5-4-4) Not really what I would call impressive.

 

[Killrose]

Originally posted by: Technonut

Originally posted by: nick1985
good thing mushkin is 2-2-2-7 stock... what other brands are like that

That is for their PC2700..... Mushkin specs their High Performance PC3200 DDR @ 2.5-4-4, and that is only available in 256MB sticks. Their 512MB sticks of "regular" PC3200 is speced at the same timings. (2.5-4-4) Not really what I would call impressive.

They seem to have better luck than that...
Here

 

[Technonut]

Originally posted by: Killrose

Originally posted by: Technonut

Originally posted by: nick1985
good thing mushkin is 2-2-2-7 stock... what other brands are like that

That is for their PC2700..... Mushkin specs their High Performance PC3200 DDR @ 2.5-4-4, and that is only available in 256MB sticks. Their 512MB sticks of "regular" PC3200 is speced at the same timings. (2.5-4-4) Not really what I would call impressive.

They seem to have better luck than that...
Here

I did not say that the Mushkin could not run well... My point is that one would have no basis for an RMA past the speced timing, where with say, Corsair XMS you do. Hell, I have a couple sticks of Kingston PC2700 Value RAM (Winbond BH-6 chips) that will boogie @ 450MHz 2-7-3-3 (Enhanced). Problem is, I could not bitch one bit or expect Kingston to RMA the RAM past it's rated PC2700 specs. Same with Mushkin....

 

[Killrose]

Yeah, it's no good if you can't warranty it like the stick of Nanya I have. Nothing quite like an easy Lifetime warranty policy.

 

[THUGSROOK]

The Mushkin completed benchmarks at its stock speed and voltage perfectly (I have never seen a stick of Mush that did not). Ramping up the voltage to 2.7 allowed me to complete testing at 453Mhz, an incredible 53Mhz over its rated 200Mhz clock speed. Clearly Mushkin are using a PC3500 quality part for their PC3200, but as it costs the same or more as everyone else?s 3500 there is little benefit.

i dont exactly agree with their testing methods here.
running benchmarks and actually being 100% stable are 2 completely different things!

YOU NEED TO RUN MEMTEST86 OR THE RESULTS ARE WORTHLESS ~ am i making myself clear?

 

[THUGSROOK]

if anyone on an AMD system can run these tests the same way i did, ill be happy to add those results to the main post.

ill assume 166fsb would be best speed for an AMD test?

any takers? LMK

 

[Soulkeeper]

too bad there is no vmem adjustment on the asus a7v333
or I would run these benches on my uncle's comp
his ocz memory won't hold the fast memory timings at 166 like it was sold to run at
ohh well

 

[THUGSROOK]

actually on many intel motherboards ~ 2.6v is default vdimm voltage.

333ddr cas222 2.6v should be a piece of cake, no?

 

[dxkj]

I will run the benchmarks when I finish putting my new system together. I still havnt ordered my ram yet though, trying to decide exactly which Corsair PC3200 model to get

 

[poMONKey]

Originally posted by: dxkj
I will run the benchmarks when I finish putting my new system together. I still havnt ordered my ram yet though, trying to decide exactly which Corsair PC3200 model to get

you mean which Corsair PC3000 to get

 

[Blain]

Wow, I would have thought the benches would have been spread wider based on those memory timings!

 

[THUGSROOK]

Blain ~ na UT2k3 demo is mean motha :Q
those are actually big increases for that game.

 

[JSSheridan]

Originally posted by: THUGSROOK
if anyone on an AMD system can run these tests the same way i did, ill be happy to add those results to the main post.

ill assume 166fsb would be best speed for an AMD test?

any takers? LMK

I'd be happy to. I've tested this memory already, and I know that running all test in memtest86 overnight at 133MHz, CAS 2/2/2/4 @ 2.6v turned up no errors. I'm going to get too it. I'll edit this post when I get results. Peace.

 

[THUGSROOK]

excellent!
those results are showing the same behavior (and same odd behavior) as my results did.
cas2224? neato

i pulled the results upto the main post.

thx

 

[THUGSROOK]

if anyone else wants me to post their results (using my paramiters) in the main post, just toss me a PM or post the results here.

results from an AMD @ 166fsb would be nice
a P4 @ 150fsb on rd4200 would be nice too.
granite bay @ 150fsb or 166fsb also.

 

[THUGSROOK]

ive got a few more testers ~ and will be posting their results as they come in

 

[Mikki]

nice...we should be doing more of this cross-platform testing...

 

[Excelsior]

Yeah, I pmed Thugs my 166Mhz FSB results. I found some interesting things also. Apparently 2T command rate is faster, at least for me, than 1T. 2.0-3-3-6 is faster than 2.0-3-3-5 by little bit.

 

[THUGSROOK]

added more results...

AMD TEST SYSTEM #2 as tested by Excelsior ~ thx

 

[Excelsior]

Keep in mind, I can do any more if you want me to. I also had 16x Aniso on. Im thinking about upping the clock speed to see jus how much of a difference it makes.

 

[THUGSROOK]

thx Excelsior

system config is really more important then CPU mhz (ahem, i mean ghz) for this particular test.
i could have run my 1.7m at pretty much any speed i wanted ~ but i think the more general "system" results will suffice.

anyone have something besides XMS memory?
that would be interesting.
i think PJ is getting some OCZ 3500.... (hint!)

 

[Toymaker]

This may help to explain it:
Lost circuits
This was written for the i815 chipsets. But, Thugs's experiments show that ddr doesn't make a difference. As frequency increases, in most cases, you'll need the extra cycles.

 

[THUGSROOK]

thx for the link ~ a very good read

 

[THUGSROOK]

got a few more ppl running tests ~ ill post results as they come in