[sccr64472]

Quote:

Originally Posted by tytlyf the 2 systems will probably perform the same, but I would say the memory controller is better on the opteron due to the server specifications and testing....that's the only difference. The opterons do clock as high as sandiegos and higher... Example: Take 100 opty 148's and 100 3700's. I would guarantee you that the % of higher clocks would belong to the opterons. I bet 50 optys would be stable at 3.0, and 20 sandiegos. Get off your high horse wax, you dont know what you are talking about.

That's not the only difference. Find 2 people with identical systems to test. I've seen it before but can't find the link. The Opterons have relaxed memory controller timings. Very close, but definitely a tad slower. Anyone here have a system similiar to someone else except San Diego versus Opteron willing to do a quick run?

 

[erauman]

Quote:

Originally Posted by tytlyf the 2 systems will probably perform the same, but I would say the memory controller is better on the opteron due to the server specifications and testing....that's the only difference. The opterons do clock as high as sandiegos and higher... Example: Take 100 opty 148's and 100 3700's. I would guarantee you that the % of higher clocks would belong to the opterons. I bet 50 optys would be stable at 3.0, and 20 sandiegos. Get off your high horse wax, you dont know what you're talking about.

Dude, they are just asking for proof of your statements. They aren't saying either chip is better than the other, only you are saying that. Why don't you just enlighten everyone by showing us where you are getting this information that optys will most likely overclock better. And Sccr64472 is right, from what i've heard in the past, the memory controllers on the 3700+ are slightly better.

 

[tytlyf]

where i get my information is from multiple forums and here. It's not hard to search around and see what people are achieving with their chips.
That's the proof, do a search....if a sandiego hits 3.0 stable, its an achievement and most people give props, if an opty hits 3.0, it's overlooked because its very common....open your eyes guys...geezuz

 

[sccr64472]

Quote:

Originally Posted by erauman Dude, they are just asking for proof of your statements. They aren't saying either chip is better than the other, only you are saying that. Why don't you just enlighten everyone by showing us where you are getting this information that optys will most likely overclock better. And Sccr64472 is right, from what i've heard in the past, the memory controllers on the 3700+ are slightly better.

Huh? I never said the memory controller was better. I just said that the Opterons have slightly relaxed timings. That will make the benchmarks a tiny bit faster, but it doesn't make it "better." There's always a tradeoff between reliability and performance.

 

[erauman]

From what i've "heard on forums", the new steppings of 3700+ are hitting 3.0 on stock coolers....
Btw...why is yours only at 2.9ghz, with the help of a nice z9500, the integrated heat sink removed, and probably some as5?

 

[erauman]

Quote:

Originally Posted by sccr64472 Huh? I never said the memory controller was better. I just said that the Opterons have slightly relaxed timings. That will make the benchmarks a tiny bit faster, but it doesn't make it "better." There's always a tradeoff between reliability and performance.

You are arguing semantics now, i'm sure you can agree that tighter timings = better system performance, and that's what we are talking about right? I didn't think this discussion had anything to do with 'reliability', maybe stability at excessive core speeds,.

 

[sccr64472]

Quote:

Originally Posted by erauman You are arguing semantics now, i'm sure you can agree that tighter timings = better system performance, and that's what we are talking about right? I didn't think this discussion had anything to do with 'reliability', maybe stability at excessive core speeds,.

I'm just saying that "better" isn't a good word to use because the Opterons' intended use is server applications and is the reasoning for the timing adjustment. They're not inferior or superior, they've just been adjusted to suit their intended use better. Not arguing with your semantics, I just don't agree with the word "better" in this case, even though it's obvious now that you meant faster.

 

[tytlyf]

Quote:

Originally Posted by erauman Btw...why is yours only at 2.9ghz, with the help of a nice z9500, the integrated heat sink removed, and probably some as5?

2.9 is plenty enough speed for me and i can run it stable with 1.5v. Voltages are lower so things are cooler. Ever since I added this second hard drive in raid things have been getting a little weird, like my chipset temperature...
I can run 3.0 just fine and loads dont break 46c. But it requires 1.6v for prime stability. So everythings warmer at that speed.
If I really wanted to I could run 3.1 prime stable, i've already posted my 32M stability at that speed. That requires 1.66v.
This chip can handle some serious voltage....I've pushed it close to 1.8v (dont try this at home) trying to reach its limits....

 

[Mother Lover]

I have ended up getting a 3700 KAB2E PIB for £135. I am totally happy with it as I would have to spend about another £20 for an opty 146 CACJE which I have read reports of only getting to 2.7-2.8 averagely. I think there was a good reason for the 'opteron fad' as some have called it due to initial cheap prices and great steppings compared with previous 3700 prices. This all now has changed. Reports are that KAB2E's are great (especially when compared to CACJEs) at the same price. 148s might kick ass but then you are talking £200. Thats nearly £70 more.

The extra cost I did not think was worth it when the results are debatable anyway. Even if optys are better, they now cost a little more and have a worse stepping.

I will let you know what I can get my KAB2E on stock before getting a Big Typhoon.