That's the only thing 1:1 can mean. The memory clock is not derived from the HTT clock, it is derived from the CPU clock by using a divider (which, when you set your DDR speed to 400MHz is equal to the CPU multiplier).
That's the only thing 1:1 can mean. The memory clock is not derived from the HTT clock, it is derived from the CPU clock by using a divider (which, when you set your DDR speed to 400MHz is equal to the CPU multiplier).
Got the Newark (in a private sale) from a user named Vandi423 over at XS forums. Perhaps the best CPU I've ever owned.
Where did anyone mention ' derivation ' ?
One of the following is NOT = 1: 1
Can you tell which one?
A. 200 : 200
B. 242 : 242
C. 234 : 234
D. 200 : 166
1:1 = a:b = Term : Term = Anything:Another Thing
If the HTT is set to 200 and the memory is set to 200 ( one of the BIOS options on many A64 mainboards ); they have a 1:1 relationship..
What is even more amazing, is that as you raise the HTT the memory clock will follow it, and the relationship will continue to be 1:1 ...
So perhaps you could explain how this fits the idea that the ONLY thing 1:1 can mean, is that the CPU clock is equal to the memory clock ?
I wish BIOS writers would just get rid of the memory config nomenclature of 1:1(200), 9:10(180), 5/6(166), etc. It causes too much confusion. Get you calculator out.
Lets say your CPU clock is 2000mhz:
BIOS setting = CPU clock/memory integer = memory speed
1:1 = 2000/10 = 200
9:10 = 2000/11 = 181.82
5:6 = 2000/12 = 166.67
3:4 = 2000/14 = 142.86
2:3 = 2000/15 = 133.33
My point is, you shouldn't look at the "memory clock" as a ratio of HTT, but as a memory integer divided into the CPU clock.
EDIT: Sorry topjimmy, my intention was not to steal the thread.
Don't be fooled as I own both the 165 Opty and 3800X2 with the 3800X2 winning hands down in response and benches...put em in the same Ultra NForce board and have an ULi sitting here so i try again
lopri
Elite Member
- Jul 27, 2002
- 13,310
- 687
- 126
People who says 165 and 180 are same are somewhat exaggerating. Clock is clock and binning is binning. Yes they are same chips but with statistically different potential. Question yourself if you would pick a 165 over 180 if given a chance to pick one (for free).
I'd rather say what matters is the stepping. Unfortunately getting a decent stepping is purely a matter of luck.
Thinking about it, I have no idea how the hell AMD bins their chips. What are they thinking? What are the criteria? I mean, their binning on Socket 939 seems almost intentional.
I'd rather say what matters is the stepping. Unfortunately getting a decent stepping is purely a matter of luck.
Thinking about it, I have no idea how the hell AMD bins their chips. What are they thinking? What are the criteria? I mean, their binning on Socket 939 seems almost intentional.
I doubt that is the case if they are running at the same clock. At the same clock speed they should be about the same, since they are essentially the same chip.
The main reason to buy an Opteron is its potential for overclocking at a given price point. That's not to say that any specific X2 3800+ could not clock higher than any specific 165. It's a crap shoot and it's all about shortening the odds in your favor.
I'm guessing that there are two factors that dictate binning policy. 1) how the chips physically perform, and 2) the demand for each model. After sorting the chips into their tested performance categories, some of the higher performing ones could be demoted to meet demand for the lower models. The best chance of this happening is when overall demand exceeds production, which has been the case for the past few months. As demand for the Opterons dies off somewhat, and production ramps up, you probably have less chance of getting a high clocker in the lower models.
Of course, I could be completely wrong.
Zim of course the Opty with the extra cache will beat the 3800x2 @ same clock but considering the price one must pay for the better RAM to OC the Opty @ the same speed ...my monies on the 3800X2 which will do 2.5GHz ...same as the Opty 165 on Air providing you have the RAM which will go @ just under 280...if not and the scenerio of RAM tops @ 250, then the 3800X2 @ 10X250 will win hands down vs the Opty @ 9X250
Just the right type of people we need to hear from. Someone that has bench tested both chips on the same hardware. But only one result is difficult to go on, for as many proponents there are for the 3800, there are oponents of the 3800 in favour of the 165. General consensus the extra L2 cache makes didly squat difference to overall system performance. Am I right here?
I would go for the 3800 due to cooler running (smaller L2 cache) and an easy OC to 2.4~2.5 and maybe even higher with the new steppings.
^
what... what L2 cache got to do with heat. also most opterons run cooler than most x2. (key word is most not all)
also the 165 ships with a very very good heat sink that compares (and beats sometimes) expenmsive after market solutions.
what... what L2 cache got to do with heat. also most opterons run cooler than most x2. (key word is most not all)
also the 165 ships with a very very good heat sink that compares (and beats sometimes) expenmsive after market solutions.
The larger L2 cache of the Opty DC's produce a great deal of heat...that's why they have so many more Million's of transistors in the silicon.
They have more transistors because the cache not because of heat. If it does increase the heat it is a very small increase.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 23K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 18K
-
-
-
Facebook
Twitter