Conroe and X2 comparison

[Stelly]

Pixie Dust

They have hired tinker bell? They are on the ball

Stelly

 

[Cyber-Mav]

Does anyone know if running a divider on the conroe is a performance killer?

G

it may hurt performance. they should have used on die memory controller. then the cpu would be truly a power monster.

 

[masterk]

The three key reasons are:

- More execution units (like 3 vector units, as opposed to just the 1 on NetBurst and the A64)

- Pipeline optimisations (compresses certain instruction groups down to one instruction)

- Shared L2 cache (allows cross core communication without hitting the FSB, and hence allows threads to switch between cores with no penalty)

- Black magic that Intel won't ever disclose

Great thanks.

So its a bit like Pepsi, with the most important ingredient being top secret?

 

[silversurfer]

Dr Pepper > Coke > Pepsi


Obviously the 6600 is the one to get but what is a 6300 equalivant to at stock, is it still in fx beating country?


And there is an even cheaper version thats 800fsb, except I dont see that on ocuk?

 

[Darkwave]

Dr Pepper > Coke > Pepsi

Coke > Pepsi >>> Dr. Pepper (which tastes rather like urine.)

Obviously the 6600 is the one to get but what is a 6300 equalivant to at stock, is it still in fx beating country?

I've been looking at some benchmarks today and it seems that he E6300 lags just a tiny bit behind the FX-62 at stock, but I dare anyone to say that's not incredible value for money. The E6400 is equal, maybe a tiny bit better, while the E6600 and X6800 are just miles ahead.

 

[Arthalen]

Just wondering if the 800FSB part is going to be the new 805 chip ..... *grin*

 

[silversurfer]

I was just looking at someone on xtreme who has clocked his 6300 to 6700 speed which is pretty impressive, 50% overclock.

There isnt any 20x multipler chip, theres a grid showing all the models on one of these threads and theres no sign of anything like that unfortunately. So to overclock needs expensive ram and a good motherboard it seems

 

[Mattus]

Does anyone know if running a divider on the conroe is a performance killer?

I doubt it. The Conroe is pretty similar in design to the Pentium-M (short pipelines, etc.) That made the P-M a low-bandwidth design, so it performed perfectly fine on single-channel 333Mhz DDR. If you're running eg. dual-channel 800Mhz DDR2 with Conroe, I think it's gonna have more bandwidth than it knows what to do with, divider or not. It's certainly not like the P4 which had a long pipeline and was very dependent on bandwidth - that's why dividers on P4 are such a pain.