Here's my E6600@3.6g running single client.
Here's the dual client speeds.
Printable View
Here's my E6600@3.6g running single client.
Here's the dual client speeds.
Guru
What are the chances of a retail chip being able to run at that speed.
Is it cherry picked ?
I got mine from the retail channel. It's also watercooled but with a good cooler and the right settings most should be able to get close to this speed. I did have a pre-production X6800 that ran stable at 3.73Ghz. The production steppings seem to overclock better then the pre-production steppings. Mine is not 100% stable but I think that's due to my waterblock. I need to reseat it. I overtightened the block and it's not making the best contact. If you want the best air cooler get the Thermaltake CL-P0114 Big Typhoon. It's very quiet and worked just as good as my watercooling. The only benefit I got from watercooling is a cooler video card and chipset. But I now have to live with more noise.
How can you tell from the Pic's that he is rinning fast or slow?
Or am I just slow:looney:
look at the text line above the pictures ... :blush:Quote:
Originally Posted by LAURENU2
You can't meen (Here's my E6600@3.6g running single client.)Quote:
Originally Posted by tnerual
That is the only text I can find above the pictures posted
I guess I am Slow:bang:
Lauren
Do you mean how can you tell how fast he is doing blocks ?
Look at the cEMS/sec 3rd box down right hand side.
The single client is doing 8 million.
With 2 clients running they are averaging about 6.5m each.
Thanks PCZ that helpedQuote:
Originally Posted by PCZ
I just compared it to the new X2 4400 I just setup
I get a score of 7,476,822.1 running 1 SOB client and 2 DPAD clients
No OC with air cooling
With the 3800 I only get 1,611,896.4 With the same clients rinning
I guess that extra 1.5 MB L2 conchthe 4400 has Does make a big difference
SoB seems to like bandwidth. Lots of cache and memory bandwidth. I have one 4 channel FB-DIMM system running 4 cores at 2.3Ghz and I get ~4.5 M per thread. With 4 3Ghz cores and two channels I was only getting ~3.5M per thread. When I dropped the clients down to two it jumped up to 4.6M per thread. So instead of 2 quad core systems I split them up to 4 dual core system. It seems that SoB needs one memory channel per core, atleast with Intel's new core cpu's.
silly question but How can u get SoB to run on client 1 and client 2 without running them as a service, is that possible.
There has been talk that FB-DIMMs are slower than DDR2 and
something about only getting the equivelant of 3 processors power out 4 cores...