- Oct 14, 2003
- 8,686
- 3,787
- 136
http://news.mydrivers.com/1/211/211710.htm
Compared to my 2600K
Cinebench R11.5
ST: +10%
MT: +10.9%
Fritz Chess
8T: +5.9%
1T: -6.5%
3DMark06: +34%(in line with Intel's estimates of 30% in 3DMark06)
Looks like Ivy Bridge has 7-8% gains in Cinebench per clock, since the 3770K has 100MHz higher Turbo for ST. Something is obviously wrong on Fritz Chess, but its an early platform so that's understandable. I wouldn't fault the score at low Turbo mode as Cinebench ST performs pretty well.
Average IPC I think will be 4-6%, Similar to what Penryn with had over Conroe. Of course I don't include games in my analysis as it varies too widely between setups to be repeatable(plus its dependent on resolution and GPU used), the #1 rule in comparing systems. Total average gain with better Turbo mode would end up being just slightly under 10%.
Intel's results showed 15% gain in Cinebench, that's also entirely possible as some 2600K owners report little less than what my system got. Of course that's almost the best case scenario. Most applications won't behave like Cinebench.
Compared to my 2600K
Cinebench R11.5
ST: +10%
MT: +10.9%
Fritz Chess
8T: +5.9%
1T: -6.5%
3DMark06: +34%(in line with Intel's estimates of 30% in 3DMark06)
Looks like Ivy Bridge has 7-8% gains in Cinebench per clock, since the 3770K has 100MHz higher Turbo for ST. Something is obviously wrong on Fritz Chess, but its an early platform so that's understandable. I wouldn't fault the score at low Turbo mode as Cinebench ST performs pretty well.
Average IPC I think will be 4-6%, Similar to what Penryn with had over Conroe. Of course I don't include games in my analysis as it varies too widely between setups to be repeatable(plus its dependent on resolution and GPU used), the #1 rule in comparing systems. Total average gain with better Turbo mode would end up being just slightly under 10%.
Intel's results showed 15% gain in Cinebench, that's also entirely possible as some 2600K owners report little less than what my system got. Of course that's almost the best case scenario. Most applications won't behave like Cinebench.
Last edited:
Facebook
Twitter