Thursday, March 09, 2006

Tech Link (Industry): "We benchmark FEAR on Conroe"

A short article is posted over at bit-tech. It's a nice little write up, though they seem to show some skepticism about Conroe's performance. Well, they never really said anything bad, I guess they just want to play safe or maybe they are just AMDroids not wanting to earn the wrath of AMDFanboys ;)...after all, AMDFanBoys have been used to comparing X2 4xxx+ vs any of Chipzilla's chips (which, any of the CPUs right now aren't equally match. Always compare GHz vs PR boyz)!
Image owned by bit-tech

We managed to grab some time with a Conroe machine that was up and running this morning.

Intel would only run the benchmarks that have been seen elsewhere on the web for us, but we managed to sneak a quick go at FEAR and we configured it the way we like it, before being marched onwards.

Our previous CPU reviews have all utilised a FEAR demo running at 640x480, with no graphical candy, to get straight down to CPU performance without being graphics bound. We configured the Conroe system to do the same, and were pretty gobsmacked with the result.

The system we looked at had a 2.6GHz CPU with DDR2-667 RAM, which - when Conroe finally launches - will be a mainstream part. New Extreme Edition chips will be clocked up at 3GHz and over, using a 1.33GHz bus and DDR2-800 RAM. The graphics cards were dual X1900s in CrossFire.

(Yes, we felt the pain of running X1900 CrossFire at 640x480, don't worry).

We also used an Intel configured AMD system, which was a 2.6GHz FX-60 overclocked to 2.8GHz and configured with low latency DDR400 memory.

The Conroe result speaks for itself, pretty much - we've included our own reference lab numbers from back home for comparison.

Of course, we'll take these results with a pinch of salt until we get to play with our own systems in the labs, which Intel has intimated will be sooner rather than later. We can't wait!

Discuss these results in the forums!

Disclaimer: Let's just make sure we understand something here: this was a quick benchmark, grabbed in a room, with systems that we didn't control the set up of. We are not presenting these numbers as defining (unlike some other sites): just as an interesting look at what we might be able to expect. Please don't over analyse them.


No comments: